A Potential Maze Solving Algorithm for a Micromouse Robot

Discretely assigned potential levels can be effectively used in making autonomous route decisions for a mobile robot to reach a goal. This paper demonstrates methods of assigning and manipulating these artificial potentials to provide locally optimized path choices while maintaining the integrity of the potentials. The basic algorithm is improved by retaining information of the number of decisions that have been made. Results from implementation and simulation of the algorithm for a Micromouse maze-solving robot are presented. Consideration is given to implementation using a limited power microprocessor. I. Introduction This paper provides experimental and simulated results of a simple, yet effective, algorithm to solve maze-type problems based upon discretely assigned virtual potentials in a restricted environment. The goal is to provide locally optimized routing choices based upon odometry and limited sensory data. The starting position of the automaton is known, as is the goal loca..