Semantic Robots – a given in the near future?
11 Sep 2015
Mobile Robotics and Olfaction Lab & Cognitive Robotic Systems, have joined forces to create a new research profile named Semantic Robots.
Robots that understand and respond to its surroundings, does it sound like science fiction? It isn’t. In fact at Örebro University two research teams, Mobile Robotics and Olfaction Lab & Cognitive Robotic Systems, have joined forces to create a new research profile named Semantic Robots. Working close with industry the goal is to develop techniques that enable semantic robots to solve problems on their own with the help of sensor systems.
Dimiter Driankov, professor of computer science and director of AASS, Centre for Applied Autonomous Sensor Systems, explains that a semantic robot should be able to create a task-related model of its environment with different human-like levels of abstraction, which are all based on data acquired by sensors. It must also be able to use the model and its different abstractions of the environment to autonomously plan and carry out the task it have been assigned.
He gives the example of a truck navigating and handling goods in a factory environment. The truck receives a vast amount of data about the incline of the floor, various obstacles to be negotiated, suitable places for loading and unloading goods and so on. It also needs to know that in order to go to another floor of the building, it needs to take the cargo lift and it needs to react safely if humans or other things unexpectedly get in its way.
– In short the robot must describe it’s surroundings in terms that makes sense for humans, like robot programmers and operators, and then produce a plan of actions to complete its mission.
The program is granted for a six-year period and is largely funded by the Swedish Knowledge Foundation (36 million SEK) in combination with industry contributors (15 million SEK) such as Optronic, Saab Dynamics, Volvo CE, Atlas Copco Rock Drills, Husqvarna, Kollmorgen, Cnet and Hothouse.
– We provide the research questions and discuss with our industry partners in terms of how they can profit from the results, that is, if our ideas can solve their problems, says Dimiter Driankov.
As one of the contributors Optronic is following the research with great curiosity, with a particular interest regarding what may be identified as the future needs and demands in the field of optical sensor systems.