Searching for an optimal path through Pasadena

Abstract

A functioning software package for a planning system that uses multiresolutional heuristic search techniques to find the optimal path over terrain with varied traversability is described. Unlike previous systems, the one considered uses the same algorithm of decision-making through all levels of planning/control operation top-down. It employs solutions previously developed for the autonomous robot Drexel-buggy and explores a number of tools which reduce the number of computations, help to overcome the trap of multiple solutions, determine streets, etc. It is demonstrated that time-efficient planning is possible based upon sophisticated input information incorporating expert knowledge. The example map discussed is of Pasadena, California.<>