Research On The Position Control Of Loader Working Device Based On Machine Learning

With the rapid development of loader technology in recent years,requirements for new function and higher efficiency of loaders have been put forward continuously.In school enterprise cooperation projects "FW50GL loader R & D",we use the main valve of electromagnetic proportional control and some sensors to realize the lifting function of boom fixed height,which can improve the working efficiency of loader.But because of the inertia of the system and the lag of response,there is a deviation when using this function,which results in extra energy loss,and the loss of energy cannot be underestimated in the long run.The way of controlling the deviation effectively and realizing the position control of the loader working device is studied in this paper.The research of this paper is based on a large number of experiments: Chapter 2 of this paper analyzes the experimental curve,studies the basic cause of the deviation,deals with experimental data and find out the influence of different factors on the deviation;Chapter 3 of this paper studies the new control strategy based on the current control method,puts forward the mathematical expression of the control strategy,and verifies the validity of the control strategy through experiments.In Chapter 4,on the basis of control strategy,the basic control framework based on machine learning is proposed.In order to reduce the influence of the blind area on the control results,the prediction of the system state is proposed.The root mean square error and the maximum and minimum error of the prediction results of different prediction schemes are studied and the best prediction scheme is determined.Based on the basic control framework,the initial learning and optimal control learning of deviation control module are studied,and a feasible and complete deviation control model based on machine learning is proposed.In Chapter 5,AMESim is used to build the boom model of loader working device which is consistent with the actual situation as much as possible.MATLAB is used to realize the control strategy studied in this paper.The proposed control model based on machine learning is verified by combining different control strategies.According to the different expression forms and prediction function of each control strategy,three verification schemes are proposed,The verification results show that the control model based on machine learning can realize the effective control of deviation,and each different mathematical expression in the control strategy directly affects the control results.When the state quantity of the system changes rapidly,the control strategy with the function of system state quantity prediction can achieve the control of deviation more effectively.