基于系统响应的履带车辆路面识别方法

In order to improve tracked vehicles' adaptability to different types of road surfaces, a dynamic simulation model of tracked vehicles and various types of road was established based on the multi-body simulation platform. The time-domain dynamic response signals of vehicle centroid were collected through the driving simulation of tracked vehicles and road models, and the signals were decomposed by wavelet transformation. Distance evaluation technique was used to extract sensitive feature vectors. A road identification method based on the above sensitive feature vectors was proposed by using BP neural network. In order to verify the validity of the method, a test system based on small tracked vehicle models was built.The vehicle model drived on the actual road to collect the time-domain dynamic response signals of tracked vehicles' body centroid, load wheels and track-terrain interaction. The results showed that the identification precision of the method is 99%. This method has a high identiciation ability for road types.