Autonomous vehicle docking using fuzzy logic.

Abstract

Autonomous control of robot vehicles has been studied recently by AI researchers with simulations and implementations emerging using both Neural Networks and Fuzzy Systems. The application is important primarily for autonomous control of planetary surface rovers and deep space vehicles where telepresence control (remote human operator) is precluded because of the long signal transit time. The problem of vehicle docking is specifically addressed in this thesis and the results of experiments on a small model truck and trailer using a mechanical based position sensing systems is reported. Fuzzy logic controls both forward and backward motion of the vehicle to successfully dock the trailer. Additionally, it is shown how the position sensing system can be improved using a small laser-based bar code engine. Future research including use of the system on other mobile robot platforms is discussed and we describe other research planned for the platform. In this thesis the author shows that the use of fuzzy logic for vehicle control simplifies the analysis of motion required to produce proper docking. In particular, the discontinuity of control produced by switching from forward to reverse is readily handled by fuzzy logic.