Research On Non-destructive And Active Grasp Technology For End-effector Of Apple Picking Robot

During the process of fruit and vegetable production,the picking is the most time-consuming part which takes up almost about 50%~70% of the total production cost.Making research on the fruit picking robot can not only solve the problem shortage of labor caused by population aging and social urbanization but also can promote the progress of agricultural science and technology in China which is very vital to realize the Chinese agricultural modernization.As the skin of fruit is thin and easy to damage and grow state is different from each other,leading a high damage rate of fruit grasp,so it is important to study the grasp planning and grasp force control strategy of the end-effector to realize non-destructive grasp.In this paper,we study the non-destructive grasp technology of picking robot from two aspects,one is steady grasp planning and other is active grasp force control.The main research contents are as follows:(1)According to the structural characteristics and mechanical properties of apple,the two-finger apple picking robot end-effector has been designed including the method of cutting the fruit stems.Then the sensory perceptual system was designed to enhance the perceived ability to environment.Finally the math model of end-effector was established.(2)The conditions of stable grasp of two fingers were analyzed based on the theory of force closure,and two fingers grasp strategies were established.At the same time,according to the three-finger crawling model,obtaining the grasp matrix about grasp position,establishing mathematical optimization model on maximum grasp force wrench,analyzing the relationship between the grasp position and the grasp force through the neural network.(3)Based on the characteristics and requirements of grasp force for the end-effector,an improved impedance control method was proposed.Through the stability analysis of the impedance control,finding the force error is related to environmental stiffness and environmental location.The RLS algorithm is adopted to realize the real-time calculation of the environment stiffness parameters.According to the forgetting factor the RLS algorithm is modified based on the force error,which improves the real-time performance of RLS algorithm.The validation of the proposed control algorithm was made with Matlab,the simulation results show a good performance of improved impedance controller.(4)Carry out grasp experiments with improved impedance control algorithm and impedance control algorithm respectively,the grasp experiment results show that the improved impedance control algorithm has a better force tracking performance,a lower damage rate and a stronger stability than impedance control,which demonstrate the controller has a huge advantages than impedance controller.