RoboHike: Autonomous Quadrupedal Robot Navigation and Hiking in Challenging Rough Terrains
Lead Research Organisation:
UNIVERSITY COLLEGE LONDON
Department Name: Computer Science
Abstract
Quadrupedal robots are gaining important capabilities, especially over the past decade, due to the rapid advancements in mechatronics, control, and planning. In scenarios that robots need to operate for either inspecting hard-to-reach areas or aiding humans in dangerous and hazardous environments, quadrupedal robots could be ideal due to their ability to deal with sparse footholds in a safe and energy efficient way. To date, quadrupedal robots are able to traverse some types of rough terrain, using usually traditional control and perception methods. However, their mobility is still far behind their natural counterparts, especially in cases that the environment is dynamically changing. Tasks such as navigating and hiking rough or rocky trails, where the environment itself is uncertain, not fully perceived, and potentially dynamically changing, remain central challenges in legged robot locomotion. RoboHike aims at introducing and developing novel high level and platform-agnostic perception and learning approaches for modeling, identifying, and mapping footholds for quadrupedal robots, such that it would be possible to achieve fast, robust, and reliable navigation and hiking skills on challenging terrains. In particular, it aims at combining various sensing systems, such as proprioceptive (e.g., inertia, speed, or joint torques) and exteroceptive (e.g., visual, range, event, or foot's force contact data) perception to reconstruct the environment and handle the uncertainty of potentially missing or inaccurate data, before and during locomotion, especially for dynamically changing terrains. This will enable novel footstep planning and robot localization in the environment. Analytic and (self-supervised and reinforcement) learning methods will leverage multi-modal sensing to allow quadrupedal robots mimic the way that animals plan footsteps when learning to walk. The developed methods will be validated experimentally on several full-size quadrupedal robots, in academic and industrial real-world use cases, for tasks such as inspection, patrolling, and maintenance. RoboHike will work towards the next-generation autonomous robotic systems in construction fields, oil&gas sites, or damaged sites after a man-made/natural disaster, where efficient navigation is required, and rough/rocky terrain, industrial stairs, pipes, and narrow passages may exist. The vision is to endow quadrupeds with environment cognition for the benefit of the public in autonomizing manual labor of hard or dangerous tasks. The impact is expected to be high in the national and industrial sectors for automated inspection, monitoring, maintenance, and disaster innervations, where terrain is arduous and the requirement for timely intervention is paramount. We intend to construct publicly shared benchmark datasets on challenging trails, bringing in this way the robotics community several steps forward in robot locomotion by enabling robots to work on challenging grounds.
Organisations
- UNIVERSITY COLLEGE LONDON (Lead Research Organisation)
- University of Cambridge (Collaboration)
- University of Leeds (Collaboration)
- Shanghai Jiao Tong University (Collaboration)
- Italian Institute of Technology (Istituto Italiano di Tecnologia IIT) (Collaboration)
- IMPERIAL COLLEGE LONDON (Collaboration)
- Norwegian University of Science and Technology (NTNU) (Collaboration)
- Ktima Biblia Chora (Collaboration)
- Amazon.com (Collaboration)
- Inria Nancy - Grand-Est research centre (Collaboration)
- Sungkyunkwan University (Collaboration)
- National Institute for Geophysics and Volcanology (INGV) (Collaboration)
- Forestry England (Collaboration)
- University of Patras (Collaboration)
- ETH Zurich (Collaboration, Project Partner)
- University of Toulouse (Collaboration)
- University of Oxford (Collaboration, Project Partner)
- Chalmers University of Technology (Collaboration)
- The Hong Kong University of Science and Technology (Collaboration)
- Osaka University (Collaboration)
- ARRIVAL LIMITED (Collaboration, Project Partner)
- Biblia Chora (Project Partner)
- INGV (Nat Inst Volcanology and Geophys) (Project Partner)
Publications
Beddow L
(2024)
Reinforcement Learning Grasping With Force Feedback From Modeling of Compliant Fingers
in IEEE/ASME Transactions on Mechatronics
Bendikas R
(2023)
Learning Needle Pick-and-Place Without Expert Demonstrations
in IEEE Robotics and Automation Letters
Changkun Liu
(2025)
AIR-HLoc: Adaptive Retrieved Images Selection for Efficient Visual Localisation
Chengxu Zhou
(2022)
TeLeMan: Teleoperative Legged Manipulation based on Body IMU-Sensing for the use of Explosive Ordinance Disposal
in arXiv
Christopher Peers
(2022)
Dynamic Camera Usage in Mobile Teleoperation System for Buzz Wire Task
Daniel Chee Hian Tan
(2024)
RL-CBF: Verifying Learned Controllers using Certificate Theory
Daniel Chee Hian Tan
(2023)
Your Value Function is a Control Barrier Function
| Title | Bedtime Stories |
| Description | A robot that will draw on the artist's back while the artist will read bedtime stories. |
| Type Of Art | Artistic/Creative Exhibition |
| Year Produced | 2022 |
| Impact | Will aim at showing the impressive AI capabilities to imagine shapes when a story is read to it. |
| Title | ELMO: A UCL Science-Fiction story |
| Description | This is a science-fiction short that reflects on the themes of aging and free will. A space mission is traveling to an unknown destination. As his health deteriorates, Elmo, the oldest astronaut in the crew will discover that he has no control over his fate. |
| Type Of Art | Film/Video/Animation |
| Year Produced | 2024 |
| Impact | Excite the general public and the new generation over robotics of the future. |
| Title | The Art of Balancing |
| Description | A robot balancing while being pushed creates some impressive trajectories. We have recorded these into a painting and presented to a UCL AI competition. |
| Type Of Art | Artwork |
| Year Produced | 2022 |
| Impact | A painting made by the robot |
| URL | https://youtu.be/lINTZHzx0YM |
| Description | Through this award, we successfully built a world-leading team of academics dedicated to advancing real-world robotics. Our research is pushing the boundaries of robotic learning, enabling autonomous legged robots to handle diverse and complex tasks. This is a crucial step toward deploying robots in challenging environments where human intervention is risky or inefficient. Key applications include: - Disaster Response: Robots that can explore and monitor disaster-stricken areas (e.g., post-earthquake inspections). - Agriculture: Automation of labor-intensive processes such as harvesting and precision farming. - Construction & Infrastructure: Assisting in hazardous work environments to improve safety and efficiency. The RoboHike project has significantly advanced how quadrupedal robots perceive and navigate difficult terrains. By developing intelligent perception and learning frameworks, we are equipping legged robots with the ability to identify stable footholds and traverse natural and unstructured landscapes safely and efficiently. This breakthrough has direct applications in search-and-rescue missions, forestry management, agriculture, and construction, where robots must navigate unstable, rugged, or hazardous environments. By enhancing robotic autonomy, RoboHike brings us closer to deployable robotic assistants that can support human workers, improve safety, and boost productivity in physically demanding or dangerous settings. |
| Exploitation Route | We successfully developed the foundation for robust, autonomous navigation in quadrupedal robots, demonstrating the feasibility of learning-based perception for challenging terrains. Our work continues to evolve, further refining models and expanding applications. The outcomes of this research will be further developed over the next two years. By publishing our findings openly and engaging with both academia and industry, we aim to: - Advance Autonomous Robotics: Contribute new insights and methodologies for robotic perception and learning. - Enable Industrial Applications: Provide technologies that industry partners can adopt for safer and more efficient operations. - Increase Public Awareness: Highlight the benefits of robotic integration in agriculture, environmental monitoring, and emergency response. By sharing our research widely, we aim to influence future developments in robotics, helping industries leverage automation while ensuring safety, sustainability, and reliability. |
| Sectors | Agriculture Food and Drink Construction Digital/Communication/Information Technologies (including Software) Education Electronics Energy Environment Government Democracy and Justice Manufacturing including Industrial Biotechology Culture Heritage Museums and Collections Security and Diplomacy |
| URL | https://rpl-as-ucl.github.io |
| Description | Since its inception in January 2022, the RoboHike fellowship has begun generating significant impact across multiple domains, fostering new research directions, industrial collaborations, public engagement, and policy discussions. Below are key areas where RoboHike has contributed to positive change: 1. Establishing a New Research Area at UCL (Academic & Institutional Impact): RoboHike has played a pivotal role in shaping a new research focus within UCL Computer Science, establishing real-world mobile robotics as a core area of expertise. This has led to: ? New Academic Programs - The launch of two degree programs: MEng and MSc in Robotics and AI. ? Cutting-Edge Research Facilities - Creation of a £1M state-of-the-art robotics lab at UCL East, supporting advanced robotic development and experimentation. ? Strategic Recruitment - Hiring of new faculty members specializing in robotics and AI, positioning UCL at the forefront of robotics research in the UK and globally. The impact of these initiatives is evident in the overwhelming response to the new programs, with over 500 applications, reflecting a growing interest in robotics and AI. In recognition of these efforts, Prof. Dimitrios Kanoulas received a departmental award in July 2022 for his contributions. 2. Public Engagement & Education (Societal Impact): The fellowship has unexpectedly sparked strong interest from schools, leading to invitations for: ? Talks, demonstrations, and lab visits to inspire young students and educators, as well as robot visits to schools and nurseries. ? Shaping public perception of robotics by addressing fears and misconceptions through direct engagement. ? These outreach efforts help promote a positive and informed understanding of robotics, fostering enthusiasm for STEM careers among younger generations. 3. Robotics in Agriculture (Economic Impact & Commercialisation): RoboHike's research has attracted significant interest from the agriculture and forestry sector, particularly in areas such as automated harvesting and field operations. Several companies have expressed a keen interest in adopting elements of our technology to enhance efficiency and productivity. To maximise this potential, the team is actively exploring the formation of a start-up, aiming to transfer RoboHike's innovations to real-world agricultural applications, boosting economic impact and technological adoption. 4. Government & Policy Engagement (Diplomatic & Strategic Impact): The fellowship has gained attention at the governmental and diplomatic level, leading to multiple high-profile engagements: ? Greek Embassy in London - Several invitations to discuss research commercialisation and strategies for supporting robotics-based start-ups between Greece and the UK, involving government ministers and the Greek Prime Minister. ? Other Diplomatic Visits - the UK Labour Party and the PM himself, the Her Royal Highness The Princess Royal, Representatives from Japan, and other national and international delegations have visited our research facilities to explore potential collaborations and understand the implications of our work. ? These interactions underscore the growing strategic interest in autonomous robotics and their role in future industries. 5. Media Recognition & Public Awareness (Shaping the AI Narrative): RoboHike has contributed to public discussions on robotics and AI, with major media outlets such as BBC reaching out for expert insights. The team has actively engaged in efforts to promote a balanced and responsible perspective on AI, advocating for "AI for Good" and highlighting the potential benefits of robotics in society. By participating in mainstream media discussions, we aim to bridge the gap between robotics research and public understanding, ensuring informed debates on the future of automation. 6. Conclusion & Future Steps: The RoboHike fellowship has already demonstrated strong academic, societal, economic, and policy-driven impact, with continued momentum expected over the next few years. Our next steps include: ? Expanding industrial collaborations and launching a start-up. ? Strengthening public engagement efforts to educate and inspire. ? Furthering government and policy discussions on robotics applications. ? Advancing the research towards fully autonomous quadrupedal navigation in extreme terrains. Through these efforts, RoboHike will continue shaping the future of intelligent legged robots in agriculture, forestry, construction, disaster response, and beyond. |
| First Year Of Impact | 2022 |
| Sector | Agriculture, Food and Drink,Education,Government, Democracy and Justice,Culture, Heritage, Museums and Collections,Security and Diplomacy |
| Impact Types | Cultural Societal Economic Policy & public services |
| Description | Demo of Robotic Dogs to Greek Ambassador in London and the Greek Prime Minister |
| Geographic Reach | Europe |
| Policy Influence Type | Participation in a guidance/advisory committee |
| Impact | Continuous contact with Greek politicians about applying robotic solutions to emerging search&rescue issues in Greece. |
| Description | Deputy Director of MSc in Robotics and AI |
| Geographic Reach | Local/Municipal/Regional |
| Policy Influence Type | Influenced training of practitioners or researchers |
| Description | Designer of MEng in Robotics and AI |
| Geographic Reach | Local/Municipal/Regional |
| Policy Influence Type | Influenced training of practitioners or researchers |
| URL | https://www.ucl.ac.uk/prospective-students/undergraduate/degrees/robotics-and-artificial-intelligenc... |
| Description | Discussion with Greek Bankers, Greek Church, and Embassy of Greece in London about academia-industry connection |
| Geographic Reach | Europe |
| Policy Influence Type | Participation in a guidance/advisory committee |
| Impact | Academic team creation for changing the ways that the Greek government is considering joint Greek-UK spin-offs with financial impact for both countries. |
| Description | Gender-based violence discussed with Lib Dem mayoral candidate during UCL East visit |
| Geographic Reach | National |
| Policy Influence Type | Participation in a guidance/advisory committee |
| URL | https://www.ucl.ac.uk/news/2024/apr/gender-based-violence-discussed-lib-dem-mayoral-candidate-during... |
| Description | HRH The Princess Royal visits UCL's new campus |
| Geographic Reach | National |
| Policy Influence Type | Participation in a guidance/advisory committee |
| URL | https://www.ucl.ac.uk/news/2025/feb/hrh-princess-royal-visits-ucls-new-campus |
| Description | Prime Minister sets out AI action plan at UCL |
| Geographic Reach | National |
| Policy Influence Type | Participation in a guidance/advisory committee |
| URL | https://www.ucl.ac.uk/news/2025/jan/prime-minister-sets-out-ai-action-plan-ucl |
| Description | Round Table for Embassy of Greece in London |
| Geographic Reach | Europe |
| Policy Influence Type | Contribution to a national consultation/review |
