多关节串联机器人的建模与运动控制

发布时间：2019-04-27 19:09

【摘要】：我国是一个制造业大国,要向着制造业强国的方向转变,工业机器人在其中发挥着非常重要的作用。在机械臂的研究中,运动学是基础,运动规划是期望轨迹来源,轨迹跟踪控制是核心。本文基于这三个方面做出了相应研究。(1)机械臂的运动学:首先,以Denso VP6242G机器人为对象,推导出正运动学方程,根据正运动学方程用传统的反变换法解出运动学逆解,并仿真验证了运动学正逆解的正确性;然后,针对传统方法对一般结构的机械臂运动学逆解无法求解的问题,研究了一种适用于任何结构的不含反三角的高精度实时求解运动学逆解的解法,该解法基于改进的万有引力粒子群算法(PSOGSA)。通过仿真验证了该解法的准确性、快速性和求解唯一性;最后,在工具坐标系标定中,针对装有工具的机械臂标定困难的偏差大的问题,在多点多姿态估算法的基础上研究了一种新的快捷标定方法并加入判误机制,通过实验验证了所研究算法的正确性。(2)机械臂的运动规划:在总结一般形式运动规划的基础上,探讨用人工势场法自动避障的路径规划,并对该方法进行了简单避障仿真。同时,研究了一种5-3-5样条曲线插值的关节轨迹规划方法,该方法在保障速度、加速度连续可导的前提下,还能保障加加速度连续且起止于零,比常规方法更有利于抑制机械臂的振动,通过Matlab仿真和Denso机械臂平台验证了该方法的有效性。(3)机械臂的轨迹跟踪控制:通过简化Puma机械臂的六轴动力学模型得到后三轴锁死的前三轴模型,并在总结分析常用关节期望轨迹生成方法的基础上,得到机械臂关节轨迹跟踪控制的通用结构。同时,研究了PD、基于重力补偿的PD、基于计算力矩补偿的PD等关节轨迹跟踪控制方法,在此基础上研究了时变PD参数的计算力矩神经网络补偿控制方法,有效改善了综合控制效果。
[Abstract]:China is a big manufacturing country. Industrial robots play a very important role in the direction of changing to a powerful manufacturing country. In the research of manipulator, kinematics is the foundation, motion planning is the expected trajectory source, trajectory tracking control is the core. This paper makes corresponding research based on these three aspects. (1) Kinematics of the manipulator: firstly, taking the Denso VP6242G robot as the object, the forward kinematics equation is deduced, according to the forward kinematics equation, the inverse kinematics solution is solved by the traditional inverse transformation method, and the kinematics inverse solution is obtained by using the traditional inverse transformation method according to the forward kinematics equation. The correctness of forward and inverse kinematics solution is verified by simulation. Then, aiming at the problem that the traditional method can not solve the inverse kinematics solution of the mechanical arm of the general structure, a high-precision real-time solution of inverse kinematics solution without anti-trigonometry which is suitable for any structure is studied. The solution is based on an improved gravitational particle swarm optimization algorithm (PSOGSA). The accuracy, rapidity and uniqueness of the solution are verified by simulation. Finally, in the calibration of tool coordinate system, aiming at the difficulty of calibrating manipulator with tools, a new fast calibration method is studied on the basis of multi-point and multi-attitude estimation algorithm, and the mechanism of error judgment is added. The correctness of the proposed algorithm is verified by experiments. (2) Motion planning of manipulator: on the basis of summarizing the general form of motion planning, the path planning of automatic obstacle avoidance by artificial potential field method is discussed. A simple obstacle avoidance simulation is carried out. At the same time, a joint trajectory planning method based on the interpolation of 5 ~ 3 ~ 5 spline curves is studied. This method can guarantee continuous acceleration and zero acceleration under the premise that velocity and acceleration can be continuously derivable. It's better than conventional methods to suppress the vibration of the manipulator, The effectiveness of the proposed method is verified by Matlab simulation and Denso manipulator platform. (3) trajectory tracking control of the manipulator: by simplifying the six-axis dynamic model of the Puma manipulator, the front triaxial model of the post-triaxial locking is obtained. On the basis of summarizing and analyzing the common methods of joint expected trajectory generation, the general structure of manipulator joint trajectory tracking control is obtained. At the same time, the PD,-based gravity compensation PD, based on the calculated moment compensation of the joint trajectory tracking control method PD, on the basis of which the time-varying PD parameters of the calculated moment neural network compensation control method is studied. The comprehensive control effect is improved effectively.
【学位授予单位】：湖北工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242

【参考文献】