Abstract
Today, specialized vehicle are widely used in many areas as construction, agriculture and forestry, military as well as for humanitarian missions and in somewhere, there are no roads for vehicle. The mobility of vehicle has great effect from wheel-terrain interaction. The analysis and identification about wheel-terrain interaction are important for capable vehicle. The soil parameters have a leading effect on the contact between the elastic tire and the terrain. Normally, the soil parameters are measured in a very time consuming way with expensive test equipment. The goal of this task is to define automatic the most fundamentals soil parameters by using dynamic values of the based on Artificial Neuronal Networks (NNs) is selected in this task to estimate soil parameters with measured values of the vehicle. The advantage of the method is the automatic identification of the soil parameters and the implementation in a real-time system. In this research, the identification method is based on a back-propagation network with three layers. The research results had been implemented to a virtual vehicle. This vehicle was used at the Institute for Automotive and Power-train Engineering.
Keywords: soil parameters identification, neuronal networks, simulation vehicle, multi-axle vehicle