| Impact | The two ministers of Greece were informed about ways to coordinate with UK-based start-ups in order to increase investment between Greece and the UK. Details: Roundtable discussion on bridging the Greek and British start-up ecosystems yesterday at the Hellenic Residency. Deputy Minister for Economic Diplomacy and Openness at the Ministry of Foreign Affairs, Mr Kostas Fragogiannis and Deputy Minister for Research and Technology at the Ministry for Development and Investments, Dr Christos Dimas discussed with distinguished members of the Greek Diaspora in the UK with diverse backgrounds (academia, investors, startupers, corporates and associations). The event was organised by the Embassy of Greece in collaboration with SEV (Hellenic Federation of Enterprises). We hope this to be the beginning of a series of events aiming at establishing a support network for Greek start-ups and for the future ventures of Greek researchers in the UK. |
| URL | https://www.facebook.com/GreeceInUK/photos/pb.100064680397456.-2207520000./5195041960539948/?type=3 |
| Description | Shadow Secretary of State for Science, Innovation and Technology visits UCL |
| Geographic Reach | National |
| Policy Influence Type | Participation in a guidance/advisory committee |
| URL | https://www.ucl.ac.uk/news/2024/nov/shadow-secretary-state-science-innovation-and-technology-visits-... |
| Description | UCL Provost and Mayor of Newham Showcase |
| Geographic Reach | Local/Municipal/Regional |
| Policy Influence Type | Participation in a guidance/advisory committee |
| Impact | Interest in showcasing the robot capabilities in the local borough. |
| Description | UCL celebrates strong Greek connections at ambassadorial reception |
| Geographic Reach | Europe |
| Policy Influence Type | Participation in a guidance/advisory committee |
| URL | https://www.ucl.ac.uk/global/news/2024/jun/ucl-celebrates-strong-greek-connections-ambassadorial-rec... |
| Description | UK-RAS White Paper on Legged Robotics: Agile and Dynamic Interaction |
| Geographic Reach | National |
| Policy Influence Type | Contribution to a national consultation/review |
| URL | https://www.ukras.org.uk/publications/white-papers/legged-robotics-agile-dynamic-interaction/ |
| Description | Autonomous Quadrupedal Robot Navigation and Hiking in Challenging Rough Terrains |
| Amount | £99,775 (GBP) |
| Funding ID | 2736841 |
| Organisation | Engineering and Physical Sciences Research Council (EPSRC) |
| Sector | Public |
| Country | United Kingdom |
| Start | 08/2022 |
| End | 09/2026 |
| Description | Bio-Robots: Crawl, Jump, and Slither! |
| Amount | £200,339 (GBP) |
| Funding ID | EP/W033755/1 |
| Organisation | Engineering and Physical Sciences Research Council (EPSRC) |
| Sector | Public |
| Country | United Kingdom |
| Start | 06/2022 |
| End | 06/2024 |
| Description | Continual Lifelong Learning and Adaptation for Legged Robots Planning and Control via Neuromorphic-based Learning |
| Amount | £99,775 (GBP) |
| Funding ID | 2728765 |
| Organisation | Engineering and Physical Sciences Research Council (EPSRC) |
| Sector | Public |
| Country | United Kingdom |
| Start | 08/2022 |
| End | 09/2026 |
| Description | Continuous Reinforcement Learning for Quadruped Robots |
| Amount | £99,775 (GBP) |
| Funding ID | 2728764 |
| Organisation | Engineering and Physical Sciences Research Council (EPSRC) |
| Sector | Public |
| Country | United Kingdom |
| Start | 08/2022 |
| End | 09/2026 |
| Description | Research on Embodied Perception and Intelligent Interaction Methods Driven by Multimodal Large Models |
| Amount | £203,079 (GBP) |
| Funding ID | 101211118 |
| Organisation | Marie Sklodowska-Curie Actions |
| Sector | Charity/Non Profit |
| Country | Global |
| Start | 08/2025 |
| End | 08/2027 |
| Description | Silicone 3D printer |
| Amount | £48,696 (GBP) |
| Organisation | University College London |
| Sector | Academic/University |
| Country | United Kingdom |
| Start | 03/2024 |
| Description | UCL Architecture Research Fund (ARF): Interactive Multi-Robot Assembly! |
| Amount | £9,989 (GBP) |
| Organisation | University College London |
| Sector | Academic/University |
| Country | United Kingdom |
| Start | 08/2022 |
| End | 06/2023 |
| Description | UCL Cities Partnerships Programme: Machine Learning for Robotic Grasping and Manipulation of Everyday Objects |
| Amount | £5,000 (GBP) |
| Organisation | University College London |
| Sector | Academic/University |
| Country | United Kingdom |
| Start | 01/2022 |
| End | 07/2022 |
| Description | UCL Grand Challenges Doctoral Students: Employing Quadrupedal Robots as an Innovative Solution for Last-Mile Logistics |
| Amount | £500 (GBP) |
| Organisation | University College London |
| Sector | Academic/University |
| Country | United Kingdom |
| Start | 01/2022 |
| End | 07/2022 |
| Description | UCL Innovation Enterprise: UCL Robotics Innovation Network |
| Amount | £29,978 (GBP) |
| Organisation | University College London |
| Sector | Academic/University |
| Country | United Kingdom |
| Start | 01/2022 |
| End | 03/2022 |
| Description | UK-RAS Pump Priming: Summer School on Robot Manipulation Learning |
| Amount | £3,500 (GBP) |
| Organisation | Engineering and Physical Sciences Research Council (EPSRC) |
| Sector | Public |
| Country | United Kingdom |
| Start | 05/2023 |
| End | 06/2023 |
| Title | Navigation Among Movable Obstacles via RL |
| Description | For first time we deal with the problem of robot navigation, where interaction with obstacle is needed. We do that via reinforcement learning -- a purely new way to deal with the problem in real time. |
| Type Of Material | Improvements to research infrastructure |
| Year Produced | 2022 |
| Provided To Others? | Yes |
| Impact | The research method is impressive and unique, as the problem is intractable and thus learning techniques had to be applied to drive a real-world solution for industrial applications, e.g., a factory that boxes need to move around to clear paths. |
| Title | New Fruit Manipulator |
| Description | We designed a new way to deal with fruit grasping, via a soft manipulators with flexible fingers and novel sensing devices. |
| Type Of Material | Improvements to research infrastructure |
| Year Produced | 2022 |
| Provided To Others? | Yes |
| Impact | This manipulator is unique as it uses a thumb-technique to encapsulate objects without damaging them, and utilizes new sensors to drive learned actions for grasping. This novel design will be submitted for a patent. |
| URL | https://ieeexplore.ieee.org/document/9635873 |
| Title | One-Shot Affordance Localization |
| Description | A novel way to find areas of interest on objects without the need to train any large neural network. |
| Type Of Material | Improvements to research infrastructure |
| Year Produced | 2022 |
| Provided To Others? | Yes |
| Impact | We managed to minimise the way of localising objects, such as grapes in a vineyard, or tools in a workshop, without the need to retrain a large network, thus, by minimising the amount of data and the energy for training a large network. |
| URL | https://sites.google.com/view/affcorrs |
| Title | UCL East Robotics Lab |
| Description | TBD |
| Type Of Material | Improvements to research infrastructure |
| Year Produced | 2023 |
| Provided To Others? | Yes |
| Impact | TBD |
| URL | https://www.ucl.ac.uk/ucl-east/study-and-research/robotics-and-autonomous-systems |
| Title | Multi-modal Data in Vineyards |
| Description | This dataset consists of multi-modal sensor data collected from three vineyards (KTIMA Gerovassiliou (Greece), Biblia Chora (Greece), and Denbies (UK)) to support research in robotic perception, precision agriculture, and AI-driven vineyard monitoring. Core Data Modalities: 1. RGB and Depth Images: Captured for grapevine detection, segmentation, and growth analysis. 2. LiDAR Point Clouds: Provides high-resolution 3D vineyard structure mapping for robot navigation and vine modeling. Key Features & Applications: ? Supports Research in Robotic Vineyard Monitoring & Navigation - Enables autonomous robots to navigate and inspect vineyards efficiently. ? Facilitates AI Training for Precision Agriculture - Provides high-quality training data for machine learning models in vine health assessment, disease detection, and yield prediction. ? Multi-Site Generalisation - Collected across three geographically distinct vineyards, enhancing model robustness and adaptability. ? Scalable for Agricultural Robotics & Environmental Monitoring - Useful for robot-assisted pruning, harvesting, and sustainable vineyard management. This dataset serves as a valuable resource for researchers in robotics, AI, and precision agriculture, enabling advancements in autonomous vineyard monitoring and intelligent decision-making for sustainable farming. |
| Type Of Material | Database/Collection of data |
| Year Produced | 2025 |
| Provided To Others? | Yes |
| Impact | ? Advancing Autonomous Vineyard Monitoring - The dataset enables AI-driven robotic perception and navigation in vineyards, improving automated monitoring, disease detection, and yield estimation. ? Facilitating AI Training for Precision Agriculture - Provides a rich multi-modal dataset that supports machine learning research in grape segmentation, vine health analysis, and environmental modelling, reducing the need for manual data collection and labelling. ? Enhancing Generalisation Across Different Vineyards - Data collected from three distinct vineyards (Gerovassiliou, Biblia Chora, Denbies) allows for cross-site validation, improving the adaptability of AI models in diverse agricultural environments. ? Enabling Real-World Deployment of Agricultural Robots - Assists in developing robotic systems for vineyard inspection, pruning, and harvesting, bringing automation to traditional farming and enhancing productivity. ? Strengthening Industry and Academic Collaborations - Facilitates interdisciplinary research between robotics, AI, and agriculture, fostering collaborations with vineyard owners, agronomists, and technology developers. This dataset is a critical step toward intelligent, data-driven agriculture, accelerating the adoption of AI and robotics in viticulture for more efficient and sustainable vineyard management. |
| Title | Navigation Among Movable Obstacles with Deep Reinforcement Learning |
| Description | It is for first time that this problem (Navigation Among Movable Obstacles) is studied from a Reinforcement Learning point of view. We produced an algorithm to solve this hard problem locally, using multiple agents in simulation and tried it on a real quadruped robot. |
| Type Of Material | Computer model/algorithm |
| Year Produced | 2023 |
| Provided To Others? | Yes |
| Impact | Impact: a completely new way of solving this problem will allow the community to solve this challenging task. We expect high impact in the research community. |
| Title | One-Shot Transfer of Affordances when Manipulating Objects with Robots |
| Description | We introduced a completely new method to finding regions on objects that have a particular functionality (e.g., cutting, holding, etc.). The novelty is that there is no need for large data to learn such functionalities for new objects appearing in a robot camera. |
| Type Of Material | Computer model/algorithm |
| Year Produced | 2022 |
| Provided To Others? | Yes |
| Impact | Impact: we expect a high impact in the research society, as we have been in contact to release the algorithm (something we did on GitHub). With this method, there is no need to label large data to allow robots handle everyday tools. |
| URL | https://github.com/RPL-CS-UCL/UCL-AffCorrs |
| Title | UMD_i Dataset |
| Description | We present UMDi- a one-shot correspondence variant of UMD, which is composed of a single instance of each object in the dataset with both RGB image and affordance ground truth annotation. The original annotations are kept to highlight the difficulties of semantic transfer, as no two human annotations are the same. The classes include common objects such as bowls, ladles, and knives, with manually labelled grasp, scoop, wrap-grasp, support, contain, cut, and pound affordances. |
| Type Of Material | Database/Collection of data |
| Year Produced | 2022 |
| Provided To Others? | Yes |
| Impact | This dataset is useful for research on self-supervised learning of manipulation tasks and handling of tools in robotics. |
| URL | https://github.com/RPL-CS-UCL/UCL-AffCorrs |
| Description | Collaboration on Human Robot Interaction with LAAS-CNRS |
| Organisation | University of Toulouse |
| Department | Laboratory for Analysis and Architecture of Systems |
| Country | France |
| Sector | Academic/University |
| PI Contribution | In robotics, human-robot interaction is an important topic with robots being among humans. We collaborate with LAAS-CNRS towards a new way to use transformers to predict human motions and contact forces when a human is physically interacting with a robot. Our contribution was mainly in PhD student supervision and methodology development. |
| Collaborator Contribution | The contribution of LAAS-CNRS was the generation of the new method of human-robot interaction based on our co-supervision and testing it on the real robotic hardware. |
| Impact | Outcomes: 1) A submitted publication to a robotics conference. Multi-Disciplinarity: No, mainly all participants were from Computer Science. |
| Start Year | 2023 |
| Description | Collaboration on Legged Robot Control with LAAS-CNRS |
| Organisation | University of Toulouse |
| Department | Laboratory for Analysis and Architecture of Systems |
| Country | France |
| Sector | Academic/University |
| PI Contribution | In robotics, legged robot control using active sensing is a difficult problem. We collaborate with LAAS-CNRS towards a new way to solve the problem of active sensing for data quality improvement in model learning. Our contribution was the supervision of the PhD student, co-developing the method, and testing it with our robots. |
| Collaborator Contribution | The contribution of LAAS-CNRS was the co-supervision of the PhD student to develop and test the method. |
| Impact | Outcomes: 1) A submitted publication to control conference. Multi-Disciplinarity: No, mainly all participants were from Computer Science. |
| Start Year | 2023 |
| Description | Collaboration on Legged Robot Locomotion with IIT |
| Organisation | Italian Institute of Technology (Istituto Italiano di Tecnologia IIT) |
| Country | Italy |
| Sector | Academic/University |
| PI Contribution | In robotics, control for locomotion is an important and difficult problem. We collaborate with the Italian Institute of Technology towards a new way to perform control (a topic that is not the primary focus of our group). Our contribution was mainly co-developing the method and co-supervising the PhD student that performed the task. |
| Collaborator Contribution | The contribution of University of the Italian Institute of Technology was the generation under our guidance of the new method to perform grasping and testing on the real robotic hardware. |
| Impact | Outcomes: 1) A submitted publication to a robotics conference. Multi-Disciplinarity: No, mainly all participants were from Computer Science. |
| Start Year | 2023 |
| Description | Collaboration on Robot Localisation and Mapping with CAM |
| Organisation | Shanghai Jiao Tong University |
| Country | China |
| Sector | Academic/University |
| PI Contribution | In robotics, localisation and mapping is an important and difficult problem. We collaborate with University of Cambridge and Shanghai Jiao Tong University towards a new way to perform the task. Our contribution was mainly capturing the dataset that was needed for training using our legged robots and sensors in the lab. |
| Collaborator Contribution | The contribution of University of Cambridge and Shanghai Jiao Tong University was the generation of the new method to perform localisation and mapping using our dataset. |
| Impact | Outcomes: 1) Software that make localisation and mapping on legged robots in dynamic environments. 2) A submitted publication to Computer Vision conference. Multi-Disciplinarity: No, mainly all participants were from Computer Science. |
| Start Year | 2024 |
| Description | Collaboration on Robot Localisation and Mapping with CAM |
| Organisation | University of Cambridge |
| Department | Cambridge Neuroscience |
| Country | United Kingdom |
| Sector | Academic/University |
| PI Contribution | In robotics, localisation and mapping is an important and difficult problem. We collaborate with University of Cambridge and Shanghai Jiao Tong University towards a new way to perform the task. Our contribution was mainly capturing the dataset that was needed for training using our legged robots and sensors in the lab. |
