The evaluation of strategies in motion planning with uncertainty

Studies motion strategies for a mobile robot moving with uncertainty using various sensors. A colored graph represents a sensory space where sensor measurements define the node's color and the authors traverse the graph with probabilistic transitions. The authors first show how to construct such graphs for geometrical scenes and various sensors, and then study the complexity of evaluating the robustness of deterministic and probabilistic strategies. The authors show upper and lower bounds for Markov, time-dependent and history-dependent strategies. In the case of finite memory, it is shown that randomized strategies are better than deterministic ones.<
>