Sense and algorithm for operation of multiple UAS
Lead Research Organisation:
Cranfield University
Department Name: Sch of Aerospace, Transport & Manufact
Abstract
As the capabilities of unmanned aircraft systems (UAS) has been developed significantly in recent decades, the application areas of UAS have been rapidly extended. In the near future, heterogeneous and multiple UAS with different tasks are expected to operate at the same time in the same airspace. Moreover, the airspace for UAS operation is also predicted to expand to the airspace for manned aircraft systems. In order for UAS to fly with numerous other manned and unmanned aircraft systems, the high performance on sense and avoid is an essential requirement. The aim of this PhD is to develop an integrated sense and avoid algorithm for UAS to make immediate decisions from the sensor measurements with minimum signal processing for fast response with high reliability.
Studentship Projects
| Project Reference | Relationship | Related To | Start | End | Student Name |
|---|---|---|---|---|---|
| EP/V519509/1 | 01/10/2020 | 30/09/2027 | |||
| 2454266 | Studentship | EP/V519509/1 | 01/10/2020 | 30/09/2024 | Bilkan Ince |
| Description | Operations involving small Unmanned Aerial Systems (sUAS) in urban environments are occurring ever more frequently as recognized applications gain acceptance, and new use cases emerge, such as urban air mobility, medical deliveries, and support of emergency services. The presence of Detect and Avoid (DAA) capability of sUAS is one of the major requirements for its safe operation in urban environments. The platform or its operator proves a full awareness of all potential obstacles within the mission, maintains a safe distance from other airspace users, and, ultimately, performs Collision Avoidance (CA) manoeuvres to avoid imminent impacts. The research addresses two sets of problems for Unmanned Aerial Systems (sUAS) operations: implementation of a hazard assessment system and implementation of a sense and avoid algorithm improving the speed at which the detection and avoidance strategies by combining both detection and avoidance algorithms together using AI specifically Reinforcement Learning. Safety/Hazard Assessment: The research results devise a DAA system's "Well Clear" recommendation combines a Remain Well Clear (RWC) and an optional Collision Avoidance (CA) function. The RWC function performs tactical manoeuvres to maintain a Well Clear status, whereas the CA function performs emergency manoeuvres to avoid mid-air collisions. Dynamic objects such as sUAS, birds, etc., are considered through the velocity of the objects. Static objects such as building, foliage, etc., safety volume is defined with a set of points apart from the obstacle with the same distance. Complex shapes safety volume is defined as the minimum size of cylinder covering the objects with a certain margin from the objects. The model was inspired by NASAs idea of a volumetric hazard assessment. Higher the level of hazard the greater the avoidance rate. Sense and Avoid utilising Reinforcement Learning: Comparison between the practical applications of reinforcement learning, the performance of a conventional reinforcement learning and performance of a robust reinforcement learning usage in real-life experimentation. The platform for simulations took place in Python and Airsim, as this provided with a better implementation of reinforcement learning using airsims several quadrotors. Software in the loop tests will comprise of several different simulation environments these including locations in Airsim/Cesium starting with DARTeC (Digital Aviation Research and Technology Centre), Cranfield, UK. Experimental tests were carried out in Snowdonia Aerospace Centre, UK. For each test mission the flight phases consisted of take-off, cruise, and landing. Simulations consider the role of robustness in reinforcement learning as the inclusion of robustness allows the consideration of model and parameter uncertainties present within the MDP (Markov Decision Process) compared to the conventional reinforcement learning. Several tests were conducted to represent the comparison of the both conventional and robust reinforcement learning During the development of the conventional reinforcement learning, the application of the algorithms was implemented to a Jetson Nano during the flight trials in Snowdonia where several ArduPilot functions were conducted and tuned for the flight of the reinforcement learning Jetson Nano. Results of the drone trial proved that more development was necessary to have better control over external uncertainties and disturbances. Conventional RL has higher probability of collision as it does not consider change in dynamics of the environment setting whereas Robust RL allows the probability of unknown disturbances to be limited to a constraint to allow better avoidance. |
| Exploitation Route | The outcomes of this research can be utilised in robotics, AI and drone applications. In addition, the research aids the certification of AI implemented small unmanned air systems (sUAS) in civil and military applications, implying the usage of autonomised drone usage with the safety factor within the general public. |
| Sectors | Aerospace, Defence and Marine,Education |