| Collaborator Contribution | The contribution of University of Cambridge and Shanghai Jiao Tong University was the generation of the new method to perform localisation and mapping using our dataset. |
| Impact | Outcomes: 1) Software that make localisation and mapping on legged robots in dynamic environments. 2) A submitted publication to Computer Vision conference. Multi-Disciplinarity: No, mainly all participants were from Computer Science. |
| Start Year | 2024 |
| Description | Collaboration on Robot Manipulation Control with Chalmers |
| Organisation | Chalmers University of Technology |
| Country | Sweden |
| Sector | Academic/University |
| PI Contribution | In robotics, manipulation is still an open problem. We collaborate with Chalmers University of Technology towards a new way to perform object grasping with robots using Implicit Reachability and Holistic Visual Servoing. Our contribution was mainly towards discussions about solving theoretically the problem. |
| Collaborator Contribution | The contribution of University of Chalmers University of Technology was the generation of the new method to perform grasping and testing on the real robotic hardware. |
| Impact | Outcomes: 1) A submitted publication to a robotics conference. Multi-Disciplinarity: No, mainly all participants were from Computer Science. |
| Start Year | 2024 |
| Description | Collaboration on Robot Mapping with HKUST |
| Organisation | The Hong Kong University of Science and Technology |
| Country | Hong Kong |
| Sector | Academic/University |
| PI Contribution | In robotics, mapping is an important and difficult problem. We collaborate with The Hong Kong University of Science and Technology towards a new way to perform the task. Our contribution was mainly developing the method and testing on mobile robots. |
| Collaborator Contribution | The contribution of The Hong Kong University of Science and Technology was the co-development of the method and supervision. |
| Impact | Outcomes: 1) Software for mapping on mobile robots in dynamic environments. 2) A submitted publication to automation jouranl. Multi-Disciplinarity: No, mainly all participants were from Computer Science. |
| Start Year | 2024 |
| Description | Collaboration on Robot Sensorimotor Learning with UPatras |
| Organisation | University of Patras |
| Country | Greece |
| Sector | Academic/University |
| PI Contribution | In robotics, any task of loco-manipulation using end-to-end learning on sensory input data is still an open problem. In particular, multi-stage tasks are a challenge for reinforcement learning methods, and require either specific task knowledge (e.g., task segmentation) or big amount of interaction times to be learned. In this collaboration we study Behaviour Policy Learning that effectively combines 1) only few solution sketches, that is demonstrations without the actions, but only the states, 2) model-based controllers, and 3) simulations to effectively solve multi-stage tasks without strong knowledge about the underlying task. We collaborate with University of Patras towards a new way to perform sensorimotor learning with stability guarantees. Our contribution is two-fold: working on the theoretical aspect with implementation on legged robots and manipulators, and tested on the real hardware. We have helped with collection of real-world data and applying the learned policies on real robots (solving the sim2real gap) in order to increase the usability and impact of the developed methods. |
| Collaborator Contribution | The contribution of University of Patras was the theoretical generation of the new method. |
| Impact | Outcomes: 1) Submitted publication to a robotics conferences/journals and workshops. Multi-Disciplinarity: No, mainly all participants were from Computer Science. |
| Start Year | 2023 |
| Description | Collaboration on Social Robotics with ICL |
| Organisation | Imperial College London |
| Country | United Kingdom |
| Sector | Academic/University |
| PI Contribution | In robotics, social robotics becomes an emerging topic with robots walking around humans. We collaborate with Imperial College London towards a new way to perform locomotion of our legged robots among humans by preserving the social distances, while compelting our tasks. Our contribution was the development of the theory and code for serving the problem. |
| Collaborator Contribution | The contribution of Imperial College London was the supervision of the method generation and help with the physical robot experiments. |
| Impact | Outcomes: 1) A submitted publication to a robotics conference. Multi-Disciplinarity: No, mainly all participants were from Computer Science. |
| Start Year | 2023 |
| Description | Collaboration on Theoretical Aspects of Reinforcement Learning with Amazon |
| Organisation | Amazon.com |
| Department | Amazon UK |
| Country | United Kingdom |
| Sector | Private |
| PI Contribution | In robotics, using reinforcement learning for completing complex tasks is a challenging problem. We collaborate with Amazon researchers towards understanding the problem and finding solutions. Our contribution was mainly on theoretically resolving the problem of offline reinforcement learning for solving various tasks, such as maze exploration. |
| Collaborator Contribution | The main contribution of Amazon was to discuss with us the current problems transformers have when using goal information and local features. |
| Impact | Outcomes: 1) A submitted publication to a Machine Learning conference. Multi-Disciplinarity: No, mainly all participants were from Computer Science background. |
| Start Year | 2024 |
| Description | Collaboration on Whole-Body Robot Teleoperation with INRIA |
| Organisation | Inria Nancy - Grand-Est research centre |
| Country | France |
| Sector | Public |
| PI Contribution | In robotics, manipulation is still an open problem. We collaborate with INRIA towards a new way to perform whole body control for teleopration. Our contribution was towards the full method development and testing on our robots. |
| Collaborator Contribution | The contribution of INRIA was the discussion of methods and code explanation for developing our own method. |
| Impact | Outcomes: 1) A submitted publication to a robotics conference. Multi-Disciplinarity: No, mainly all participants were from Computer Science. |
| Start Year | 2023 |
| Description | Industrial Inspection - Arrival LTD |
| Organisation | Arrival Limited |
| Country | United Kingdom |
| Sector | Private |
| PI Contribution | Our team focused on developing a robotic solution for on-the-spot inspection of electric vehicles, exploring how mobile robots can autonomously assess vehicle integrity and manufacturing quality. |
| Collaborator Contribution | Industry partners provided key insights into the requirements of Generation 3 vehicle manufacturing, specifically identifying inspection tasks that could be automated to improve efficiency and accuracy in production. |
| Impact | ? Use-Case Definition: Established a well-defined industrial application for robotic vehicle inspection, ensuring alignment with real-world manufacturing needs. ? Multi-Disciplinary Collaboration: Brought together expertise in robotics, automation, and the automotive industry, bridging research and industrial requirements. ? Economic Impact: Contributed to the exploration of cost-effective, autonomous inspection solutions, enhancing efficiency in vehicle manufacturing. This collaboration was governed by formal agreements, including material transfer and confidentiality agreements, ensuring secure knowledge exchange between academic and industry partners. |
| Start Year | 2022 |
| Description | Mine Inspection with Forestry England |
| Organisation | Forestry England |
| Country | United Kingdom |
| Sector | Public |
| PI Contribution | We started preparing methods for inspecting private coal mines in Whales. Such tasks make sure that the mines are well maintained without risking the life of engineers. |
| Collaborator Contribution | Forestry England is our main contact with private coal miners to allow us collect data and try our robots in their mines. |
| Impact | So far we have discussions on what inspection in mine means and arranged trials in the mine. |
| Start Year | 2023 |
| Description | Promoting Field and Legged Robotics in the Research Community |
| Organisation | Norwegian University of Science and Technology (NTNU) |
| Country | Norway |
| Sector | Academic/University |
| PI Contribution | This is an effort to promote the benefits of legged robots in the research community via various ways: 1) editorials in good journals, 2) white papers, 3) robot competitions |
| Collaborator Contribution | Helping in initiating the efforts, e.g., inviting us to editorials as Editors, or including us to the organisation team of legged robot competitions in good conferences (ICRA, IROS). |
| Impact | 1) Editorial Paper in Frontiers in Robotics and AI on Rising Stars in Field Robotics that promote the new generation of researchers in the field. 2) A robot competition organised yearly at IEEE ICRA conference.Quadruped Robot Challenges (QRC) ICRA 2024 Yokohama, Stage 1+ |
| Start Year | 2023 |
| Description | Promoting Field and Legged Robotics in the Research Community |
| Organisation | Sungkyunkwan University |
| Country | Korea, Republic of |
| Sector | Academic/University |
| PI Contribution | This is an effort to promote the benefits of legged robots in the research community via various ways: 1) editorials in good journals, 2) white papers, 3) robot competitions |
| Collaborator Contribution | Helping in initiating the efforts, e.g., inviting us to editorials as Editors, or including us to the organisation team of legged robot competitions in good conferences (ICRA, IROS). |
| Impact | 1) Editorial Paper in Frontiers in Robotics and AI on Rising Stars in Field Robotics that promote the new generation of researchers in the field. 2) A robot competition organised yearly at IEEE ICRA conference.Quadruped Robot Challenges (QRC) ICRA 2024 Yokohama, Stage 1+ |
| Start Year | 2023 |
| Description | Robot Training - ETH |
| Organisation | ETH Zurich |
| Country | Switzerland |
| Sector | Academic/University |
| PI Contribution | Our team engaged in research collaboration on legged robot locomotion, focusing on advancing control strategies, perception, and adaptation for real-world deployment. |
| Collaborator Contribution | Our partners at ETH Zurich provided hands-on training for our researchers, offering practical experience with real-world robotic systems, including the ANYmal quadrupedal robot, to support the fellowship's objectives. |
| Impact | ? Researcher & Student Training: Two-day training on the UCL ANYmal robot. One PhD student attended a summer school at ETH Zurich, gaining advanced knowledge in legged robot locomotion. |
| Start Year | 2022 |
| Description | Robot Training - ORI |
| Organisation | University of Oxford |
| Department | Oxford Robotics Institute (ORI) |
| Country | United Kingdom |
| Sector | Academic/University |
| PI Contribution | Our team has been actively engaged in research collaboration on legged robot locomotion, focusing on enhancing mobility, control, and perception for autonomous quadrupedal robots in real-world environments. |
| Collaborator Contribution | Our partners provided access to key robotic resources, including: 1) Legged robotic platforms for experimental validation. 2) Software for localisation and mapping, enabling improved navigation capabilities. |
| Impact | ? EPSRC Research Proposal - A proposal has been submitted to EPSRC, leveraging insights from this collaboration to expand research in legged robot navigation and control. |
| Start Year | 2022 |
| Description | Tele-operated Legged Robots - Univ. of Leeds |
| Organisation | University of Leeds |
| Country | United Kingdom |
| Sector | Academic/University |
| PI Contribution | We collaborated closely with Univ. of Leeds (Dr Chengxu Zhou) to generate legged robots that can be tele-operated in a whole-body fashion, and allow user to use VR headsets to feel the environment when performing loco-manipulation tasks. We tried our method for a Explosive Ordnance Disposal tasks, in collaboration with DASA/Dstl. |
| Collaborator Contribution | We developed the visual and virtual reality system to deal with teleoperation and telexistence when controlling a legged robot manipulator. |
| Impact | We manage to produce a method to control quadruped manipulators using tele-operation and tele-existence. We produced 4 papers (2 in 5th UK Robotics and Autonomous Systems Conference, 1 in IEEE Transactions on Human-Machine Systems and 1 in arXiv). |
| Start Year | 2022 |
| Description | Vineyard Harvesting/Pruning - Bilbia Chora |
| Organisation | Ktima Biblia Chora |
| Country | Greece |
| Sector | Private |
| PI Contribution | Our team developed a novel approach for grape recognition and identification that eliminates the need for large, manually labeled datasets, significantly reducing data annotation efforts. This method was successfully presented in our CoRL'22 publication. |
| Collaborator Contribution | Our partners provided valuable domain expertise and field support, including: 1) Hands-on training in vineyard processes. 2) Demonstration of harvesting and pruning techniques to enhance agricultural robotics research. 3) Collection of real-world vineyard data to support algorithm development and testing. |
| Impact | ? Technical Output: Developed and released code for one-shot grape recognition/localisation in RGB images, advancing agricultural automation. ? Academic Contribution: Published a peer-reviewed paper at CoRL'22, disseminating our findings to the broader research community. This collaboration is inherently multi-disciplinary, combining expertise in robot perception, computer vision, machine learning, and agricultural sciences to drive innovation in precision viticulture. |
| Start Year | 2022 |
| Description | Volcanic Inspection - INGV |
| Organisation | National Institute for Geophysics and Volcanology (INGV) |
| Country | Italy |
| Sector | Public |
| PI Contribution | Our team initiated discussions on robotic navigation methods for challenging volcanic terrains, exploring the sensor requirements and data collection strategies necessary for effective autonomous inspection of Mountain Vesuvius. |
| Collaborator Contribution | Collaborating volcanologists provided expert insights into the requirements for volcano inspection, detailing specific environmental challenges and potential strategies for robotic deployment. |
| Impact | ? Use-Case Definition: Established a framework for autonomous robotic inspection of volcanic environments, identifying key challenges and required sensing modalities. ? Multi-Disciplinary Collaboration: Combined expertise in robotics, environmental monitoring, and volcanology, bridging the gap between scientific fieldwork and autonomous exploration technologies. |
| Start Year | 2022 |
| Description | Weiwei Wan - Osaka Univ. |
| Organisation | Osaka University |
| Country | Japan |
| Sector | Academic/University |
| PI Contribution | Our team conducted research on tele-operation, focusing on enhancing remote control capabilities for robotic systems, particularly in complex manipulation tasks. |
| Collaborator Contribution | Our partners contributed through research on bi-manual manipulation, developing coordinated control strategies for two robotic arms to handle heavy object manipulation effectively. |
| Impact | ? Journal Publication - A research paper was published on bi-manual manipulation of heavy objects using two robots, contributing to advancements in robotic manipulation and automation. |
| Start Year | 2022 |
| Title | ASFM: Augmented Social Force Model for Legged Robot Social Navigation |
