| The autonomous operating system of intelligent harvester has the nonlinear characteristics of time-varying load,complex actuator,obvious difference of working environments and so on.The research on the nonlinear motion control method of intelligent harvester can effectively improve the accuracy,robustness and stability of the harvester motion control.Aiming at the nonlinear problems in nonlinear motion control technology of intelligent harvester,such as input saturation,input time delay and uncertainty of model parameters,a combination model of nonlinear motion control system,an integrated navigation compensation method based on GNSS/INS,a neural network nonlinear motion control considering input saturation,and a prescribed performance control method considering input time delay are researched,realizing the autonomous operation of the intelligent harvester.The main contributions of the paper are as follows:(1)Aiming at the problem that a single model of harvester motion control system cannot take into account both nonlinear analysis and model lightweight,a combined model of nonlinear motion control system is proposed.A tracking error function is constructed,which can comprehensively measure the lateral position error and heading angle error of the harvester.The geometric model,kinematic model and one-dimensional dynamic model of the harvester control system are effectively combined.The tracking error source and input nonlinear characteristics of the combined model of the control system are analysed.The methods are researched to reduce the tracking error source and compensate the input nonlinear characteristics,which lays the foundation for the nonlinear motion control method with high precision,good robustness and strong stability.(2)Aiming at the problem that the navigation information accuracy of the GNSS/INS integrated navigation system of the harvester is reduced due to the lack of global navigation satellite system(GNSS)signal,a compensation method for the integrated navigation system of the harvester is proposed during GNSS signal outages.According to the missing time of GNSS signal or the characteristics of harvester motion state change,a prediction model based on discrete grey neural network and a prediction model based on long short-term memory network are selected.The position information of GNSS and de-noising information of the inertial measurement units by empirical mode decomposition are used as the inputs for the prediction models respectively to predict the position information of the intelligent harvester during GNSS signal outages.The predicted position information is fused with the inertial navigation system(INS)navigation information,which can effectively restrain the divergence of INS error and provide reliable current navigation information for nonlinear motion control system of harvester.The vehicle experimental results verify that compared with the traditional integrated navigation compensation methods,the navigation accuracy of the proposed compensation method is improved by more than 19% under different motion state changes and different GNSS signal missing time.(3)Aiming at the problem that input saturation reduces the stationarity of harvester motion control,a neural network nonlinear motion control method considering input saturation is proposed.A combined model of control system with uncertain model parameters is established.With the tracking error and tracking error changing rate as inputs,an adaptive neural network estimator is designed to estimate uncertainty of model parameters in real time.It can improve the accuracy of control system model parameters,reduce the tracking error of harvester,and avoid the phenomenon of input saturation.An input saturation compensation variable is designed by using the steering error of the harvester actuator,and the control input is corrected by the internal feedback of the intermediate controller,reducing the influence of input saturation on the control system.The simulation experimental results demonstrate the neural network nonlinear motion control method considering input saturation can effectively reduce the tracking error of the harvester control system,restrain the input saturation phenomenon of the control system,and improve the stationarity of the harvester motion control.The field experimental results show that compared with the backstepping-based motion control method,the standard deviation of the lateral position error of the proposed method is reduced from 2.49 cm to 1.68 cm,and the control accuracy is improved by 32.53%,which verifies the effectiveness of the neural network optimization-based motion control method.(4)Aiming at the problem of input oscillation of harvester control caused by input time delay,a nonlinear motion control method with prescribed performance considering input time delay is proposed.A combined model of control system is established with input time delay.The tracking error is constrained by the prescribed performance function,and the state variable is constructed satisfying the prescribed performance constraints.Based on the state variable,a nonlinear motion controller with prescribed performance considering input time delay is designed.An input time delay compensation variable is designed by using the control inputs in the delayed period and Lyapunov-Krasovskii function.The control system model related to the input delay time is optimized to a control system model independent of the input delay time to suppress the input oscillation of the harvester.The simulation experimental results illustrate the nonlinear motion control method with prescribed performance considering the input delay can keep the tracking error within the prescribed range,and ensure the dynamic and steady performance of the harvester motion control.The field experimental results show that the maximum absolute error,average absolute error and standard deviation of curve tracking experiments are 5.56 cm,2.31 cm and 2.31 cm respectively,which verifies the feasibility of the input delay compensation motion control method. |
PDF Full Text Request