A fuzzy algorithm for path selection in autonomous vehicle navigation

This article describes the application of fuzzy logic to the path selection algorithm of an autonomous roving vehicle. A simulated laser-detector configuration provides the imprecise terrain information which constitutes the input to the path selection algorithm. The algorithm determines the most suitable path to a given target point. The vehicle's motion is then simulated, more terrain information is obtained and a new decision is made. At every decision point, the algorithm determines the safety of alternative paths and their direction with respect to the target point, and finally selects a path by performing some trade-offs between path safety and heading toward the target. A graphic display of the terrain and the vehicle's path summarizes the result of the algorithm.