基于BP神经网络优化算法的工程车辆挡位判断的训练及仿真

为了克服工程车辆重载作业时，传动系统效率大幅下降的问题，一般采用自动换挡的办法，以实现对负载的适应，提高传递效率。所以选择一种合适的换挡规律是非常重要的。由于车辆的状态与最佳挡位之间存在非线性，若其关系以相应的数据的形式给出，可直接用所获得的数据对人工神经网络进行离线训练，讨论了建立在人工神经网络基础上的工程车辆挡位判断方法，以三参数换挡规律为例对网络进行了训练，并给出了MATLAB仿真结果。

TRAINING AND SIMULATION ON GEAR POSITION DECISION FOR VEHI-CLE BASED ON OPTIMAL ALGORITHM OF BP NETWORK

In order to solve the problem that efficiency of the powertrain of the vehicle is very low, especially under the heavy load, the method of automatic shifting is realized to suit for the load, and improve the efficiency. It is important to find a suitable shifting strategy. Since the nonlinear function between the status of vehicle and optimal gear position, and corresponding data is given, neural network can be trained away from line with the acquisition data. The method of gear position decision for the vehicle based on optimal algorithm of BP neural network is studied. With the example of three parameter shifting strategy, the neural network is trained and the result of the simulation using MATLAB is given.