| Description | This software provides a socially-aware navigation framework for legged robots, enabling them to move efficiently and naturally in human-populated environments. It is based on an augmented social force model, which refines traditional pedestrian interaction models by incorporating visual perception for human localisation and adaptive avoidance behaviours. Core Functionality: 1) Social Force Model for Legged Robots: Implements a refined interpretation of repulsive forces and pedestrian avoidance strategies, ensuring safe and smooth robot movement in dynamic human environments. 2) Vision-Based Human Localisation: Uses visual perception techniques to detect and track pedestrians, adapting navigation based on human actions and motion patterns. 3) Target Following Mechanism: Integrates an intelligent path-following system, allowing the robot to efficiently reach goals while respecting pedestrian comfort. Key Features: ? Human-Aware Navigation - Balances pedestrian comfort with efficient path-finding, allowing seamless integration of robots into human environments. ? Enhanced Avoidance Behaviours - Dynamically adapts repulsive force interpretations based on real-world pedestrian interactions. ? Validated in Real-World Scenarios - Tested on a quadruped robot in varied settings, including oncoming pedestrian interactions, crowds, and obstructed paths. ? Performance Improvements Over Baselines - Demonstrates shorter path lengths, improved average velocity, and reduced time-to-goal, surpassing traditional social navigation approaches. This framework advances the state-of-the-art in social navigation, offering a robust, perception-driven solution for legged robots operating in human-centred spaces, with applications in service robotics, public environments, and assistive technologies. |
| Type Of Technology | Software |
| Year Produced | 2024 |
| Open Source License? | Yes |
| Impact | ? Enhanced Robot Integration in Human Environments - Enables legged robots to navigate seamlessly among pedestrians, improving their ability to function in public spaces, workplaces, and service environments. ? Improved Socially-Aware Navigation - Advances human-aware path planning by refining repulsive force interpretations and avoidance strategies, leading to more natural and comfortable interactions between robots and pedestrians. ? Increased Efficiency in Navigation - Demonstrates shorter path lengths, improved average velocity, and reduced time-to-goal, surpassing traditional social navigation baselines. ? Potential for Real-World Deployment - The system's robustness and adaptability make it suitable for autonomous service robots, security patrols, and last-mile delivery applications, where navigating human-populated areas is crucial. ? Advancement in Human-Robot Interaction Research - Provides a scalable framework for social navigation, contributing to further developments in robotics for smart cities, assistive technology, and public safety applications. By enabling robots to move efficiently and respectfully within human spaces, this software brings us closer to the seamless integration of legged robots in everyday life. |
| URL | https://rpl-cs-ucl.github.io/ASFM/ |
| Title | Common low level control framework for quadruped robots |
| Description | We developed a framework to control several different quadrupedal robots in the same way to make our algorithms robot-agnostic. |
| Type Of Technology | Software |
| Year Produced | 2022 |
| Impact | Impact: we aim at releasing the software soon in order to make an impact to the research society, by providing a framework that works independently of the robot model. |
| Title | Learning-based Grasping of Grocery Items using Caging-inspired Gripper Design |
| Description | This software provides a novel grasping approach for reliable and stable robotic grocery handling, addressing the challenges posed by diverse object shapes, delicate packaging, and dynamic deformable items. It integrates compliant caging gripper design with force feedback reinforcement learning, ensuring precise and adaptive grasp control. Core Functionality: 1. Compliant Caging Gripper Simulation: A validated physics-based model of the compliant caging gripper, enabling accurate simulation-to-real transfer. 2. Force Feedback Reinforcement Learning: A grasping controller trained using RL, optimizing grasp stability while keeping in-grasp forces minimal to prevent item damage. 3. Robust Grasp Evaluation Metrics: Measures grasp success rates, force stability, and disturbance resistance, ensuring reliable handling of a variety of grocery items. Key Features: ? High Grasp Success Rate - Achieved 95.4% grasp success on 42 grocery items, with grasps resisting up to 3.6N disturbances. ? Superior Performance Over Existing Methods - Exceeded state-of-the-art grocery grasping benchmarks, improving bin clearance rates by 18%. ? Validated Sim-to-Real Transfer - Demonstrated a 3.1% success rate gap between simulated and real-world execution, proving model reliability. ? Durability & Large-Scale Testing - The gripper design successfully performed over 5000 grasps across multiple experiments, ensuring robustness. This work represents a significant step forward in autonomous grocery grasping, providing a reliable, adaptive, and efficient solution for applications in e-commerce fulfillment, warehouse automation, and robotic grocery handling. |
| Type Of Technology | Software |
| Year Produced | 2024 |
| Open Source License? | Yes |
| Impact | ? Advancing Autonomous Grocery Handling - The development of this software and grasping approach significantly improves robotic handling of varied grocery items, enabling more reliable and damage-free grasping in e-commerce fulfillment and warehouse automation. ? State-of-the-Art Performance in Grocery Grasping - Demonstrated a 95.4% grasp success rate, outperforming existing methods with a median improvement of 11% and achieving a 93% bin clearance rate, an 18% improvement over prior benchmarks. ? Improved Stability and Delicate Object Handling - The force feedback reinforcement learning approach ensures that grasps remain stable while minimizing in-grasp forces, reducing the likelihood of damaging fragile grocery items. ? Scalable and Robust for Real-World Deployment - The compliant caging gripper design was tested extensively, completing over 5000 grasps in large-scale experiments, demonstrating durability and real-world feasibility. ? Validated Simulation-to-Real Transfer - The quantitatively validated gripper simulation model enables accurate transfer of learned grasping policies from simulation to real-world execution, reducing development costs and training time. ? Potential for Broader Applications - Beyond grocery handling, this approach has implications for automated packaging, warehouse logistics, and industrial robotic grasping, where adaptive, safe, and efficient manipulation is crucial. This work represents a significant advancement in robotic grasping, making grocery automation more reliable, scalable, and commercially viable. |
| URL | https://lukebeddow.github.io/ |
| Title | LiteVLoc: Map-Lite Visual Localization for Image Goal Navigation |
| Description | LiteVLoc is a hierarchical visual localization framework designed for efficient and scalable camera pose estimation in large-scale environments. Unlike traditional methods that rely on detailed 3D representations, LiteVLoc employs a lightweight topo-metric map, significantly reducing storage overhead and computational costs. Core Functionality: 1) Coarse-to-Fine Camera Pose Estimation: Uses a three-stage hierarchical approach to progressively refine localization accuracy. 2) Learning-Based Feature Matching: Reduces reliance on dense 3D models by leveraging deep-learning-driven feature extraction and geometric solvers. Map-Free Relocalization Dataset: Introduces a novel benchmark dataset specifically designed to evaluate localization performance without prior detailed maps. Key Features: ? Storage-Efficient Localization - Reduces the need for dense 3D maps, making it ideal for large-scale deployments. ? High-Precision Pose Estimation - Achieves accurate localization using lightweight representations and feature-based matching techniques. ? Real-World and Simulated Validation - Demonstrates robust navigation and localization performance across diverse environments. ? Scalable for Large-Scale Applications - Suitable for robotics, autonomous navigation, and AR/VR applications, where efficient relocalization is critical. LiteVLoc provides a highly efficient, scalable, and storage-conscious alternative to conventional localization techniques, making it well-suited for real-world mobile robotics and autonomous navigation. |
| Type Of Technology | Software |
| Year Produced | 2025 |
| Open Source License? | Yes |
| Impact | ? Improved Efficiency in Large-Scale Visual Localisation - By utilising a lightweight topo-metric map instead of dense 3D representations, LiteVLoc significantly reduces storage and computational requirements, making it scalable for large environments. ? Advancement in Map-Free Relocalisation - Introduces a novel benchmark dataset, enabling the research community to develop and evaluate localization systems without the need for pre-built 3D maps, fostering innovation in autonomous navigation and robotics. ? Enhanced Real-World Deployment - The framework's precision and efficiency have been validated in both simulated and real-world scenarios, demonstrating potential for use in mobile robotics, AR/VR, and autonomous vehicles. ? Increased Accessibility of Localisation Technologies - Reduces reliance on heavy computational resources, making advanced localisation more accessible for edge devices and resource-constrained robotic systems. ? Potential Applications Across Multiple Domains - LiteVLoc is applicable to robotic navigation, augmented reality (AR), virtual reality (VR), autonomous vehicles, and drone-based exploration, offering a versatile and efficient localisation solution. By improving accuracy, efficiency, and scalability, LiteVLoc represents a significant step forward in visual localisation research, supporting a wide range of real-world applications where fast and lightweight relocalization is essential. |
| URL | https://rpl-cs-ucl.github.io/LiteVLoc/ |
| Title | Navigation Among Movable Obstacles with Deep Reinforcement Learning |
| Description | In this code we allow RL agents to be trained to solve the challenging problem of Navigation among Movable Obstacles, using Reinforcement Learning. |
| Type Of Technology | Software |
| Year Produced | 2023 |
| Open Source License? | Yes |
| Impact | Impact: a novel method to solve a problem that is open for several years in the research community. |
| Title | Python-based framework for rapid prototyping of RL and optimal controller for quadruped robots |
| Description | We work on a CasADi-based python framework to allow fast and less complex code development for quadruped robots either they are trained via Reinforcement Learning or Optimal Control. |
| Type Of Technology | Software |
| Year Produced | 2023 |
| Impact | Impact: we aim for open-sourcing the code in order to allow the research community develop faster methods to control robots, without the need to hard code their algorithms. |
| Title | SDS - See it, Do it, Sorted: Quadruped Skill Synthesis from Single Video Demonstration |
| Description | SDS is a novel pipeline for intuitive quadrupedal skill learning from a single demonstration video, eliminating the need for manual reward engineering and large-scale training datasets. Core Functionality: Video-to-Reward Learning: SDS leverages GPT-4o's vision capabilities to process input videos and extract key motion patterns using a chain-of-thought prompting technique (SUS). Automated Reward Function (RF) Generation & Evolution: SDS autonomously constructs executable reward functions, refining them iteratively using fitness metrics and training footage. Reinforcement Learning Integration: The PPO-based RL policy learns locomotion skills in the NVIDIA IsaacGym simulator, adapting motions for real-world deployment. Key Features: ? Intuitive Imitation Learning - Robots can learn complex locomotion skills (trotting, bounding, pacing, hopping, etc.) from a single video demonstration. ? Autonomous Reward Evolution - GPT-4o continuously optimizes reward functions, reducing dependence on handcrafted domain knowledge. ? Seamless Sim-to-Real Transfer - Successfully validated on the Unitree Go1 quadrupedal robot, achieving high imitation fidelity and locomotion stability. ? Improved Task Adaptability - Outperforms state-of-the-art (SOTA) reinforcement learning methods, enabling fast skill acquisition with minimal human intervention. SDS represents a major advancement in quadrupedal robot learning, making skill acquisition faster, more efficient, and highly adaptable for real-world applications. |
| Type Of Technology | Software |
| Year Produced | 2024 |
| Open Source License? | Yes |
| Impact | ? Advancing Intuitive Skill Learning in Quadrupedal Robotics - SDS enables robots to learn complex locomotion skills from a single video demonstration, significantly lowering the barrier for skill acquisition compared to traditional reinforcement learning methods. ? Reduction in Reward Engineering Effort - By automating the generation and evolution of reward functions (RFs), SDS reduces the need for manual tuning and domain-specific knowledge, making robotic skill learning more efficient and accessible. ? Improved Task Adaptability & Imitation Fidelity - SDS achieves higher locomotion stability and imitation fidelity across a range of gaits, including trotting, bounding, pacing, and hopping, outperforming state-of-the-art (SOTA) approaches in skill generalisation. ? Seamless Integration of GPT-4o for Continuous Reward Optimization - The method introduces an autonomous reward evolution mechanism, where GPT-4o iteratively refines reward functions based on training footage and performance metrics, leading to progressively improved robotic behavior. ? Bridging the Gap Between Simulation and Reality - SDS demonstrates effective sim-to-real transfer on the Unitree Go1 robot, showcasing its real-world applicability without requiring extensive real-world training data. ? Potential Applications in Diverse Domains - SDS has broad implications for legged locomotion research, assistive robotics, industrial automation, and autonomous exploration, enabling robots to quickly adapt to new tasks and environments with minimal human intervention. This work marks a significant step towards intuitive, efficient, and scalable robot learning, pushing the boundaries of video-based skill acquisition and reinforcement learning for autonomous quadrupeds. |
| URL | https://rpl-cs-ucl.github.io/SDSweb/ |
| Title | UCL-AffCorrs |
| Description | Given a single reference image of an object with annotated affordance regions, can we segment semantically corresponding parts within a target scene? We investigate this question with AffCorrs, an unsupervised model that combines the useful properties of pre-trained DINO-ViT's image descriptors and cyclic correspondences. AffCorrs is able to find corresponding affordances both for intra- and inter-class one-shot semantic part segmentation. |
| Type Of Technology | Software |
| Year Produced | 2022 |
| Open Source License? | Yes |
| Impact | Finding corresponding affordances both for intra- and inter-class one-shot semantic part segmentation. is more difficult than supervised alternatives, but enables future work such as learning affordances via imitation and assisted teleoperation. |
| URL | https://sites.google.com/view/affcorrs |
| Title | Watch your STEPP: Semantic Traversability Estimation using Pose Projected Features |
| Description | This software provides a novel learning-based approach to terrain traversability estimation, enabling autonomous legged robots to navigate complex, unstructured environments by learning from human walking demonstrations. Core Functionality: 1) Feature Extraction with Vision Transformers: Utilises DINOv2, a vision Transformer model, to generate dense, pixel-wise terrain embeddings for robust scene understanding. 2) Encoder-Decoder MLP for Terrain Analysis: Processes terrain segments using a neural network-based reconstruction framework, identifying familiar and unfamiliar terrain based on reconstruction error. 3) Anomaly Detection for Safe Navigation: Differentiates between navigable terrain and hazardous regions, helping the robot adapt locomotion strategies accordingly. Key Features: ? Human Demonstration-Based Learning - The system learns terrain traversability from human walking behaviour, reducing reliance on manually labeled datasets. ? Vision-Based Terrain Analysis - The pixel-wise embeddings enable a fine-grained understanding of terrain types, improving perception accuracy. ? Adaptive and Scalable - Can generalize across indoor and outdoor environments, supporting navigation in forests, rocky landscapes, and urban terrains. ? Validated on ANYmal Legged Robot - Successfully tested on a real quadrupedal robot, demonstrating effective navigation in real-world scenarios. This approach offers a significant advancement over traditional occupancy mapping, allowing legged robots to perceive and respond to terrain variations more intelligently, enhancing their autonomous mobility in unstructured environments. |
