A pathfinding algorithm for a myopic robot

Abstract

A pathfinding algorithm has been developed for an autonomous myopic roving vehicle.With the assumption that a path exists between two points, two propositions are proved.Proposition 1 is concerned with no obstacles and proposition 2 is concerned with at least one obstacle.Various conditions are established and terms defined.Three algorithms, necessary to direct the robot, are main, left scan, and right scan.The decisions for the navigation, control, and obstacle avoidance of the rover are based mainly on myopic (local) information of the terrain.The use of Gaussian density functions to simulate topography is considered to be novel.A FORTRAN IV computer program was written to implement the pathfinding algorithm, and the results of the computer-program runs were satisfactory.