A robot swarm the size of a city will coordinate itself to move goods from local producers to local consumers, then collect and redistribute the waste. Today’s logistics chains rely on centralised warehouses and long-distance transport, which generate emissions and waste. This project tackles the engineering gap between small lab swarms and the large, trustworthy swarms needed for real urban logistics. The researchers must solve two problems: how to give each robot enough intelligence to react locally while still acting as part of a collective, and how to let humans monitor and override the swarm in real time without losing its efficiency. If successful, this work could reshape urban supply chains. Neighbourhoods might produce, distribute, and reuse goods using fleets of cheap, capable robots that require no special infrastructure. The team will test their designs in a physical testbed, a digital twin of a city, and a real Bristol neighbourhood, working with social scientists and the public. They also plan to spin out a startup. The project will produce best-practice guidelines for deploying smart machines in cities, shared through a new “real robotics” network.
View original technical description
Local production, distribution, and reuse of goods facilitated using robot swarms will enable a more sustainable future through the lowering of transport and waste. By powering local communities to be part of the life-cycle and local economy, we can use technology as an equaliser. This vision requires trusted swarms of robots useable out-of-the-box, with little setup or infrastructure. These swarms will become ubiquitous, cooperating with each other and humans. We focus on the core research questions in swarm engineering required to achieve this vision: 1) Beyond minimal robots: How can we design distributed cognition and action for highly capable next generation swarms? Advances in individual robot hardware mean we can now mass-produce low-cost highly capable robots with significant sensing, computation, communication, and mobility. We will develop robot cognition specifically for distributed systems that will allow these robots to react to their local environment in a capable manner, thereby increasing the breadth of swarm behaviours that may emerge and moving them closer to real-world applications. 2) Trustworthy design: How can we enable real-time human monitoring and control of trustworthy swarms? Real-world swarms will need to be useful, and trusted. New metrics for swarm operations, including key performance indicators (KPIs), swarm performance indicators (SPIs), and trust performance indicators (TPIs), will enable the real-time automatic design of swarm strategies for user requirements using Quality-Diversity optimisation of human-readable controllers. We will also design ways to monitor swarm operations, and control their behaviour in real time. Discoveries will be validated using cyber-physical infrastructure for human-centric stories using digital twins, a swarm logistics testbed, and living labs. Living labs are real-world environments where technology can be trialled, in our case the Bristol Robotics Laboratory, and Paintworks neighbourhood in Bristol. We will work with the ESRC Centre for Sociodigital Futures to co-design use cases for city-scale logistics with users, and will engage with policy makers and the public through a dedicated programme of activities. Through this project we also aim to explore translation opportunities including founding a startup. The result of this journey will be the discovery of best practice to responsibly deploy smart machines in cities. This best practice will be shared with the research community through workshops and a new 'real robotics' network, promoting a positive change in the community towards deploying robots in reality (plus component of this fellowship).
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know