Evaluation of the Autonomous Parking System Based on BP Neural Network

The autonomous parking system is an intelligent system which can park a car into a parking spot. Recently, more and more studies pay attention on this system. However, research on evaluation of autonomous parking system in these studies is not enough. Because most of evaluation methods of intelligent vehicle system are based on task completion degree, the evaluation of autonomous parking system cannot show its real intelligent level. In order to ensure the feasibility and accuracy of evaluation of autonomous parking system, the paper proposes an effective evaluation method based on BP neural network. The evaluation method is established based on autonomous parking system evaluation index system, using BP neural network evaluation model. Using samples collected from the autonomous parking system in intelligent vehicles, the model is trained by MATLAB software and the evaluation model is tested by samples. The result shows that the simulation values are very close to the expected ones, which means the feasibility and accuracy of the evaluation model.