中文摘要：

工程车辆传动系统复杂多变，传动效率低下，难以用传统的方法解决．文中通过对工程车辆传动系统进行建模分析，推导出了工程车辆传动系统的动力学模型，在此基础上对四参数自动换挡策略进行了研究．针对常规BP神经网络收敛速度慢、存在“局部最小值”等的缺陷，文中采用了变步长法和动量梯度下降反向传播算法对常规BP神经网络进行改进，并将其应用于自动换挡控制．最后利用换挡控制试验数据对改进BP算法网络进行训练，并进行了仿真试验．结果表明，文中提出的换挡策略能提高传动系统效率，采用改进BP神经网络能够缩短网络训练时间，并能根据车辆运行状态确定最佳挡位．

英文摘要：

The transmission system of construction vehicle is difficult to describe via the traditional mathematical model due to its high complexity and low transmission efficiency. In order to solve these problems, a typical transmission system of construction vehicle is modeled and analyzed, and a dynamic model is established. Based on the model, a four-parameter automatic shift strategy is presented. Moreover, in order to overcome the low convergence rate and the local minimum of the conventional back propagation （BP） neural network, the method of changing step length and the reverse transmission algorithm of the momentum gradient reduction are adopted to improve the BP neural network for automatic shift control. Some test data of shift control are used to train the improved BP algorithm and a simulation is finally performed. The results indicate that the proposed shift strategy improves the transmission efficiency, and that the improved BP neural network effectively shortens the training time and determines the optimal shift according to the driving condition of vehicle.