The Analysis And Implementation Of Maze Algorithm Based On IEEE Micro-mouse Contest

With the continuous development of computer technology and research of artificial intelligence, the study of intelligent robot attracts more and more attention. Actually micro-mouse is a tiny intelligent mobile robot which is controlled by microprocessor and can automatically navigate a maze. It can automatically memory the path and choose it, find export and achieve at destination finally in the maze. Micro-mouse combines mechanical design and manufacturing, motor technologies and applications, sensing and measurement technology, electronic engineering, control theory, artificial intelligence and other aspects of professional knowledge, performing the characteristic of multidisciplinary knowledge fusion.Thus the Micro-mouse is important for the integration, development and innovation of the above-mentioned disciplines.This topic researches into the robot path programming and related content based on the above background,including path planning of known map method which covers view of the law, the free space approach, topological approach, grid approach, neural network approach and so on, also including path planning of unknown map approach which comprises artificial potential field approach, optimal control approach, genetic algorithm and so on. Finally, a testbench is built to apply the algorithm provided in the thesis, and to test them.Combined with IEEE micro-mouse competition rules and the relevant theory of robot path planning, Micro-mouse competition is divided into three stages:maze searching stage, maze returning stage and maze sprinting stage. In view of sprint stage, this topic analyzes the commonly used method of robot path planning:A*algorith and Flood-Fill algorithm. In the searching stage, this topic provides a maze search algorithm based on probability distance. By generating the static probability distance. map and processing the map dynamically, it realizing efficient maze searching. Meanwhile, it changes the status that currently used algorithms lack theoretical basis. In the returning stage, the issue provides a return algorithm based on partitioning Flood-Fill, it can assist micro-mouse find better path with generating partition Flood map. Finally this topic implements design of micro-mouse maze algorithm and also verifies its feasibility.