Rule BaseInference ?

This paper proposes a method for the design of a dynamic path planner which utilizes fuzzy logic for navigation of mobile robots in uncertain environments. The method consists in identifying a dynamical system, with a fuzzy logic controller structure, from the examples generated by an \expert". The generator of examples will be a search algorithm on a known map of the working environment, that is, a model-based path planning. The dynamic path planner maps the ultrasonic sensors readings and the odometer measures to the robot control commands {its heading velocity and steering angle{. The correct mapping is found by identifying a set of fuzzy rules from desired input-output data pairs. It is wanted with this idea to use the optimality qualities that the geometrical path planners have, integrating them in an sensor-based architecture, better adapted for the real-time speciications. The eeectiveness of the proposed method is veriied by a series of simulations.