Acquisition of obstacle motion patterns to improve mobile robot motion planning

Mobile robots for advanced applications have to act in environments which c ontain moving obstacles (humans). Even though the motions of such obstacles are not precisely predictable, usually they are not completely random; lon g-term observation of obstacle behavior may thus yield valuable knowledge a bout prevailing motion patterns. By incorporating such knowledge as statist ical data, a new approach called statistical motion planning yields robot m otions which are better adapted to the dynamic environment. To put these id eas into practice, an experimental system has been developed. Cameras obser ve the workspace in order to detect obstacle motion. Statistical data is de rived and represented as a set of stochastic trajectories. This data can be directly employed in order to calculate collision probability, i.e. the pr obability of encountering an obstacle during the robot's motion. Further as pects of motion planning are addressed: path planning which minimizes colli sion probability, estimation of expected time to reach the goal and reactiv e planning.