Lawn Mowing System for Known Areas

This paper deals with and simulates the mowing of a lawn, which is a continuous planar area, by a mobile robot. The robot is given as input a bitmap of a known geometric area and derives an optimal covering path by implementing and improving the well known on-line Full Scan Spanning Tree Covering (STC) algorithm.The optimal path for continuous area coverage is calculated using a defining Depth First Search (DFS)spanning tree. We enhanced the STC algorithm by optimizing the number of U-turns along the path and allowing the mowing direction to be shifted. Furthermore, because our method uses different sensor information than used by the original algorithm, errors are reduced. The results of our improved algorithm are presented in an example wherein a 3D simulation program mimics the robotpsilas path and provides statistical calculations for testing optimality.