At the AI Summit London, Nick Hawes, director of the Oxford Robotics Institute, highlighted the potential for autonomous robots to become crucial data sources for industrial companies. These robots can provide consistent, repeatable measurements that surpass human capabilities.
Hawes discussed how his team has utilized autonomous robots, such as Boston Dynamics' robot dog Spot, to collect data at challenging sites, including a former nuclear fusion reactor. Over 35 days, Spot gathered radiation data with minimal human intervention, showcasing the efficiency of robotic autonomy in industrial settings.
The Oxford Robotics Institute comprises seven research groups working on various aspects of robotics, from dexterous control manipulators to fully AI-driven robots in industrial environments. Hawes noted that 80% of the Institute's work is related to AI, both in fundamental research and practical applications for clients.
Early research from the Institute contributed to the development of autonomous vehicles through Oxbotica, now known as Oxa. Currently, the focus is on deploying robots in industrial environments, enabling them to reach places inaccessible to humans.
Hawes pointed out that while industrial companies are exploring autonomy, they often prefer a human-in-the-loop approach. This method allows operators to collaborate with robots and take control when necessary, balancing the benefits of autonomy with the need for human oversight.
Transitioning to full autonomy will be gradual, as systems need to improve in handling unpredictability. "If the world was fixed and 100% predictable, we’d write a Python script and be done," Hawes explained. "The reason you need AI on a robot is because it needs to respond to changes."
Hawes described several deployments, including using quadrupedal robots from Boston Dynamics to patrol industrial sites. For instance, his team equipped Spot robots with lidar and advanced 3D mapping technologies, enabling them to navigate industrial sites autonomously and carry mission-specific payloads, such as devices for monitoring radiation levels.
In one notable deployment, Spot operated autonomously for 35 days at the U.K.'s former fusion reactor site. It gathered alpha radiation data with minimal human involvement, programmed to return to its charging station when its battery was low. "The robot had a little script that said, here are the six places I want you to look at and this is what your battery looks like," Hawes said. "The robot was able to then plan and optimize for that information to get up every day, walk around the site, gather the information, and go back home."
Autonomous robots operating over long periods can generate multiple maps of their environments, providing valuable data for operators. These maps can be used to create digital twins—virtual representations of industrial environments—that can aid in planning and emergency responses. Hawes demonstrated how data from the fusion reactor site could create a virtual environment for emergency services to understand before entering in case of a disaster.
Describing artificial general intelligence (AGI) as “nonsense,” Hawes emphasized that his team's focus is on practical autonomy—robots performing tasks or measuring data reliably and repeatedly. Instead of pursuing AGI, they aim to develop robots that can consistently deliver actionable data over long periods, enhancing reliability and efficiency in industrial operations.
“A robot is something that can deliver actionable data over long periods of time,” Hawes concluded. “Humans struggle because they get bored and point the camera at the wrong place or forget to take an image whereas a robot gives you a bit more reliability.”