From Sensing to Collaboration: Engineering, Exploring and Exploiting the Building Blocks of Embodied Intelligence - An EPSRC Programme Grant

Lead Research Organisation: University of Oxford
Department Name: Engineering Science

Abstract

Our robots are too specialised, too impoverished in their sensing, too uncooperative and too unsafe to be productive at scale. To contribute to productivity in strategically important areas such as social care, manufacturing, logistics, service, inspection or agriculture, future generations of robots need to be able to sense, interpret, act, navigate, coordinate and collaborate with an hitherto unprecedented acuity.

VISION: The overarching aim of this research programme is to deliver autonomous systems which amplify human capacity and potential. These robots must be capable of performing a broad array of bespoke tasks effectively, and with a minimum of operator intervention. In a sustainable national centre of excellence we will grow the technology and people substrate for robust embodied intelligence, i.e. the science and technology to enable robots to robustly and flexibly act, interact and collaborate in the real world.

STRATEGY: Our focus is on engineering, exploring and exploiting the building blocks of integrated embodied intelligence to deliver autonomous systems which, over the course of their life-time, acquire the sensing, perception, manipulation, navigation, collaboration and problem solving abilities required to allow them to operate unaided while significantly enhancing human productivity. We will significantly expand the reach and versatility of robots in domains of strategic and commercial value by exploiting synergies across research disciplines, which only emerge when deploying robot systems. In doing so we are driven by both fundamental science questions and real-world applications. Together with our partners, we have a clear scope in mind: versatile, collaborative robots whose societal and economic footprint is vast. Our work will underpin a national strategic aim with a carefully considered and coherent programme of research: from sensing to collaboration.

Planned Impact

Robust, versatile embodied intelligence is crucial for RAS technology contributing to industry and society. It is a pivotal enabler in a number of large industrial sectors in the global economy with an ability to reach a truly vast audience. To realise this potential, it is imperative to combine academic excellence with an appropriate grounding of the research in real industrial need. We run a "petal" impact model. At the heart of the endeavour is research staff surrounded by expert engineering support. Within this team are knowledge exchange engineers and a project delivery coordinator who are charged with managing the flow of knowledge between industry partners and the researchers. Membership to the cadre of industrial partners is offered on a sector-by-sector basis. In this way our proposal will create lasting, catalysing impact on UKPLC.

Industrial Impact - the true impact, reach, ambition and richness of this proposal is best illustrated by some of our research Flagships: domain-specific, physical instantiations of integrated competencies developed across the entire research consortium. Examples include pick-and-carry operations and dressing assistance in Social Care (supported by Oxfordshire County Council and Scan Computers); small-batch, interactive manufacturing tasks (with Accenture, RACE and the Shadow Robot Company); stock-taking, bin picking and bin packing tasks in Logistics (with Ocado and Bossa Nova); precision farming applications (with Saga Robotics); and long-term, multi-agent inspection tasks and autonomous fault intervention (with BP, Navtech L3-ASV and Oxbotica). Our long-term plan is to act as a hub, which serves as a catalyst for investors, existing businesses and new enterprises to inform, engage in and exploit the next generation of RAS technologies.

Academic Impact - the award of a Programme Grant will stimulate the transition of robotics technology from an academic pedigree to real-world applications of industrial significance. It will consolidate and strengthen the UK research base across disciplines such as Robotics, HRI and Cognitive Science.

Societal Impact - we aspire for our robots to work with and alongside us in productive symbiosis - or even care for us. While scientific boundaries are receding, issues surrounding trust and acceptance of this technology into society become ever more apparent. Immediately, then, questions of Responsible Innovation must be asked. The Anticipate-Reflect-Engage-Act) paradigm, as set out by EPSRC, is of course apt here. The Programme Grant will: Anticipate results of related research on the impact of automation on employment; Reflect on and promote ethical thinking around intelligent machines and the societal benefits of interactions between robots and people; Engage with stakeholders and the general public to provide feedback surrounding the effects of our innovation; and Act to test examples of safe and effective human-robot interaction.

Publications

10 25 50