| Description | Findings were used in a UKRI project (INMED-Enabling Infrastructure for Medical Drone Deliveries) funded by HEROTECH8, Bluebear, NHS Foundation Trust and Intelsius, the project delivered medical packages containing blood samples between NHS owned hospitals. The delivery of medical drones, comprising a low-maintenance launch platform for drones with integrated automatic recharging, a drone with a temperature-controlled housing, which is approved and compliant for the medical transport of blood/blood products, operating procedures and a management application. The drone loading, take-off, landing and unloading will be demonstrated in the project at the Milton Keynes University Hospital site, while the flight from the blood bank to the hospital will be simulated for analysis of risks and development of risk mitigation measures related to airspace management, communication, and navigation. |
| First Year Of Impact | 2022 |
| Sector | Aerospace, Defence and Marine,Digital/Communication/Information Technologies (including Software),Transport |
| Impact Types | Societal |
| Description | INMED - Enabling Infrastructure for Medical Drone Deliveries |
| Organisation | Blue Bear |
| Country | United Kingdom |
| Sector | Private |
| PI Contribution | Devise a safety-critical criteria for manned and unmanned drone operations. Utilizing Unreal Engine and MATLAB analysis of the findings and simulation results leads to a holistic approach to implementation of sUAS operations in urban environments, focusing on extracting critical DAA capability for safe mission completion. The proposed approach forms a valuable asset for safe operations validation, enabling better evaluation of risk mitigation for sUAS urban operations and safety-focused design of the sensor payload and algorithms. |
| Collaborator Contribution | The integration of drones into the UK transport system for their widespread use for commercial deliveries. For this purpose, we focus on development and integration of physical and digital infrastructure into the practice of medical supplies delivery between blood banks and UK hospitals, paying special attention on integration of the drone operations into medical operational practices. |
| Impact | Conference paper - DOI: 10.1109/DASC52595.2021.9594407 |
| Start Year | 2021 |
| Description | INMED - Enabling Infrastructure for Medical Drone Deliveries |
| Organisation | Cranfield University |
| Country | United Kingdom |
| Sector | Academic/University |
| PI Contribution | Devise a safety-critical criteria for manned and unmanned drone operations. Utilizing Unreal Engine and MATLAB analysis of the findings and simulation results leads to a holistic approach to implementation of sUAS operations in urban environments, focusing on extracting critical DAA capability for safe mission completion. The proposed approach forms a valuable asset for safe operations validation, enabling better evaluation of risk mitigation for sUAS urban operations and safety-focused design of the sensor payload and algorithms. |
| Collaborator Contribution | The integration of drones into the UK transport system for their widespread use for commercial deliveries. For this purpose, we focus on development and integration of physical and digital infrastructure into the practice of medical supplies delivery between blood banks and UK hospitals, paying special attention on integration of the drone operations into medical operational practices. |
| Impact | Conference paper - DOI: 10.1109/DASC52595.2021.9594407 |
| Start Year | 2021 |
| Description | INMED - Enabling Infrastructure for Medical Drone Deliveries |
| Organisation | King's College London |
| Country | United Kingdom |
| Sector | Academic/University |
| PI Contribution | Devise a safety-critical criteria for manned and unmanned drone operations. Utilizing Unreal Engine and MATLAB analysis of the findings and simulation results leads to a holistic approach to implementation of sUAS operations in urban environments, focusing on extracting critical DAA capability for safe mission completion. The proposed approach forms a valuable asset for safe operations validation, enabling better evaluation of risk mitigation for sUAS urban operations and safety-focused design of the sensor payload and algorithms. |
| Collaborator Contribution | The integration of drones into the UK transport system for their widespread use for commercial deliveries. For this purpose, we focus on development and integration of physical and digital infrastructure into the practice of medical supplies delivery between blood banks and UK hospitals, paying special attention on integration of the drone operations into medical operational practices. |
| Impact | Conference paper - DOI: 10.1109/DASC52595.2021.9594407 |
| Start Year | 2021 |
| Description | INMED - Enabling Infrastructure for Medical Drone Deliveries |
| Organisation | Milton Keynes Hospital NHS Foundation Trust |
| Country | United Kingdom |
| Sector | Public |
| PI Contribution | Devise a safety-critical criteria for manned and unmanned drone operations. Utilizing Unreal Engine and MATLAB analysis of the findings and simulation results leads to a holistic approach to implementation of sUAS operations in urban environments, focusing on extracting critical DAA capability for safe mission completion. The proposed approach forms a valuable asset for safe operations validation, enabling better evaluation of risk mitigation for sUAS urban operations and safety-focused design of the sensor payload and algorithms. |
| Collaborator Contribution | The integration of drones into the UK transport system for their widespread use for commercial deliveries. For this purpose, we focus on development and integration of physical and digital infrastructure into the practice of medical supplies delivery between blood banks and UK hospitals, paying special attention on integration of the drone operations into medical operational practices. |
| Impact | Conference paper - DOI: 10.1109/DASC52595.2021.9594407 |
| Start Year | 2021 |