Motion in an unknown environment

This chapter considers the methods of probabilistic motion planning in the unknown environment widely known as probabilistic robotics. It explains Bayesian motion planning and considers the algorithms that resolve the uncertainties using informational heuristics. For the known model of the robot’s motion, the motion planning is based on the predicted positions of the robot and observation results, and for unknown motion model the motion planning is based on general Bayesian updates of the robot’s beliefs or of the occupancy grid. In the case of offline potion planning, an optimal or near optimal path of the robot is usually obtained following general optimization techniques, for example by value iteration or policy iteration algorithms, and for online motion planning some heuristic methods are applied. The chapter also considers examples of online motion planning and mapping using informational heuristics.