British researchers are developing robots capable to work with people and act intelligently in real-world environments.
Funded through EU’s Seventh Framework programme, the STRANDS project has been awarded £7.2m to design and build mobile robots that could serve as indefatigable assistants to security guards or care home nurses.
Overseen by the University of Birmingham, the STRANDS project has brought together computer scientists from the University of Lincoln, UK, with a security company G4S and scientists from the Academy of Ageing Research in Austria.
The University of Lincoln team has now started developing software that would enable the robots to process real-life experience the machines will encounter.
“The idea is to create service robots that will work with people and learn from long-term experiences. In a security scenario a robot will be required to perform regular patrols and continually inspect its surroundings for variations from its normal experiences,” explained Tom Ducket, director of the Lincoln Centre for Autonomous Systems Research.
The robot’s recognition systems will be based on so called 4D maps – a technology that uses 3D mapping over an extensive period of time. The robot will learn what the normal settings of its working area are, and should be able to detect deviations suggesting something might have gone wrong.
“Certain changes such as finding a person in a restricted area may indicate a security violation or a burglary. In a care home a robot will be required to act as an assistant for elderly patients, fetching and carrying things while also being alert to incidents such as people falling over,” Ducket said.
The unique feature of the concept is that it should enable the robots to refine their perception of changes in the environment by learning from their own experience. The team believes such technology would have many other possible applications.
“The main idea is to deploy robots that run for a long time so they have the chance to develop a common-sense attitude on how the world should be and be able to spot the deviations,” said Marc Hanheide, of the team members.
“The robots are curious to learn about the environment - they will see if something has changed and whether that’s a one-off or a regular occurrence. Our robots will be active for long periods in dynamic and changing environments,” he said.
Apart from mapping the environment, the robots will be equipped with advanced person detection, tracking and activity recognition systems. The combination of these technologies will hopefully make the machines capable to interact properly with human users.
“Currently industry robots can run for 24 hours a day and are incredibly reliable in well-controlled environments, but they don’t use long-term experience to adjust or improve in any way,” Hanheide said.
“Cognitive robotics systems can learn and adapt, but most are used for just one experiment. We want to build a bridge between the two by creating robots that can run for long periods of time and also make use of life-long learning capabilities to adapt to the needs of different users.”
The team promises to showcase the results at museums, public demonstrations and trade shows, once the development is concluded.