FPGA-based autonomous parking of a car-like robot using fuzzy logic control

Abstract

Cutting edge technology is continuously being developed and used in vehicles for both safety and comfort. Driver Assistance Systems are using the latest technology to make driving both easier and safer. These include systems such as autonomous parking, which aims to simplify the task of parking and is supremely useful in cases such as parking in tight spaces or when visibility is an issue. This project focuses on this particular problem. The system developed, works in such a way that it scans for a valid parking spot and upon detection, performs the necessary manoeuvres to park the car within the spot. This is done for both parallel and perpendicular parking. Fuzzy Logic Control provides a way of processing imprecise information and obtaining an output based on linguistic variables and rules. Consequently, complex systems can be implemented in a very efficient manner. As a result, Fuzzy Logic Control was used to solve the autonomous parking problem. The system was primarily prototyped in a custom-built simulation environment and subsequently implemented on a Field Programmable Gate Array (FPGA). The primary motive for choosing an FPGA as the target device, was the good compromise that FPGAs provide between special purpose Application-Specific Integrated Circuits (ASICs) and general purpose processors This hardware implementation was then tested on a car-like mobile robot (CLMR) specifically designed for this purpose. The final product is an efficient, cost effective solution that fully achieves the initial goals. Furthermore, it is scalable in a way that it would be very appropriate for implementation on an actual vehicle, given certain adaptations.