| Type Of Technology | Software |
| Year Produced | 2024 |
| Open Source License? | Yes |
| Impact | ? Enhanced Autonomous Navigation in Complex Terrains - Enables legged robots to intelligently assess and adapt to unstructured environments, improving mobility in forests, rocky landscapes, and urban settings. ? Reduced Reliance on Manually Labeled Data - By learning from human walking demonstrations, the system eliminates the need for extensive, labor-intensive terrain labeling, making traversability estimation more scalable and efficient. ? Improved Safety & Robustness for Legged Robots - The anomaly detection mechanism helps robots identify and avoid hazardous terrain, leading to safer and more stable locomotion in dynamic environments. ? Advancement in AI-Driven Perception - Demonstrates how self-supervised learning with vision transformers (DINOv2) can be leveraged for fine-grained terrain analysis, setting a foundation for future applications in autonomous exploration. ? Real-World Validation & Deployment - Successfully tested on the ANYmal quadrupedal robot, proving its effectiveness in both indoor and outdoor scenarios, with potential applications in disaster response, environmental monitoring, and autonomous inspection. This development represents a significant leap in terrain-aware navigation, making legged robots more adaptive, efficient, and capable of operating in real-world, unstructured terrains. |
| URL | https://rpl-cs-ucl.github.io/STEPP/ |
| Description | 4x Invited talks General Mobile Robot Navigation System for Unstructured Environments |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Postgraduate students |
| Results and Impact | Delivered a series of invited talks in June 2024 at Shenzhen University, Beijing Institute of Technology, Shanghai University, and Hunan University, hosted by Professors Baoding Zhou, Shilei Li, Jianlin Chen, and Yi Zhou. Presented advancements in autonomous mobile robot navigation, focusing on AI-driven perception, adaptive locomotion, and decision-making in unstructured environments. Engaged with faculty, researchers, and students to discuss challenges and emerging solutions in mobile robotics, particularly in natural terrain navigation, disaster response, and industrial automation. Aimed to strengthen research collaborations between China and the UK, fostering future joint projects, student exchanges, and interdisciplinary cooperation in robotics and AI. Outcomes & Impacts: ? Increased Interest in AI-Driven Mobile Robotics - Sparked discussions on learning-based navigation, multimodal perception, and real-time planning, leading to new research ideas for students and faculty. ? Strengthening International Research Ties - Opened collaborative opportunities for joint publications, PhD exchanges, and funding proposals between Chinese and UK institutions. ? Engaging the Next Generation of Roboticists - Inspired students and early-career researchers to explore robotics and AI, enhancing participation in research on autonomous navigation. ? Industry and Real-World Applications - Highlighted how mobile robots can be applied in logistics, agriculture, and search-and-rescue missions, fostering interest from academic and industrial stakeholders. These talks contributed to bridging research communities between China and the UK, promoting innovation in mobile robot autonomy and intelligent navigation systems. |
| Year(s) Of Engagement Activity | 2024 |
| Description | AI Showcase |
| Form Of Engagement Activity | Participation in an open day or visit at my research institution |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Postgraduate students |
| Results and Impact | AI Showcase with our quadruped robots performing for PhD students, the Deans and Provost of UCL. |
| Year(s) Of Engagement Activity | 2022 |
| Description | BBC Digital Planet podcast |
| Form Of Engagement Activity | A broadcast e.g. TV/radio/film/podcast (other than news/press) |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Public/other audiences |
| Results and Impact | In brief: BBC interview about robotic dogs and their ethical aspects. Impact: public engagement about legged robots and how useful they can be in our society, as well as discussed the problem of weaponising robots. |
| Year(s) Of Engagement Activity | 2022 |
| URL | https://www.bbc.co.uk/sounds/play/w3ct31yt |
| Description | BMVA'24 Workshop on "Robotics Foundation World Models" |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Postgraduate students |
| Results and Impact | Co-organized and participated in the BMVA 2024 Workshop on "Robotics Foundation World Models", focusing on the role of foundation models in robotic perception, decision-making, and control. Engaged with experts in robotics, AI, and vision-based learning, discussing how world models can enable robots to generalize across tasks and environments. Explored advances in deep learning, reinforcement learning, and self-supervised learning for improving robotic autonomy. Outcomes & Impacts: ? Advancing AI-Driven Robotics - Highlighted the role of world models in autonomous robot learning, shaping future research directions in robotic intelligence and adaptation. ? Strengthening Industry & Academic Collaborations - Fostered discussions between academic researchers and industry professionals, promoting potential joint research efforts. ? Engaging Early-Career Researchers - The workshop attracted PhD students and postdocs, encouraging new perspectives on integrating foundation models into robotics. ? Defining Future Challenges - Identified key research questions and challenges in robot learning, data efficiency, and cross-domain generalization, paving the way for further innovation in autonomous systems. This workshop contributed to bridging AI and robotics research, reinforcing the importance of scalable world models for future intelligent robotic systems. |
| Year(s) Of Engagement Activity | 2024 |
| URL | https://www.bmva.org/meetings/24-10-30-RobotWorldModels.html |
| Description | Conference paper presentation at CoRL 2022 |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Postgraduate students |
| Results and Impact | In brief: presentation of paper on "One-Shot Transfer of Affordance Regions? AffCorrs!", at the 6th Annual Conference on Robot Learning. Impact: outreaching to academics working in the same research area. |
| Year(s) Of Engagement Activity | 2022 |
| URL | https://corl2022.org/ |
| Description | Conference paper presentation at IROS 2022 |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Postgraduate students |
| Results and Impact | In brief: presentation of paper on "Navigation Among Movable Obstacles with Object Localization using Photorealistic Simulation", at the 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems. Impact: outreaching to academics working in the same research area. |
| Year(s) Of Engagement Activity | 2022 |
| URL | https://iros2022.org/ |
| Description | Conference paper presentation at IROS 2022 |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Postgraduate students |
| Results and Impact | In brief: presentation of paper on "Autonomous Mobile 3D Printing of Large-Scale Trajectories", at the 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems. Impact: outreaching to academics working in the same research area. |
| Year(s) Of Engagement Activity | 2022 |
| URL | https://iros2022.org/ |
| Description | Conference paper presentation at IROS 2022 |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Postgraduate students |
| Results and Impact | In brief: presentation of paper on "Robust Contact State Estimation in Humanoid Walking Gaits", at the 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems. Impact: outreaching to academics working in the same research area. |
| Year(s) Of Engagement Activity | 2022 |
| URL | https://iros2022.org/ |
| Description | Conference paper presentation at UK-RAS 2022 |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Postgraduate students |
| Results and Impact | In brief: presentation of paper on "Analysis of VR Usability in Mobile Manipulator Teleoperation", at the 5th UK Robotics and Autonomous Systems Conference. Impact: outreaching to academics working in the same research area in the UK. |
| Year(s) Of Engagement Activity | 2022 |
| URL | https://www.ukras.org.uk/news-and-events/uk-ras/ukras22-the-5th-uk-robotics-and-autonomous-systems-c... |
| Description | Conference paper presentation at UK-RAS 2022 |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Postgraduate students |
| Results and Impact | In brief: presentation of paper on "Dynamic Camera Usage in Mobile Teleoperation System for Buzz Wire Task", at the 5th UK Robotics and Autonomous Systems Conference. Impact: outreaching to academics working in the same research area in the UK. |
| Year(s) Of Engagement Activity | 2022 |
| URL | https://www.ukras.org.uk/news-and-events/uk-ras/ukras22-the-5th-uk-robotics-and-autonomous-systems-c... |
| Description | Demos/Posters at the 2023 UCL Robotics Workshop |
| Form Of Engagement Activity | Participation in an open day or visit at my research institution |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Industry/Business |
| Results and Impact | In brief: We ran demos with our robots to promote our research, we presented student posters of our work, and gave an invited talk about the fellowship to more than 200 participants, that included industry (e.g., BP), research centres (e.g. DeepMind),academics from all over the country, and media/policymakers. Impact: Changed the view of robotics towards a safer society. |
| Year(s) Of Engagement Activity | 2023 |
| URL | https://twitter.com/matchymaucl/status/1627701885044985859 |
| Description | Disruptive Thinker Video |
| Form Of Engagement Activity | A broadcast e.g. TV/radio/film/podcast (other than news/press) |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Other audiences |
| Results and Impact | In Brief: How can we bring limbed robots into real world environments to complete challenging tasks? Dr Dimitrios Kanoulas and the team at UCL Computer Science's Robot Perception and Learning Lab are exploring how we can use autonomous and semi autonomous robots to work in environments that humans cannot. Impact: public engagement, academic outreaching |
| Year(s) Of Engagement Activity | 2023 |
| URL | https://www.ucl.ac.uk/engineering/research/disruptive-thinkers-video-series#How%20can%20we%20bring%2... |
| Description | Evening Standard Tech and Science Daily by Mark Blunden about RoboHike |
| Form Of Engagement Activity | A broadcast e.g. TV/radio/film/podcast (other than news/press) |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Public/other audiences |
| Results and Impact | In brief: a podcast about the need of robots in our society Impact: discussed the economical impact of robots helping in tasks, such as agriculture. |
| Year(s) Of Engagement Activity | 2022 |
| URL | https://open.spotify.com/episode/6AUzNIgvT0EExs2Vt2zLyu?si=9zzRkZzxQdW4FLYX_X8MKw&dl_branch=1&nd=1 |
| Description | Film about Disruptive Thinkers |
| Form Of Engagement Activity | A press release, press conference or response to a media enquiry/interview |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Public/other audiences |
| Results and Impact | An interview about how can we bring robots into challenging environments, demonstrating how legged robots can help in several dangerous jobs. |
| Year(s) Of Engagement Activity | 2023 |
| URL | https://www.ucl.ac.uk/engineering/research/disruptive-thinkers-video-series#How%20can%20we%20bring%2... |
| Description | ICRA 2022 Workshop: "Towards Real-World Deployment of Legged Robots" |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Industry/Business |
| Results and Impact | In brief: organised the sixth version of "Towards Real-World Deployment of Legged Robots" workshop at ICRA 2022 (the larger robotics conference), where we have several speakers from academia and industry discussing legged robots, with invited presentations, demos, and posters. Impact: this is the largest legged robot event of the year, that brings together all roboticists that work on legged robots |
| Year(s) Of Engagement Activity | 2022 |
| URL | https://www.youtube.com/playlist?list=PLY45TGWcpM7yohYe33biZd1cEzwLK78dh |
| Description | ICRA 2023 Workshop on "Towards Real-World Deployment of Legged Robots" |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Postgraduate students |
| Results and Impact | In brief: We organised this workshop in the top conference in robotics to engage with researchers and industry in legged robotics. This is the best known workshop in legged robots worldwide. Impact: Brought together all roboticists working on legged robots. |
| Year(s) Of Engagement Activity | 2023 |
| URL | https://leggedrobots.org/ |
| Description | ICRA 2023-2024 Competition: "Autonomous Quadruped Robot Challenge (AQRC)" |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Postgraduate students |
| Results and Impact | The IEEE RAS Quadruped Robot Challenges (QRC) are robotic competitions mainly supported by IEEE RAS. The ICRA 2023 QRC is the first event in a series of challenges that will evolve from remote control of quadruped robots at a first-person-view station, to autonomous traversing on challenging terrain, and ultimately to racing of multiple quadruped manipulator robots. The QRC has great potential to lead the robotics community in technology advancement, nurture field engineers, and foster interactions with industries, ultimately leading to the creation of practical services for the public. |
| Year(s) Of Engagement Activity | 2023,2024 |
| URL | https://quadruped-robot-challenges.notion.site/Quadruped-Robot-Challenges-bdc4af35638c4036817c3212e6... |
| Description | ICRA'24 Competition on "Autonomous Quadruped Robot Challenge (AQRC)" |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Postgraduate students |
| Results and Impact | Co-Organized the ICRA 2024 Autonomous Quadruped Robot Challenge (AQRC), an international competition focused on advancing the autonomy, perception, and control of quadrupedal robots. Collaborated with leading experts in robotics and AI, contributing to the design, testing, and evaluation of autonomous quadruped systems in challenging environments. Aimed to push the boundaries of quadrupedal mobility, adaptive navigation, and real-world deployment, exploring applications in search-and-rescue, disaster response, and industrial automation. Outcomes & Impacts: ? Advancing Quadruped Autonomy - The competition provided a benchmark for evaluating state-of-the-art algorithms in robot perception, locomotion, and AI-driven decision-making. ? Encouraging Innovation & Collaboration - Fostered global collaboration among researchers from academia, industry, and robotics labs, leading to potential joint research efforts and technology advancements. ? Inspiring the Next Generation of Roboticists - Engaged PhD students, postdocs, and early-career researchers in hands-on AI and robotics development, strengthening interest in legged autonomy. ? Real-World Impact - The advancements demonstrated in AQRC contribute to autonomous robotics applications, particularly in hazardous environments, autonomous exploration, and robotic assistance. This competition played a key role in shaping the future of quadrupedal robotics, driving breakthroughs in AI-driven locomotion and real-world deployment of legged robots. |
| Year(s) Of Engagement Activity | 2024 |
| URL | https://ieee-qrc.agilertc.com/ |
| Description | IEEE Spectrum - Video Friday |
| Form Of Engagement Activity | A press release, press conference or response to a media enquiry/interview |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Postgraduate students |
