This paper addresses the problem of planning the motions of a circular mobile robot moving amidst polygonal obstacles with uncertainty i n robot control and sensing. T h e robot is equipped with sensors which, if properly used, m a y provide information t o overcome the uncertainty accumulated during the motions. T h e position sensor is based o n dead-reckoning, the error t h e n results i n t o a cumulative uncertainty. A proximity sensor m a y be used t o localize the robot with respect to the obstacles of the environment. T h e robot can also gain information by entering inside landmark areas where the position error is assumed to be bounded. W e describe a planner which produces robust motion strategies composed of sensor-based m o t i o n comm a n d s which guarantee that, given a n explicit model of the error accumulated b y t h e m o t i o n commands, the robot canreach safely its goal with a n error lower than a pre-specified value. It is based o n a propagation of a numericalpotential and o n a geometric unalysis of the reachability of environmental features. T h i s planner exhibits a set of powerful Capabilities: while it allows t o consider motion primitives which accumulate uncertainty, it is able, whenever possible, t o navigate without relocalizing the robot when the task does n o t impose it, and also to m a k e a proper use of the sensors. Several examples r u n with the planner are presented a t the end of the paper.

This paper presents an operational framework to bridge the gap between planning with uncertainty and real-time sensor-based motion control. The environment being known, a planner produces a plan composed of free space and sensor-based motion commands. The representations of uncertainty and its evolution, environment landmarks, and actions generated at the planning level are discussed. Sensor-based actions and command de nitions for a nonholonomic mobile robot based on a Task-Potential eld approach are developed. These various elements are integrated in a system that actually generates the motions of the Hilare2 mobile robot.

No abstract