A new study suggests that autonomous delivery robots that run smoothly on the streets largely owe their successful navigation to humans, who subtly modify their behaviour to avoid obstacles posed by the robots. The researchers argue that this aspect of 'invisible human work' should be factored into designing robot routes.

The study was conducted by a combined team from the University of Nottingham's School of Computer Science and Horizon Digital Economy Research, Linkoping University in Sweden, and the University of York. Researchers observed and analysed humans interacting with delivery robots in two UK cities. Their findings revealed important behavioural patterns associated with how humans adjust and provide a way for these robots on the streets.

Autonomous delivery robots are being tested in various UK cities, including Milton Keynes, Manchester, Cambridge, and Northampton. Robots powered by artificial intelligence and GPS navigation facilitate short-distance deliveries. Collaborating with diverse companies, these trials aim to revolutionise delivery systems.

For the study, the researchers followed and filmed several robots in Milton Keynes and Northampton over several days and studied their encounters along set routes. The robots navigated myriad urban obstacles such as bins, pets, parked cars, and human interaction.

According to the findings, the robot is an obstacle needing continuous and subtle human accommodation. For instance, people adjust their walking pace, relocate to assist the robot's passage or signal the robot to keep moving when it slows down. These instances of human assistance show people's important role in facilitating the successful operation of robots in public spaces.

The study, co-authored by Dr Stuart Reeves, Hannah Pelikan, and Marina Cantarutti, was highlighted at the 2024 ACM/IEEE International Conference in Human-Robot Interaction, where it was named the best paper. Dr Reeves iterated that the study's findings should be considered when designing routes and programming robots, explaining, "What we showed is that for a robot to navigate a route successfully, it relies on the accommodation of people for it to be successful and not become an obstacle itself."

The research, part of a Responsible AI UK-funded international partnerships project, seeks to investigate robot autonomy in public spaces. The team aims to understand the unique characteristics of public spaces and how they affect robot deployment. Dr Reeves hopes that these insights will allow councils and designers to shape public spaces and robot technology design better to accommodate such technologies' growing use.

The study conducted by a combined team from the University of Nottingham, Linkoping University, and the University of York sheds light on the significant role humans play in facilitating the successful navigation of autonomous delivery robots in public spaces. The findings emphasise the importance of considering human behaviour and interaction when designing robot routes and programming.

As autonomous delivery robots become increasingly integrated into urban environments, understanding these dynamics is crucial for ensuring efficient and safe operations. The research highlights the collaborative relationship between humans and robots and provides valuable insights for future advancements in robot technology and urban planning.