基于遗传算法的工程车辆自动变速神经网络控制

Neural Network Control of Automatic Shift for Construction Vehicle Based on Genetic Algorithm

为了提高工程车辆传动系统效率,保持液力变矩器工作在高效区,提出了基于径向基函数神经网络的工程车辆自动变速控制方法,利用车辆传动试验台换挡控制试验的数据作为学习样本,采用遗传算法对径向基网络进行训练,并进行了验证性的仿真试验。仿真结果表明:该算法收敛速度快,能够满足工程车辆对换挡实时性的要求,可以根据车辆运行状态确定最佳挡位,并能够保证液力变矩器经常工作在高效区。

In order to improve transmission efficiency of construction vehicle, and to keep hydraulic torque converter working in high efficient area, authors presented a method of shiftdecision using a radial basis function （RBF） neural network and using genetic algorithm （GA） trains the RBF by optimized shift test data of construction vehicle, and did simulation experiment to validate it. The simulation results show that GA is effective for intelligent control and can judge the optimal gear shift position during vehicle running, and keep hydraulic torque converter always working in high efficient area.