工程机械模糊神经网络挡位控制试验研究

结合工程机械（车辆）的工作特点，提出了一种可应用于工程机械换挡决策的模糊神经网络（FNN）方法，改造后的模糊控制系统具有了知识自动获取功能，能更好地适应工况环境。仿真试验结果表明：该挡位决策方法可以根据操作工况环境实现正确的变速箱挡位决策，并依据工况环境的变化调整换挡策略，为智能化工程机械的开发打下了基础。

Experimental Study on Gear-shifting Control of Construction Machine Based on Fuzzy Neural Network

A method of shift decision based on the technology of fuzzy neural network (FNN) has been put forward in the paper, which can be applied to determine which gear in a transmission box should be shifted according to the on site operation of construction vehicle. The relevant simulative experiments on the structure of FNN and the shifting decision were carried out on a test bench which was equipped with a power train of ZL50 wheel loader. The experimental result shows that FNN shift decider can realize the gear selection of transmission box correctly with the shift strategy learned from the performance by itself. According to the operation knowledge acquired through the off line fuzzification, the normal fuzzy shifting controller is able to work well on the condition that the operation knowledge can recur or occur at least approximately. The great performance error of the fuzzy controller, however, will be produced if the operation environment changes, namely there isn't the corresponding operation knowledge in the knowledge base of the controller to deal with the change. We have transformed the structure of the fuzzy controller into that of neural network in order to make the controller possess the adaptability and the adjustability to the changing environment. The key technology of controller is none but that the shape and position of fuzzy sub function are modified according to the operation environment. Finally, the design of efficient hardware ensures the speediness, accuracy and reliability of controller.