六自由度采摘机械臂运动及任务规划研究

发布时间：2018-04-14 15:34

本文选题：采摘机械臂 + 正逆运动学　；参考：《重庆理工大学》2017年硕士论文

【摘要】：近年来,我国的水果种植面积和总产量逐年增加,在水果生产作业过程中,50%-70%的生产成本源于采摘。采摘作业重复性强且劳动力需求大,但目前用于果实采摘的机械装备较少且智能化水平低,不能有效地将人从繁杂的农业生产中解放出来,不能有效的节约劳动力资源、提高生产率。因此,研究智能、高效的果实采摘机器人有助于将人从农业生产中解放出来,并降低生产成本、提高劳动生产率。机械臂是采摘机器人重要的执行部件,它的运动规划和任务规划直接影响着采摘机器人的采摘成功率和效率,因此有必要对采摘机械臂的运动规划和任务规划进行研究。本文主要的研究工作如下:(1)对采摘机械臂进行正逆运动学求解分析,利用解析法求出了机械臂的八组逆解,采用“最小能量”准则对逆解进行了选取,借助Robotics toolbox编写了基于MATLAB GUI的机械臂正逆运动学仿真界面,仿真结果表明:正逆运动学求解正确。(2)提出一种基于第五关节分离的采摘姿态规划方法,通过理论分析及MATLAB仿真表明该方法能够先判断果实是否在采摘空间内,然后确定机械臂的采摘姿态;针对关节空间的轨迹规划,对比了几种轨迹规划方法的运动特性,最终选择摆线轨迹规划;结合姿态规划和轨迹规划,进行了单果采摘仿真实验,实验结果表明:机械臂能够自适应规划姿态和轨迹,满足机械臂实际采摘的运动要求。(3)针对采摘机械臂连续采摘多个果实的作业任务,提出了一种基于“关节角加权最小”的任务规划方法,采用蚁群算法进行求解,编写了机械臂连续采摘多个果实的仿真界面,通过多组实例仿真结果表明:该方法不仅能够满足机械臂连续采摘多个果实的要求,而且当果实个数大于三个时,该方法比“路径最短”方法更节约能量。(4)在实验室内搭建了采摘实验平台,进行了单果采摘及多果采摘实验,单果采摘实验表明:运动规划能使机械臂完成单果采摘的运动要求,平均采摘时间为25.5s/个;多个果实采摘实验表明:基于“关节角加权最小”的任务规划能够使机械臂自动完成多个果实的连续采摘任务,且该方法的采摘时间比“路径最短”方法的采摘时间更短。
[Abstract]:In recent years, the planting area and total output of fruit in China have been increasing year by year, and 50% -70% of the production cost is due to plucking in the process of fruit production.Picking operations have strong reproducibility and large labor demand, but at present the machinery and equipment used for fruit picking is less and the level of intelligence is low, which can not effectively liberate people from the complex agricultural production, and can not effectively save labor resources.Increase productivity.Therefore, the study of intelligent and efficient fruit picking robot is helpful to liberate people from agricultural production, reduce production cost and improve labor productivity.Robot arm is an important executive part of picking robot, its motion planning and task planning directly affect the success rate and efficiency of picking robot, so it is necessary to study the motion planning and task planning of picking robot.The main research work of this paper is as follows: (1) the forward and inverse kinematics analysis of picking manipulator is carried out. Eight sets of inverse solutions of the manipulator are obtained by using analytical method, and the inverse solution is selected by using the "minimum energy" criterion.With the help of Robotics toolbox, the forward and inverse kinematics simulation interface of manipulator based on MATLAB GUI is compiled. The simulation results show that the kinematics solution of forward and inverse kinematics is correct. (2) A picking attitude planning method based on the separation of the fifth joint is proposed.The theoretical analysis and MATLAB simulation show that the method can determine whether the fruit is in the picking space and then determine the picking attitude of the manipulator. The kinematic characteristics of several trajectory planning methods are compared for the trajectory planning of joint space.Finally, the cycloidal trajectory planning is selected, and the simulation experiment of single fruit picking is carried out combining attitude planning and trajectory planning. The experimental results show that the robot arm can plan the attitude and trajectory adaptively.According to the task of picking multiple fruits continuously, a task planning method based on "minimum joint angle weighting" is proposed, which is solved by ant colony algorithm.The simulation interface of continuous picking of multiple fruits with mechanical arm is compiled. The simulation results show that the method can not only meet the requirements of continuous picking of more than three fruits, but also can be used when the number of fruits is more than three.This method saves more energy than the "shortest path" method. (4) the experimental platform of picking is set up in the laboratory, and the experiment of picking single fruit and multiple fruit is carried out.The results of single fruit picking experiment show that the motion planning can make the robot arm complete the motion requirement of single fruit picking, and the average picking time is 25.5s/;Multiple fruit picking experiments show that the task planning based on "joint angle weighted minimum" can make the robot arm complete the continuous picking tasks automatically, and the picking time of this method is shorter than that of the "shortest path" method.
【学位授予单位】：重庆理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP241

【参考文献】