Automation and technology, of course, have already made a huge impact on the way we make things and live our lives. Robots have played a part in this, principally in factory automation. But factory robots are simple puppets, continually repeating the actions scripted for them by human controllers.
The coming years will see a revolution in the use of autonomous robots – which have some freedom in how they achieve their human-defined goals. But to do this, robots will need to be able to perceive their environments (seeing objects and people), work out what to do next (which route to take through a building or where to search for people), and then execute the resulting actions (driving through the building, avoiding people, picking up objects).
In the School of Computer Science at the University of Birmingham, we are working on a range of Artificial Intelligence (AI) and robotics technologies to give autonomous robots the ability to work with and for humans, in both industry and the home.
We’ve developed systems to allow robots to identify objects in images by breaking them down into small fragments; robot manipulation systems that can learn to grasp new objects after being shown how to pick up a similar object; curious robots that can explore buildings to create maps and ask questions about the things they don’t understand; and systems for detecting faults in autonomous submarines. Underlying these examples are recent developments in AI like reasoning with uncertain information and planning actions (particularly those that involve gathering more information) in dynamic, changing environments.
Existing robots allow tasks to be carried out repeatedly (and indefinitely) to a predictable standard. Autonomous, intelligent robots will allow this approach to be applied to a wider variety of tasks, from warehouses to care homes. Examples on the horizon include flexible assembly robots that can be taught to use new tools, or to assemble new prototype products; logistics robots that can safely choose the best routes through a busy warehouse; and cleaning robots that can learn which areas get dirty and when. For businesses, this means that more tasks can be automated (and thus predictably and accurately performed). And by using learning software, robots could be retrained to work on new problems as required.
As AI and robotics technology advances, society will clearly benefit. Robotics has long been promoted as a solution for dull, dirty and dangerous jobs, and removing humans from such situations will improve the lives of many. But there is also the chance that robots will offer technological solutions to current open problems.
Two examples we are actively exploring are nuclear decommissioning and caring for an ever-growing ageing population. For the nuclear industry, we are researching autonomous robots, which can be guided by a human operator to help clean up highly radioactive facilities in which humans can only spend limited periods. For elderly care, we are developing portering robots, which can learn to assist limited nursing staff in care homes. The robots will understand what is happening (whether it’s time for lunch, medicine or bingo) and support patients by allowing overworked staff to perform more caring duties. This work is part of a multi-million pound European project called STRANDS.
Beyond the scientific challenges our field is actively pursuing, at least two other issues must be addressed for these results to be realised. The first is the issue of legislation. As robots become increasingly autonomous, and are able to learn and act on their own, laws must be created to determine who is responsible for their actions should mistakes and accidents happen. What happens when a driverless car accidentally runs someone over, for example?
The second problem is the acceptance of robots in our world. While we are all comfortable with robots building cars, only by understanding their benefits and limitations (and through sensitive deployment) will we have a similar level of comfort with them assisting the elderly, or working in homes and workplaces. But despite the remaining challenges, the potential advantages are extremely exciting, and we look forward to helping to build this future alongside scientists and users alike.
Nick Hawes is a senior lecturer in intelligent robotics at The University of Birmingham. www.nickhaw.es. He will be speaking at the RE.WORK summit, held in London this September – bringing together science, technology and entrepreneurism to solve global challenges. For more information or to book tickets go to www.re-work.co