改进LM算法的液力机械传动车辆神经网络自动换档系统

The developed LM algorithm of hydraulic mechanical transmission vehicle neural network automatic shiftting controling system

通过神经网络对液力机械传动车辆"节能换档规律"进行系统建模和控制,采用改进的Levenberg-Marquardt算法(LM算法)对神经网络的训练过程进行了优化,大大提高了网络的收敛速度,减小了网络收敛于局部最小点的可能性,可以实现对自动变速器的档位判断进行模拟和预测,并给出了MAT-LAB仿真结果。

Set up system model and controling for hydraulic mechanical transmission vehicle 'energy conservation shift schedule' was based on neural networks.Adopting developed energy conservation shift for improving the neural networks training process can develop the convergence rapid of the net and reduce the possibility of net convergence to be local smallest point.It can simulate and predict shift judgement of the automatic transmission.The MATLAB simulation result was available.