dr. Jens Kober, assistant professor at the Delft Center for Systems and Control
op 09-03-2016 om 20:00 uur in PrinsenKwartier, Sint Agathaplein 4 te Delft
In robotics, concepts from decision making and control are applied on all levels. Examples range from low level control of the motors, for example when the controller sends currents to the motor in order to reach the desired robot position, to high level control where the robot needs to take a decision given the current situation perceived through its sensors, for example “Can I move forward or do I need to open the door first?”. The control problems become even more involved if the robot starts to interact with its environment, other robots, or humans.
Traditionally, decision making and control are programmed by hand. The resulting behaviour is usually very efficient but also extremely task specific, i.e., the robot cannot deal with slightly varying situations. The process is also extremely time-consuming and expensive, which constitutes a major bottleneck e.g., in industrial, household, care, and agricultural settings. As an alternative, we could imaging a robot that is able to acquire new skills autonomously, that is, via learning and artificial intelligence (AI).
In this debate we are going to explore the opportunities and challenges posed by autonomous, learning robots. (Where) would you like to see such intelligent robots in your daily life?
Deze activiteit valt onder het thema Techniek en samenleving.