Path planning and dynamic objects detection
István Szőke, Gheorghe Lazea, Levente Tamás, Mircea Popa, András Majdik · International Conference on Advanced Robotics · 2009
This paper describes the path planning for the mobile robots, based on the Markov Decision Problems and the detection of dynamic objects using stereo-vision. The presented algorithms are developed for resolving problems with partially observable states. The algorithm is applied in an office environment and tested with a skid-steered robot. The created map combines two mapping theory, the topological respectively the metric method. The main goal of the robot is to reach from the home point to the door of the indoor environment using algorithms which are based on Markovian decisions. In case if a dynamic object is detected the agent must replan the previous rout.