Abstract: The SLAM map building method generally relies on human-computer interaction platform in environmental modeling, which has high labor cost and poor independence. This paper proposes an unknown environment map construction algorithm based on dynamic matrix. The algorithm builds a two-dimensional raster map and uses A* algorithm for backtracking in completely unknown environment, so as to realize the full coverage collection of map information independently. Traditional local coverage path planning algorithm has high repetition rate and low operation efficiency. To solve these problems, the presented algorithm deals with the closed area immediately, and combines the edge tracking and boustrophedon coverage methods in the path planning of sub-region. The algorithm was designed using MATLAB, and further verified by the Webots robot simulation platform. Simulation results show, compared with the traditional local coverage algorithm, that the algorithm has significant improvements in the reduced number of sub regions, the total length of backtracking path and the path repetition rate..

Key words: raster map modeling, path planning, dynamic matrix, A* algorithm, edge tracking