| Results and Impact | In brief: the disruptive thinking video appeared to IEEE Spectrum and outreached to several colleagues worldwide. Impact: few colleagues showed interest in multi-disciplinary worth with us towards novel research and impact outcomes, e.g., scanning tunnels. |
| Year(s) Of Engagement Activity | 2023 |
| URL | https://spectrum.ieee.org/video-friday-drones-in-trees |
| Description | IEEE Spectrum Video Friday |
| Form Of Engagement Activity | A press release, press conference or response to a media enquiry/interview |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Postgraduate students |
| Results and Impact | In brief: appeared in the IEEE Spectrum (Video Friday), with our methods and outreaching activities. In this example making art using a quadruped robot. Impact: outreaching to robotics around the world with our research and outreach activities |
| Year(s) Of Engagement Activity | 2022,2023 |
| URL | https://spectrum.ieee.org/video-friday-robot-soccer-finals |
| Description | IROS'22 Workshop: "The Role of Uncertainty and How it is Tackled in Robotic Grasping and Manipulation" |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Postgraduate students |
| Results and Impact | We organised this workshop in one of the top conferences in robotics to engage with researchers and industry in robotic manipulation and promote state of the art research in the area. |
| Year(s) Of Engagement Activity | 2022 |
| URL | https://sites.google.com/view/the-role-of-uncertainty |
| Description | IROS'24 Competition on "Autonomous Quadruped Robot Challenge (AQRC)" |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Postgraduate students |
| Results and Impact | Co-organized the IROS 2024 Autonomous Quadruped Robot Challenge (AQRC), an international competition focused on advancing the autonomy, perception, and control of quadrupedal robots. Collaborated with leading experts in robotics and AI, contributing to the design, testing, and evaluation of autonomous quadruped systems in challenging environments. Aimed to push the boundaries of quadrupedal mobility, adaptive navigation, and real-world deployment, exploring applications in search-and-rescue, disaster response, and industrial automation. Outcomes & Impacts: ? Advancing Quadruped Autonomy - The competition provided a benchmark for evaluating state-of-the-art algorithms in robot perception, locomotion, and AI-driven decision-making. ? Encouraging Innovation & Collaboration - Fostered global collaboration among researchers from academia, industry, and robotics labs, leading to potential joint research efforts and technology advancements. ? Inspiring the Next Generation of Roboticists - Engaged PhD students, postdocs, and early-career researchers in hands-on AI and robotics development, strengthening interest in legged autonomy. ? Real-World Impact - The advancements demonstrated in AQRC contribute to autonomous robotics applications, particularly in hazardous environments, autonomous exploration, and robotic assistance. This competition played a key role in shaping the future of quadrupedal robotics, driving breakthroughs in AI-driven locomotion and real-world deployment of legged robots. |
| Year(s) Of Engagement Activity | 2025 |
| URL | https://quadruped-robot-challenges.notion.site/IROS-2024-Quadruped-Robot-Challenges-d3a7de99caf842b7... |
| Description | Innovation Day event at the Hellenic Centre |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | Regional |
| Primary Audience | Industry/Business |
| Results and Impact | The Greek Embassy put together an educational day dedicated to innovations or apps that would be of interest to the wider community. We were invited to demonstrate our findings. |
| Year(s) Of Engagement Activity | 2024 |
| Description | Interview (broadcast), Ant1, Greece |
| Form Of Engagement Activity | A press release, press conference or response to a media enquiry/interview |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Public/other audiences |
| Results and Impact | This broadcast interview on ANT1 Greece highlighted the work of the Robotic Perception and Learning (RPL) team at UCL, showcasing advancements in legged robotics and AI-driven autonomy. The interview focused on Spot, the quadrupedal robot from Boston Dynamics, and its potential applications in industry, agriculture, and disaster response. Intended Purpose: 1. To educate the public about advancements in robotics and AI and their real-world applications. 2. To highlight Greek contributions to global robotics research, particularly the work of Professor Dimitrios Kanoulas and the UCL RPL team. 3. To engage with industry and policymakers on how robotic technology can enhance productivity and safety in various sectors. Outcomes & Impacts: ? Increased Public Awareness - The interview helped demystify robotics and highlight its positive societal impact, reaching a broad Greek-speaking audience. ? Potential for Industry Collaboration - Showcased the robot's real-world capabilities, fostering interest from industry, agriculture, and emergency response sectors. ? Strengthening Greek Tech Networks - The event at the Hellenic Center in London, where Spot was presented, promoted collaboration between Greek researchers and technology startups. ? Inspiring Future Robotics Talent - The media exposure increased interest in robotics and AI among students and researchers, encouraging further engagement in STEM fields. This broadcast and event reinforced the role of robotics in solving real-world challenges, supporting further collaborations between academia, industry, and Greek tech networks. |
| Year(s) Of Engagement Activity | 2024 |
| URL | https://www.antenna.gr/tech/article/4/909947/skylos-rompot-o-ellinas-poy-programmatise-ton-spot-mila... |
| Description | Interview (broadcast), Hellenic Tech Network, UK |
| Form Of Engagement Activity | A broadcast e.g. TV/radio/film/podcast (other than news/press) |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Public/other audiences |
| Results and Impact | Featured in a broadcast interview with the Hellenic Tech Network, UK, discussing robotics, AI, and the future of autonomous systems. Highlighted ongoing research in robotic perception, learning, and mobility, particularly in legged robotics and AI-driven navigation. Aimed to engage the Greek tech and startup community, promoting collaboration between academia, industry, and entrepreneurs. Outcomes & Impacts: ? Increased Public Awareness of AI & Robotics - Helped inform a wider audience about cutting-edge advancements in robotics and their real-world applications. ? Strengthening Greek Tech Ecosystem - Encouraged dialogue between Greek researchers, industry leaders, and startups, fostering potential collaborations and innovation partnerships. ? Inspiring Future Innovators - Sparked interest among students, engineers, and entrepreneurs, encouraging engagement in robotics, AI, and deep tech ventures. ? Bridging Research & Industry - Positioned robotic perception and learning technologies as key drivers for innovation in sectors such as automation, agriculture, and disaster response. This interview contributed to enhancing the visibility of AI and robotics research, reinforcing the role of technology in solving real-world challenges and strengthening Greek-UK tech collaborations. |
| Year(s) Of Engagement Activity | 2024 |
| URL | https://youtu.be/WXzoafdc7TE?si=a6MEowxoIdrJmj6q&t=684 |
| Description | Interview 2023 for BBC |
| Form Of Engagement Activity | A broadcast e.g. TV/radio/film/podcast (other than news/press) |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Public/other audiences |
| Results and Impact | A broadcast about UCL East opening with respect to legged robots and automation, influencing the general public about good AI. |
| Year(s) Of Engagement Activity | 2023 |
| URL | https://www.bbc.co.uk/iplayer/episode/m001qpxz/bbc-london-evening-news-18092023 |
| Description | Interview 2024 at Institution of Mechanical Engineers |
| Form Of Engagement Activity | A press release, press conference or response to a media enquiry/interview |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Media (as a channel to the public) |
| Results and Impact | Interview for the Institution of Mechanical Engineers on embodied AI via humanoid robots, focusing on challenges and future directions. |
| Year(s) Of Engagement Activity | 2024 |
| URL | https://www.imeche.org/news/professional-engineering-digital-exclusives/stepping-out-of-the-shadows |
| Description | Interview for CDT of Foundational AI |
| Form Of Engagement Activity | A magazine, newsletter or online publication |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Postgraduate students |
| Results and Impact | 60sec with a Supervisor about several topics regarding AI and robotics, to influence the PhD students and their view about the future of embodied AI. |
| Year(s) Of Engagement Activity | 2023 |
| URL | https://www.linkedin.com/feed/update/urn:li:activity:7100479693501411328 |
| Description | Interview, Institution of Mechanical Engineers, UK |
| Form Of Engagement Activity | A press release, press conference or response to a media enquiry/interview |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Public/other audiences |
| Results and Impact | Participated in an interview for the Institution of Mechanical Engineers (IMechE) in October 2024, discussing advancements in fluid circuit-based control systems for soft robotics. Explored how intelligent fluid circuits could address key challenges in soft robotics, particularly in actuation, compliance, and control. Highlighted the potential applications of these systems in biomedical devices, adaptive robotic grippers, and autonomous soft robots operating in unstructured environments. Outcomes & Impacts: ? Public Engagement in Robotics & Engineering - The interview helped raise awareness about emerging technologies in soft robotics, making advanced engineering concepts accessible to a broader audience. ? Bridging Mechanical Engineering and AI - Showcased how fluid circuit control mechanisms can enhance robot adaptability, contributing to interdisciplinary robotics, AI, and mechanical design research. ? Encouraging Industrial and Academic Interest - Sparked discussions on the feasibility of soft robotic systems in industrial automation, medical applications, and space exploration, potentially influencing future research directions. ? Inspiring Next-Generation Engineers - The interview reached mechanical engineers, roboticists, and students, encouraging further exploration of soft robotics technologies and their applications. This interview contributed to increasing visibility of fluid circuit-driven soft robotics, promoting innovation at the intersection of mechanical engineering, AI, and robotics. |
| Year(s) Of Engagement Activity | 2024 |
| URL | https://www.imeche.org/news/news-article/very-clever-fluid-circuits-could-solve-soft-robot-challenge... |
| Description | Invited Keynote Talk, QMUL, 60 years of robotics at QMUL, UK |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Postgraduate students |
| Results and Impact | Delivered an invited keynote talk at Queen Mary University of London (QMUL) for the 60 Years of Robotics celebration in December 2024. Presented on "Legged Robots - What Happened in the Past 60 Years," tracing the evolution of legged robotic systems from early mechanical walkers to modern AI-driven quadrupeds. Highlighted key advancements in robot perception, control strategies, and learning-based locomotion, with a focus on how AI and machine learning have transformed legged robotics. Engaged with students, researchers, and industry professionals, discussing future challenges and opportunities in the field. Outcomes & Impacts: ? Educational Impact - Provided a historical and technical perspective on six decades of legged robot development, sparking engagement and discussion among attendees. ? Inspiring Future Robotics Research - Encouraged students and early-career researchers to pursue advancements in autonomous legged mobility, influencing research directions in robotics and AI. ? Strengthening Academic & Industry Collaboration - Connected with robotics experts, alumni, and industrial partners, fostering potential future collaborations in robotics and autonomous systems. ? Shaping the Next Decade of Legged Robotics - Provided insights into emerging technologies, AI-driven control, and bio-inspired designs, offering a vision for the future of legged robotics. This keynote contributed to celebrating the legacy of robotics at QMUL and in London in general, while reinforcing the importance of interdisciplinary research and innovation in legged robotic systems. |
| Year(s) Of Engagement Activity | 2024 |
| URL | https://www.qmul.ac.uk/events/upcoming-events/items/60-years-of-robotics-at-qmul.html |
| Description | Invited Lecture in Open Robotics Summer School |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Schools |
| Results and Impact | In brief: the invited talk at University of Patras in Greece, focused on "Humanoids, Quadrupeds, and other Legged Robots", to introduce school students to legged robotic research and development. Impact: Initiated a set of round table discussion with pupils, that showed interest in working with legged robots (have asked their instructor to consider a legged robot for next year's school). |
| Year(s) Of Engagement Activity | 2022 |
| URL | https://www.youtube.com/live/61eg3islhLU?feature=share |
| Description | Invited Panelist, Hellenic Tech Network, UK |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | Regional |
| Primary Audience | Industry/Business |
| Results and Impact | Participated as an invited panelist in the Hellenic Tech Network (HTC) event in February 2025, discussing "Robotics and AI: Shaping Our Daily Lives?" Explored the impact of robotics and AI on society, highlighting advancements in autonomous systems, smart automation, and human-robot interaction. Engaged with industry leaders, researchers, and policymakers to discuss the opportunities and challenges of AI integration in daily life, including ethics, trust, and automation in the workforce. Aimed to raise awareness of AI-driven technologies, foster collaborations between academia and industry, and encourage public discourse on AI adoption and governance. Outcomes & Impacts: ? Public Engagement & AI Awareness - Helped demystify robotics and AI by addressing common concerns and misconceptions, promoting informed discussions on AI ethics and applications. ? Strengthening Industry-Academia Collaborations - Encouraged dialogue between research institutions and tech companies, fostering potential research partnerships and innovation projects. ? Encouraging Thought Leadership in AI & Robotics - Engaged with professionals and policymakers to shape discussions on responsible AI deployment and its long-term societal impact. ? Inspiring Future Innovators - Sparked interest among students, early-career researchers, and entrepreneurs, encouraging them to pursue AI and robotics-driven careers. This panel discussion reinforced the importance of AI and robotics in shaping the future, while highlighting the role of ethical, human-centered AI in everyday life and industry. |
| Year(s) Of Engagement Activity | 2025 |
| URL | https://www.linkedin.com/posts/hellenic-tech_hellenictech-london-robotics-activity-72900006416686202... |
| Description | Invited Talk (Keynote), ICRA'24 Workshop on Agile Robotics, Japan |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Postgraduate students |
| Results and Impact | Delivered a keynote presentation at the ICRA 2024 Workshop on Agile Robotics, discussing advancements in robot perception, learning, and locomotion for agile robotic systems. Explored state-of-the-art methods in adaptive locomotion, whole-body control, and perception-driven navigation, focusing on legged robots operating in complex, unstructured environments. Engaged with leading researchers, industry professionals, and students to discuss emerging trends, challenges, and future directions in agile robotics. Outcomes & Impacts: ? Influencing Research in Agile Robotics - Provided insights into cognitive perception and adaptive motion planning, inspiring new research in autonomous, high-mobility robots. ? Fostering International Collaborations - Strengthened connections between academia, industry, and research labs, encouraging potential joint projects and funding opportunities. ? Knowledge Exchange with the Global Robotics Community - Contributed to discussions on how AI-driven perception and control techniques can enhance robot agility and robustness. ? Engaging the Next Generation of Roboticists - Inspired PhD students and early-career researchers, shaping the future landscape of agile robotics. This keynote talk helped advance the global conversation on agile robotics, reinforcing the importance of intelligent perception and adaptive locomotion in the next generation of autonomous robotic systems. |
