Development of artificial neural network based autonomous parking system

Driving assistance systems provide life safety for the driver and passengers, as well as create a comfort zone in vehicles. In this study, the autonomous parking system, which is one of the most common driving support systems, was developed. Theoretical analyzes of the autonomous parking system were carried out and then its performance was evaluated by applying it on a 1/10 scale model car. Autonomous parking software consists of vehicle control card software and control computer software. An artificial neural network model was used in control algorithm to determine parking path planning. The artificial neural network was trained according to vehicle kinematic equations and different parking starting positions. The trained network was integrated into the control computer software. A test parkour was set according to vehicle geometric boundaries to test the developed system. A number of parking tests were carried out on test parkour and the system performance was determined. The test results reveal the success of proposed autonomous parking system.