Research on Road Unevenness Recognition Method Based on Off-road Vehicle Driving Characteristic

By using road surface recognition technology to identify the level of road surface unevenness when the vehicle is in motion and thus adjusting the active suspension control parameters, vehicle smoothness can be improved. This paper developed a pavement unevenness recognition model based on decision tree algorithm and random forest algorithm, and a soft soil stochastic uneven pavement and off-road vehicle dynamics model based on off-road vehicle driving characteristics. The vehicle is simulated to drive on the soft soil pavement, and the vertical acceleration data of the wheels are collected as the data set of the pavement recognition model. The recognition performance of the two pavement recognition models is compared. The results show that the ensemble learning algorithm performs better in generalization and achieves an accuracy of 98% for recognizing different road surfaces. The proposed pavement identification method is essential for improving the smoothness of off-road vehicles driving on the soft soil surface.