Research Of Robot Control Strategy For Skewer Tomatoes Picking Manipulator

Skewer tomatoes are deeply loved by consumers in the fruit and vegetable market because of their rich nutritional value and convenient consumption.However,there is a situation that the skewer tomatoes bear fruit at the height of the plant,which leads to low efficiency in picking skewer tomatoes.Research on picking robots can effectively improve production efficiency and reduce production costs,which is of great significance to the realization of agricultural machinery automation.n this paper,the control system is designed and researched on the skewer tomato picking robot,which shows that the RBF neural network control optimized by the particle swarm algorithm has excellent trajectory tracking performance and excellent stability.The main contents are as follows:(1)Establish a model of a skewer tomato picking robot.By collecting data through field investigation,a four-degree-of-freedom tandem joint picking robot is designed,and completed three-dimensional modeling by using Solidoworks software.The D-H parameter method of the threedimensional model is used to establish a mathematical model in Matlab,and the kinematics and dynamics models are verified by simulation.The picking trajectory is planed and simulated by the fifth-order polynomial method.to provide input and controlled objects for the control system.there provide input and controlled objects for the control system.(2)Establish RBF neural network control system.Aiming at the uncertainty of the picking robot model in the moving process,an adaptive control method based on the gradient descent method of the RBF neural network is adopted by using the advantages of the RBF neural network’s nonlinear approximation.It Approximate and compensate the robot system uncertainty through the RBF neural network,and use the gradient descent method to realize the online optimization of the RBF neural network parameters,and design an adaptive law to adjust the control for these problems,and construct the Lyapunov function pair The cont rol system conducts stability analysis.The simulation results in Matlab show that the adaptive control of the adopted RBF neural network has a good effect on the uncertainty compensation of the skewer tomato picking robot model,and has a higher trajectory tracking accuracy and a faster response speed to the expected output.(3)Optimize the RBF neural network control system.In the adaptive control output of the RBF neural network based on the gradient descent method,the trajectory tracking error at the initial moment is large and the speed fluctuation is obvious.The analysis is caused by the shortcomings of the gradient descent method.The particle swarm algorithm is introduced to replace the gradient descent method to train the network and optimize the adaptive control system of the RBF neural network.The simulation results in Matlab show that the particle swarm optimization optimization of the adaptive control of the RBF neural network overcomes the shortcomings of the large trajectory tracking error and obvious speed fluctuation of the RBF at the initial moment.