Parallel parking a car-like robot using fuzzy gain scheduling

Abstract

A fuzzy gain scheduling controller is proposed to parallel park a car-like robot. At first a fuzzy sliding mode controller (FSMC) embedded by driving experience is developed to locally track the most typical path for the parallel parking. In order to extend the controlled region, several typical paths are formed and can be pieced together to constitute a complete parking path in a large region. Then, a fuzzy gain scheduler with knowledge-based structure is proposed to decide the best parking path and to generate the proper control gains.