| Year(s) Of Engagement Activity | 2024 |
| URL | https://agile-robotics-workshop.github.io/icra2024/ |
| Description | Invited Talk (Keynote), UCL Robotics Institute Workshop, UK |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Other audiences |
| Results and Impact | Delivered a keynote presentation at the UCL Robotics Institute Workshop in September 2024, providing an overview of robotics research at UCL Computer Science (UCL-CS). Highlighted ongoing projects, key advancements, and interdisciplinary collaborations in areas such as robot perception, learning, autonomous systems, and human-robot interaction. Aimed to foster collaboration between research groups, industry, and students, encouraging discussions on future directions for robotics at UCL and beyond. Outcomes & Impacts: ? Strengthening Research Collaboration - Engaged faculty, researchers, and students in discussions about robotics research at UCL, fostering potential new collaborations. ? Raising Awareness of Robotics at UCL-CS - Provided an opportunity to showcase the department's work, increasing its visibility within the broader robotics and AI community. ? Inspiring the Next Generation of Roboticists - Encouraged students and early-career researchers to explore robotics research opportunities, leading to increased participation in robotics projects. ? Industry and Academic Engagement - Facilitated conversations with industry representatives and external research groups, helping to align academic research with real-world applications. This keynote talk contributed to shaping the strategic direction of robotics research at UCL, reinforcing the importance of interdisciplinary collaboration and innovation in autonomous systems. |
| Year(s) Of Engagement Activity | 2024 |
| Description | Invited Talk (Robotics and AI Workshop), AIxSET, USA. |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Postgraduate students |
| Results and Impact | Delivered an invited talk on "Cognitive Legged Robots for Real-World Applications" at the AIxSET Robotics and AI Workshop in October 2024, hosted in the USA. Presented cutting-edge research on AI-driven perception, adaptive locomotion, and real-time decision-making for legged robots operating in complex environments. Highlighted real-world applications in search and rescue, autonomous exploration, environmental monitoring, and industrial automation, emphasizing how cognitive AI models enhance robot autonomy. Aimed to foster international collaborations and explore how robotics and AI researchers can bridge the gap between academic research and real-world deployment Outcomes & Impacts: ? Advancing AI-Driven Legged Robotics - Showcased state-of-the-art methods for perception-aware locomotion and autonomous decision-making, inspiring new research directions in robotic mobility and cognitive intelligence. ? Strengthening Global Research Collaborations - Engaged with leading AI and robotics experts from both academia and industry, fostering discussions on collaborative projects and technology transfer. ? Inspiring Next-Generation Researchers - Encouraged PhD students and early-career researchers to pursue AI-integrated robotics research, strengthening interest in cognitive legged robots. ? Industrial and Academic Impact - Opened opportunities for cross-sector collaboration, particularly in robotic automation for hazardous environments, logistics, and precision agriculture. This talk contributed to shaping the future of AI-enhanced robotic mobility, reinforcing the importance of cognitive perception and adaptive control in real-world legged robotics applications. |
| Year(s) Of Engagement Activity | 2024 |
| URL | https://www.aixset.org/roboticsai |
| Description | Invited Talk 2024: Dawes Centre for Future Crime |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Postgraduate students |
| Results and Impact | 50 students of the EPSRC CDT in Cybersecurity attended the talk on how robots are and will be used in security and defense. |
| Year(s) Of Engagement Activity | 2024 |
| Description | Invited Talk at Archimedes/Athena, Greece |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Postgraduate students |
| Results and Impact | Delivered an invited talk on "Cognitive Legged Robots for Real-World Applications" at Archimedes/Athena Research Center, Greece, in September 2024. Presented advancements in AI-driven perception, locomotion, and decision-making for legged robots operating in unstructured environments. Discussed real-world applications in agriculture, search and rescue, environmental monitoring, and industry, highlighting the importance of cognitive capabilities in robotic autonomy. Aimed to foster collaboration between Greek and international AI/robotics researchers, aligning research efforts with practical, high-impact applications. Outcomes & Impacts: ? Strengthening Research Ties with Greece - Encouraged collaboration between UCL and Greek research institutions, particularly in cognitive robotics and AI. ? Advancing Legged Robotics for Real-World Deployment - Provided insights into new methodologies for adaptive locomotion, perception-driven navigation, and decision-making in autonomous robots. ? Engaging with Researchers and Industry Experts - Initiated discussions on how AI-powered legged robots can support critical industries, such as agriculture, disaster response, and infrastructure inspection. ? Inspiring Future Researchers - Sparked interest among early-career researchers and PhD students, leading to potential joint projects and academic exchanges. This talk played a key role in bridging cognitive robotics research with real-world needs, strengthening international collaboration, and highlighting the impact of AI-driven legged robots in practical applications. |
| Year(s) Of Engagement Activity | 2024 |
| Description | Invited Talk at Arrival LTD |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | Regional |
| Primary Audience | Industry/Business |
| Results and Impact | An invited talk about cognitive real-world legged robotics in industrial inspection, that showcased the advancements we had in my group. |
| Year(s) Of Engagement Activity | 2023 |
| Description | Invited Talk at University of Southern California |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Postgraduate students |
| Results and Impact | An invited talk to PhD students about Cognitive Real-World robotics, that showcased the advancements we had in my group. |
| Year(s) Of Engagement Activity | 2023 |
| Description | Invited Talk at University of Waterloo |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Postgraduate students |
| Results and Impact | An invited talk to PhD students about Cognitive Real-World robotics, that showcased the advancements we had in my group. |
| Year(s) Of Engagement Activity | 2023 |
| Description | Invited Talk to the Royal Veterinary College in the UK |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Postgraduate students |
| Results and Impact | In brief: A multi-disciplinary invited talk at the RVC, titled "Robots with Legs & Arms: Cognition for Real-World Loco/Manipulation in Complex Environments", to target collaboration between animal researchers and roboticists. Impact: Grant proposal writing and collaboration on animal/robot study. |
| Year(s) Of Engagement Activity | 2022 |
| Description | Invited Talk, Dawes Centre for Future Crime, UCL, UK |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | Regional |
| Primary Audience | Policymakers/politicians |
| Results and Impact | Title: "Police Robots: Crime Fighting and Law Enforcement vs. Trustworthiness and Ethics" Intended Purpose: 1. To explore the role of robotics in law enforcement, discussing how autonomous systems can support policing efforts in crime prevention, surveillance, and response. 2. To examine ethical and trust-related challenges, including public perception, accountability, and potential biases in AI-driven policing. 3. To engage interdisciplinary audiences from criminology, law, AI ethics, and robotics in a critical discussion on the future of autonomous systems in law enforcement. Outcomes & Impacts: ? Interdisciplinary Dialogue on AI & Policing - The talk sparked discussions on balancing security benefits with ethical concerns, influencing perspectives on robot-assisted crime prevention strategies. ? Engagement with Policymakers & Researchers - Brought together law enforcement professionals, criminologists, and AI experts, fostering potential collaborations for responsible technology deployment. ? Increased Awareness of Ethical AI in Law Enforcement - Encouraged deeper exploration of human oversight, transparency, and trust in robotic policing technologies. ? Potential Future Research Directions - Inspired discussions on policy frameworks, legal implications, and AI governance for robotic law enforcement. This talk contributed to shaping the discourse on robotics in policing, highlighting both technological possibilities and the ethical responsibilities associated with autonomous law enforcement systems. |
| Year(s) Of Engagement Activity | 2024 |
| Description | Invited Talk, UCL Dawes Centre for Future Crime, UK |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Policymakers/politicians |
| Results and Impact | Delivered an invited talk at the Sandpit on Social Robotics and Crime, hosted by the UCL Dawes Centre for Future Crime in December 2024. Explored the role of social robots in crime prevention, policing, and public safety, discussing both technological advancements and ethical challenges. Examined human-robot interaction (HRI) in law enforcement, including issues of trust, accountability, and public perception. Facilitated an interdisciplinary dialogue between experts in AI, robotics, criminology, law enforcement, and ethics, aiming to define future research directions and policy frameworks for social robotics in crime prevention. Outcomes & Impacts: ? Shaping Research in Social Robotics & Crime Prevention - Encouraged new research on AI-driven crime deterrence and law enforcement support systems. ? Fostering Interdisciplinary Collaboration - Engaged experts from robotics, criminology, psychology, and AI ethics, fostering future partnerships on socially responsible AI applications. ? Contributing to AI & Robotics Policy Development - Provided insights into the governance of robotic law enforcement, informing ethical and regulatory discussions on AI in policing. ? Engaging Law Enforcement & Policy Stakeholders - Increased awareness among policymakers and law enforcement agencies about potential applications and limitations of robotics in crime prevention. This event helped to bridge the gap between robotics, ethics, and crime prevention, reinforcing UCL's leadership in AI governance and future crime research. |
| Year(s) Of Engagement Activity | 2024 |
| Description | Invited lecture to the UCL Science Centre - School Outreach Lectures |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Schools |
| Results and Impact | In brief: each year, giving a talk about ``Real-World Legged Robots'' at UCL Science Centre lectures (the talk was sent to about 300 schools, with great engagement). Impact: several pupils contacted after the lecture (either via email or LinkedIn) to ask further info or questions on the related subject area. |
| Year(s) Of Engagement Activity | 2022,2023 |
| URL | https://www.ucl.ac.uk/physics-astronomy/outreach/science-centre-lectures |
| Description | Invited talk at Imperial College London |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Postgraduate students |
| Results and Impact | An invited talk to PhD students and Professors about Cognitive Real-World robotics, that showcased the advancements we had in my group. |
| Year(s) Of Engagement Activity | 2023 |
| URL | https://youtu.be/49Wb3XGryDo?si=fbrSGK-taWYB9kFc |
| Description | Invited talk at Science Centre Lectures |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Schools |
| Results and Impact | An important online talk for schools about robotics and the advancements of my group in legged robots, that influences the direction of study for hundreds of pupils. |
| Year(s) Of Engagement Activity | 2022,2023 |
| URL | https://www.ucl.ac.uk/physics-astronomy/outreach/science-centre-lectures |
| Description | Invited talk at UCL Robotics Workshop |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Postgraduate students |
| Results and Impact | A talk at the yearly organised workshop of robotics at UCL, with invited people from the unveristy and industry, where we also demonstrated our recent method on the legged robots to the public. |
| Year(s) Of Engagement Activity | 2023 |
| URL | https://www.ucl.ac.uk/robotics/events/2023/feb/ucl-robotics-23 |
| Description | Invited talk for Newham CIEAG Network |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Schools |
| Results and Impact | A talk to school teachers about robotics and the development at UCL East and the local neighborhood in terms of automation and future works. |
| Year(s) Of Engagement Activity | 2023 |
| Description | Invited talk for Spring into STEM 2.0 at UCL |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | Regional |
| Primary Audience | Schools |
| Results and Impact | 50 pupils attended for a talk about robotics in making a difference through engineering, which sparked questions and discussion afterwards, and the school reported increased interest in related subject areas. |
| Year(s) Of Engagement Activity | 2023,2024 |
| URL | https://www.ucl.ac.uk/engineering/study/spring-stem-20-making-difference-through-engineering |
| Description | Invited talk to 2023 UCL Robotics Workshop |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Industry/Business |
| Results and Impact | In Brief: An invited talk about the UKRI FLF to academics and industry partners from all the UK, titled: "Robots with Legs & Arms: Cognition for Real-World Loco/Manipulation in Complex Environments". Impact: Engagement with industry, as well as outreaching to academics from other institution that work on similar topics. |
| Year(s) Of Engagement Activity | 2023 |
| URL | https://www.ucl.ac.uk/robotics/events/2023/feb/ucl-robotics-23 |
| Description | Invited talk to UK-RAS Early-Career Seminar |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Postgraduate students |
| Results and Impact | In brief: UK-RAS Early-Career team invited Dr Kanoulas to share the path towards the UKRI FLF, and accept questions from early career academics that aim at similar paths. Impact: received several emails to help with early carer paths. |
| Year(s) Of Engagement Activity | 2023 |
| URL | https://youtu.be/9M8j87XMZac |
| Description | Invited talk to University of Colorado at Boulder |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Postgraduate students |
| Results and Impact | In Brief: A PhD seminar talk about Cognitive Real-World Loco-Manipulation. Impact: Outreaching to international academics about our work resulted by the FLF and seek for collaborations. |
| Year(s) Of Engagement Activity | 2023 |
| Description | Media coverage on the fellowship |
| Form Of Engagement Activity | A press release, press conference or response to a media enquiry/interview |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Public/other audiences |
| Results and Impact | In Brief: The UKRI Future Leaders Fellowship: RoboHike appeared to more than 30 newspapers, blogs, journals, online news, etc., including Telegraph and Yahoo!, while featured by UKRI and the UK Government itself. Impact: the news reached to all possible audiences and familiarised the UK and the world with the new type of robots that will help humans in real-world environments. |
| Year(s) Of Engagement Activity | 2022 |
| URL | https://www.gov.uk/government/news/next-generation-of-uk-science-leaders-backed-with-113-million-to-... |
