Orthogonal Wall Following and Obstacle Avoidance by an Autonomous Vehicle

摘要

The purpose of this thesis was to integrate a wall following motion mode for a rigid body autonomous vehicle. Yamabico, an autonomous vehicle located at the Naval Postgraduate School, was used as the test and evaluation platform. To implement the new motion mode, the vehicle was required to follow a straight wall with minor variations, navigate around corners, and avoid obstacles in its path while maintaining a specified offset distance from continuously connected wall segments. Sonar transmitter/receiver pairs were used to sense the environment and collect positional data for analysis. Modifications to pre-existing motion and sensor software libraries on board Yamabico were performed to achieve the motion goals. One of the major contributions from these modifications was the addition of a linear fitting algorithm using a decay factor. The algorithm produced quick response by the vehicle to changing conditions in its environment. The experimental results by Yamabico were successful with the algorithm developed by the author. The result of this thesis is that an autonomous vehicle can be given the capability to perform smooth and efficient motion adjustments to an environment composed of orthogonal wall segments and obstacles.