Path Planning Of A 6-DOF Apple Picking Manipulator Based On Stability Control

At present,apple picking has many disadvantages,such as high cost,low efficiency and high labor intensity.In order to solve these problems,the research on intelligent and efficient apple picking robot can effectively reduce labor intensity,reduce production cost and improve productivity.Therefore,this thesis focuses on the study of mobile picking robotic arm,which mainly includes the track tracking of mobile platform,recognition and three-dimensional positioning of fruits,and planning of picking sequence,specifically including the following parts:(1)Accurate trajectory tracking is a prerequisite for the mobile manipulator to carry out subsequent picking tasks.Therefore,a sliding mode control(SMC)method based on long and short term memory network(LSTM)and quasi-sliding mode is studied to complete trajectory tracking control.Firstly,the kinematic model and dynamic model of the tracked vehicle are given,and the sliding mode control system is established based on the dynamic model.Then,LSTM network based on deep learning method is designed to control and compensate unknown interference items,reduce the influence of external interference,and reduce the tremor phenomenon by combining the advantages of LSTM network and quasi-sliding mode.The stability analysis of the control system is given.Finally,MATLAB/Simulink simulation experiments are carried out on different tracks,and compared with the existing method,which proves its superiority.(2)In order to solve the problem that it is difficult to recognize and locate when multiple apples are adjacent to each other,a recognition and positioning method based on AP-K-means clustering and three-point circle method contour fitting based on conditional constraints is studied.Firstly,the color image and depth image were preprocessed to get fruit contour pixels.Then AP-K-means clustering was used to complete the preliminary classification of contour pixels.Then,the three-point circle drawing method was used to obtain multiple centers for each type of contour pixel points,and these centers were screened according to the constraints.At the same time,AP-K-means clustering was used to cluster the selected centers,and the addition and average of the centers of each cluster were obtained to fit the center and contour of the fruit.Finally,for the two-dimensional pixel coordinates in the image coordinate system,the three-dimensional coordinates in the robot arm base coordinate system are obtained through coordinate transformation.The correctness rate of fruit recognition and root mean square error(RMSE)of location results were used as indexes to verify and analyze the algorithm.The accuracy rate of recognition was 85.7%,the mean square error of X-axis coordinate was 0.24 cm,the mean square error of Y-axis coordinate was 0.35 cm,and the mean square error of Z-axis coordinate was0.17 cm.(3)A multi-objective optimization method based on simulated annealing algorithm was studied to improve the picking efficiency when faced with multi-fruit picking tasks.Firstly,distance and energy consumption were taken as optimization objectives,optimization functions were designed,and the weights were obtained by entropy weight method.Then,the optimal planning of multi-fruit picking was obtained by simulated annealing algorithm.Finally,experimental verification was carried out and compared with the simulated annealing algorithm based on distance optimization,simulated annealing algorithm based on energy consumption optimization and mountain climbing algorithm.Experimental results show that the proposed algorithm has certain optimization in picking distance and energy consumption.In this thesis,from the three aspects of the trajectory tracking of the moving platform,the recognition and three-dimensional positioning of the fruit,and the planning of the picking sequence,the relevant research and experiment of the motion path planning of the6-DOF apple picking robot arm based on the stability control was completed.From the experimental results,some conclusions for reference were obtained.