| Description | Mentors panel formation - Andrew Davison (ICL), Marc Deisenroth (UCL) |
| Form Of Engagement Activity | A formal working group, expert panel or dialogue |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Study participants or study members |
| Results and Impact | In brief: We formed a group of mentors for the fellowship, that are experts in various topics, including team management, research in the field of Machine Learning and Robotics, as well as formation of spin-offs. Impact: Initial discussions with the mentors, provided clear understanding of the good practices regarding the fellowship and helped with the formation of the working group and starting development of research methods. |
| Year(s) Of Engagement Activity | 2022,2023 |
| Description | Outreach Video for UCL-CS new MEng programma in Robotics and AI |
| Form Of Engagement Activity | A broadcast e.g. TV/radio/film/podcast (other than news/press) |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Schools |
| Results and Impact | A video about the new MEng programme of Robotics and AI at UCL with a focus to legged robots, attracting the interest of several students (500 applications were reported). |
| Year(s) Of Engagement Activity | 2023 |
| URL | https://youtu.be/pZQOAgESIcs?si=2ZaA8d382Fr3VEeI |
| Description | Panelist (Workshop), 13th Conference of the Hellenic Artificial Intelligence Society (SETN 2024), UK |
| Form Of Engagement Activity | A formal working group, expert panel or dialogue |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Postgraduate students |
| Results and Impact | Participated as a panelist in the "Cognitive Robotics and AI" workshop at SETN 2024, discussing the intersection of robotics, artificial intelligence, and cognitive systems. Explored how AI-driven perception, learning, and decision-making are shaping the future of autonomous robots in dynamic and unstructured environments. Engaged in discussions with leading experts, including Prof. Yiannis Aloimonos (University of Maryland), AI researchers, and roboticists, on emerging trends, ethical concerns, and challenges in cognitive robotics. Outcomes & Impacts: ? Advancing Cognitive Robotics Research - Contributed to the exchange of ideas on AI-powered autonomy, human-robot interaction, and cognitive perception, helping define future research directions. ? Strengthening International Collaboration - Encouraged interdisciplinary partnerships between Greek and international AI and robotics researchers, fostering potential future collaborations. ? Engaging with the AI and Robotics Community - Provided insights into state-of-the-art research in cognitive robotics, sparking discussions on AI-driven decision-making and real-world deployment. ? Inspiring Early-Career Researchers and Students - Encouraged young researchers to explore cognitive robotics and AI, increasing participation in the field and enhancing future talent pipelines. This panel discussion helped to bridge AI and robotics research, reinforcing the role of cognitive systems in advancing autonomous, intelligent robotics. |
| Year(s) Of Engagement Activity | 2024 |
| URL | https://robotics.pme.duth.gr/workshop_cogrob/ |
| Description | Panelist for UCL Engineering, UK |
| Form Of Engagement Activity | Participation in an open day or visit at my research institution |
| Part Of Official Scheme? | No |
| Geographic Reach | Regional |
| Primary Audience | Schools |
| Results and Impact | Participated as a panelist in the "Future: Battle of Ideas" discussion at the Festival of Engineering (UCL) in July 2024. Engaged in critical discussions on the future of technology, AI, and robotics, exploring their impact on society, industry, and ethics. Provided expert insights into autonomous systems, AI-driven robotics, and human-robot interaction, contributing to debates on innovation, policy, and ethical challenges. Outcomes & Impacts: ? Public Engagement & Knowledge Exchange - Encouraged thought-provoking discussions among students, researchers, and industry professionals, sparking critical reflections on the role of AI and robotics in shaping the future. ? Interdisciplinary Collaboration - Brought together experts from diverse fields, fostering cross-disciplinary dialogue on engineering, AI ethics, and societal implications. ? Inspiration for Future Engineers & Researchers - The session engaged schools, general public, students and early-career researchers, increasing interest in robotics, AI, and autonomous systems. ? Shaping the Future of AI & Robotics Policy - Contributed to discussions on responsible AI, human-centered design, and sustainable robotics, influencing perspectives on technology governance and societal impact. This event played a key role in bridging the gap between research, policy, and public discourse, reinforcing the importance of ethical and forward-thinking technological development. |
| Year(s) Of Engagement Activity | 2024 |
| URL | https://www.ucl.ac.uk/engineering/festival-engineering |
| Description | Podcast Interview about the new IM@UCL |
| Form Of Engagement Activity | A broadcast e.g. TV/radio/film/podcast (other than news/press) |
| Part Of Official Scheme? | No |
| Geographic Reach | Regional |
| Primary Audience | Public/other audiences |
| Results and Impact | In brief: a podcast on "Robotics and Simulation: For or Against Humanity", promoting AI for Good and how robots can be beneficial in our society. Impact: change the view of people afraid of robots and the future with them. |
| Year(s) Of Engagement Activity | 2023 |
| URL | https://soundcloud.com/uclsound/ep-5-robotics-and-simulation-for-or-against-humanity |
| Description | Railway Industry Association (RIA) Innovation Conference 2024 |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Industry/Business |
| Results and Impact | This year, RIA introduced the "Future Focus Zone", a dedicated space within the conference aimed at highlighting groundbreaking innovations in fields like robotics, artificial intelligence, nanotechnology, virtual reality, and immersive technologies. Their goal is to spotlight the pioneering work being done in academic and research institutions, which might not typically be featured in a commercial exhibition setting. We were invited to demonstrae our work. |
| Year(s) Of Engagement Activity | 2024 |
| URL | https://www.riagb.org.uk/RIA/Events/2024/RIC_2024/RIA_Innovation_Conference_2024.aspx |
| Description | RoboHike Working Group |
| Form Of Engagement Activity | A formal working group, expert panel or dialogue |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Postgraduate students |
| Results and Impact | In brief: we have formed a group of research fellows, research assistants, PhD students, MSc students, and BSc/MEng students to work collaboratively on the project, by meeting monthly and solving one task at a time. Impact: Quick progress on the fellowship's goals and aims. |
| Year(s) Of Engagement Activity | 2022,2023 |
| Description | Robotic 3D Building news |
| Form Of Engagement Activity | A press release, press conference or response to a media enquiry/interview |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Public/other audiences |
| Results and Impact | In brief: several online news covered our work on 3D printing buildings Impact: excitement on the achievement showing proof of value for robots on the next generation. This led to discussions with Arup on funding a PhD student on the topic. |
| Year(s) Of Engagement Activity | 2022 |
| URL | https://www.fabbaloo.com/news/robotic-3d-print-research-shows-large-scale-possibilities |
| Description | Robotics and Artificial Intelligence MEng Information Session |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Schools |
| Results and Impact | In brief: expert view from our Information Session webinar (Oct 2022) for our new undergraduate programme, launching Sept 2023. Impact: pupils showed high interest in robotics, with special focus to under-represented in engineering groups, e.g., female or black students. |
| Year(s) Of Engagement Activity | 2022 |
| URL | https://youtu.be/_LJqDmljw94 |
| Description | Round Table Discussion to Greek Embassy in London |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | Regional |
| Primary Audience | Public/other audiences |
| Results and Impact | A round table discussion about the AI Revolution, with respect to practical capabilities and challenges, with several start-ups and academics being involved. It resulted to discussions over the future of AI with diverse audience. |
| Year(s) Of Engagement Activity | 2023 |
| URL | https://www.eventbrite.co.uk/e/the-ai-revolution-opportunities-challenges-in-the-greek-ecosystem-tic... |
| Description | School Visits |
| Form Of Engagement Activity | Participation in an open day or visit at my research institution |
| Part Of Official Scheme? | No |
| Geographic Reach | Regional |
| Primary Audience | Schools |
| Results and Impact | We have an extensive request for visits at the robotics lab at UCL East, to demonstrate the capabilities of our legged robots, with tens of Schools, industrial partners, and academics every week. This influences their decisions and view of robotics in the real world. |
| Year(s) Of Engagement Activity | 2023,2024 |
| Description | Scientific Advisors: Lourdes Agapito (UCL), Kostas Alexis (NTNU), Maren Bennewitz (Uni-Bonn), Enrico Mingo Hoffman (PAL Robotics), John Hutchinson (RVC), Simon Julier (UCL), Christopher Kilburn (UCL), Nikos Tsagarakis (IIT) |
| Form Of Engagement Activity | A formal working group, expert panel or dialogue |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Study participants or study members |
| Results and Impact | In brief: We formed a group of scientific advisors for the fellowship, that are experts in various topics, firstly to support the research and developments and secondly to foster further collaborations Impact: Initial discussions with the scientific advisors, regarding particular problems in the research resulted in to potential solutions. Moreover, we had collaboration initiatives with K. Alexis (mapping), J. Hutchinson (animal study and EPSRC grant proposal), and C. Kilburn (volcano inspection). |
| Year(s) Of Engagement Activity | 2022,2023 |
| Description | Summer School Organisation and Talk at Queen Mary University of London |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | Regional |
| Primary Audience | Postgraduate students |
| Results and Impact | The main organiser of a summer school in Robotics as a collaboration between the 4 big universities in London, UCL, QMUL, KCL, and ICL. This is yearly and in 2024 it will be at UCL, affecting the local research and industrial community. |
| Year(s) Of Engagement Activity | 2023,2024 |
| URL | https://www.linkedin.com/posts/dkanoulas_london-is-a-hub-in-robotics-and-thus-we-started-activity-71... |
| Description | UCL AI Showcase |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Postgraduate students |
| Results and Impact | Demo by our legged robots for the AI Showcase (November 2022). Audience included the UCL Provost and UCL Deans. |
| Year(s) Of Engagement Activity | 2022 |
| URL | https://blogs.ucl.ac.uk/faicdt/2022/11/04/cdt-students-shine-at-poster-showcase-event/ |
| Description | UCL Gatsby and SWC (Sainbury Welcome Centre) Workshop |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Postgraduate students |
| Results and Impact | Every year, the UCL Gatsby and SWC (Sainbury Welcome Centre) PhD intake students participate in a 3-week bootcamp. During the bootcamp they unpack "blackboxes", i.e., everyday subsystems and concepts which we take for granted e.g., transistors, computer architecture, motors, sensors, control etc. The students spend the 3-weeks tinkering and interacting with these "blackboxes" and eventually build mobile robots with arduino's and raspberry pi's. At the end of the bootcamp, the robots are showcased to the entire institute. This year, the students are joined by the Go1 robot from UCL's Robohike project to play follow the leader. Here, the smaller robots will follow the larger Go1 robot around the room using bespoke infrared sensors. |
| Year(s) Of Engagement Activity | 2022 |
| Description | UCL News Spotlight |
| Form Of Engagement Activity | A press release, press conference or response to a media enquiry/interview |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Other audiences |
| Results and Impact | In brief: a spotlight interview around the fellowship that was published at UCL. Impact: several emails about potential collaborations towards the fellowship's goals, including research and outreaching |
| Year(s) Of Engagement Activity | 2022 |
| URL | https://www.ucl.ac.uk/news/2022/aug/spotlight-dimitrios-kanoulas |
| Description | UCL OVPA Case Study |
| Form Of Engagement Activity | A magazine, newsletter or online publication |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Postgraduate students |
| Results and Impact | An interview about my work with legged robots that would probably influence collaborators, industrial partners, and donor. |
| Year(s) Of Engagement Activity | 2023 |
| URL | https://www.ucl.ac.uk/giving/case-studies/2023/jun/ucl-east-dr-dimitrios-kanoulas-ukri-future-leader... |
| Description | UCL Spotlight News |
| Form Of Engagement Activity | A press release, press conference or response to a media enquiry/interview |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Media (as a channel to the public) |
| Results and Impact | Spotlight interview to engage people with UK robotics and developments at UCL East. |
| Year(s) Of Engagement Activity | 2022 |
| URL | https://www.ucl.ac.uk/news/2022/aug/spotlight-dimitrios-kanoulas |
| Description | UCL-CS News Spotlight |
| Form Of Engagement Activity | A magazine, newsletter or online publication |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Postgraduate students |
| Results and Impact | tbc |
| Year(s) Of Engagement Activity | 2022 |
| URL | https://liveuclac.sharepoint.com/sites/ComputerScienceIntranet/SitePages/Staff-Spotlight.aspx |
| Description | UKRI visit - 27 March - Robot dogs |
| Form Of Engagement Activity | Participation in an open day or visit at my research institution |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Policymakers/politicians |
| Results and Impact | Participated in the UKRI Advisory Infrastructure Committee visit to UCL East on 27 March 2024, engaging with key academic leads and infrastructure experts. Provided insights into ongoing robotics and AI research, showcasing advancements in autonomous systems, intelligent perception, and mobile robotics. Discussed UCL's cutting-edge research facilities, emphasizing the role of robotics and AI in shaping the future of research infrastructure and innovation. Facilitated discussions on potential funding opportunities, research infrastructure needs, and interdisciplinary collaboration. Outcomes & Impacts: ? Strengthening Research Infrastructure Discussions - Contributed to strategic planning for future investments in robotics, AI, and autonomous systems research facilities. ? Engagement with UKRI Leadership - Provided an opportunity to highlight UCL's expertise in robotics and AI, influencing funding and policy decisions. ? Showcasing UCL East as a Hub for Innovation - Reinforced UCL's position as a leading institution in research and technology development, fostering potential collaborations. ? Networking & Future Collaborations - Engaged with key decision-makers, creating opportunities for future partnerships and funding discussions. This visit played a crucial role in aligning UCL's research priorities with national infrastructure strategies, ensuring continued support for robotics and AI innovation in the UK. |
| Year(s) Of Engagement Activity | 2024 |
| Description | Visiting School at the UCL Science Museum - Demo |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Schools |
| Results and Impact | In brief: we ran some demos of our legged robots to ~50 pupils, as part of their visit to the Science Museum, to promote our research findings and the "AI for Good", also in order to consider Computer Science as their next field of study. Impact: pupils (female of under-represented groups) showed high interest in robotics and applying for computer science. |
| Year(s) Of Engagement Activity | 2022,2023 |
| URL | https://twitter.com/SBonnellSchool/status/1618998641544364032 |