双臂机器人拟人化动作实现与协调控制方法研究

发布时间：2018-01-13 21:07

本文关键词：双臂机器人拟人化动作实现与协调控制方法研究　出处：《中国科学技术大学》2017年博士论文　论文类型：学位论文

【摘要】：拟人双臂机器人在精密打磨、精密装配、大负载搬运等工业领域有其独特的优势,它是下一代工业机械臂智能化发展的方向。拟人双臂机器人是仿照人手臂的自由度构成和分配而设计的一种机械臂系统,其具备操作空间大、灵活性高以及协同能力强等特点而被广大学者研究;由于冗余自由度的存在,双臂机器人系统运动控制研究水平一定程度上标志着一个国家在机械臂研究领域的学术地位。本文以拟人双臂机器人为研究对象,以人体双臂运动机理、双臂机器人运动规划方法以及双臂协调运动控制为研究内容,针对传统运动规划方法操作复杂、效率低下以及双臂机器人运动控制系统存在的非线性等问题开展相关研究。本文研究的主要内容如下1.提出了基于直角三角函数的手臂关节旋转的评价方法。从人类手臂的特点出发,结合人体解剖学,分别对肩部、肘部与前臂、腕部进行了详细阐述及运动学分析,构建了肩关节-肘关节-手腕组成的人类手臂运动链,并对其运动学进行了建模。分析了手臂关节角的运动轨迹、角度变化以及肌肉的作用力特征,解决了手臂在旋转过程中肩部和肘部关节力矩的计算问题。2.提出了基于人体手臂运动捕获数据的对偶四元数空间向量法关节角计算方法。针对传统人体双臂运动捕获方法存在的不足,描述了光学运动捕捉系统的优点以及当前在运动捕捉数据的分析与处理方法中有待解决的问题。应用空间向量法对人体肩部和肘部关节角进行了计算,通过对偶四元数对原型标记点位置的逆运动学进行了求解,避免了欧拉角的奇异点问题。3.采用了欧几里得群结构与最优控制逆解相结合的方法,建立了人机匹配模型。围绕机器人复现人体双臂运动的目标,通过对捕获人体双臂运动轨迹上的速度和关节角度变化数据进行修正,并将修正后的数据与机器人模型进行运动学匹配,解决了人体运动数据到双臂机器人的最佳模型匹配问题。实现了仿人双臂机器人的拟人化运动。4.双臂机器人位置跟踪精度不同控制策略的对比研究。在经典反演控制器设计的基础上,建立了一种改进型的双臂协调反演控制策略。该策略有三个特点:第一,采用模糊逼近反演控制算法,对非线性不确定性进行估计,进一步提高末端位置跟踪的精度。基于力和位置的混合控制系统,模糊逼近末端期望轨迹,完成了双臂的协调抓取任务。第二,采用扩展的自适应反演控制方法来控制双臂稳定工作,充分利用反演法的优点,使用一阶滤波器减少控制器的复杂度。基于力矩补偿的模糊自适应反演控制方法,对机器人双臂实现准确的轨迹跟踪。第三,采用滑模自适应反演控制方法来减少系统总干扰,结合非线性扰动观测器的特点,提出了一种非线性扰动观测的滑模自适应反演控制方法。该方法削弱了滑模控制的抖振,提高了仿人双臂机器人系统的控制性能,保证双臂系统稳定工作。5.设计了双臂机器人拟人化控制和双臂协调稳定操作的完整方案。针对机器人双臂实现拟人化操作问题,应用目前通用的仿真平台V-REP(Virtual Robot Experimentation Platform)从仿真、实验角度验证了本论文提出方法的有效性,为仿人机器人双臂协调稳定控制以及双臂拟人化复现操作提供了一个新的参考方向和方法,拓展了双臂机器人的应用,具有积极的学术指导价值和重要的工程实际意义。
[Abstract]:Polish, humanoid dual arm robot in precision precision assembly, large load handling and other industrial fields has its unique advantages, it is the industrial manipulator intelligent the direction of the next generation. A manipulator system is modeled on the humanoid dual arm robot arm of the degree of freedom of allocation and the structure and design, with its large operation space the characteristics of high flexibility and coordination ability and the majority of scholars; the redundant degree of freedom, the research level to some extent marks the academic status of a country in the research field of mechanical arm control system. Based on the motion of dual arm robot humanoid dual arm robot as the research object, taking the human arm movement mechanism, the motion planning method of arms the robot arms and coordinated motion control as the research content, the traditional motion planning method of complicated operation, low efficiency and robot motion control system Research on nonlinear problems. The main contents of this paper are as follows: 1. put forward the evaluation method of rotating arm joint angle based on trigonometric function. Based on the characteristics of the human arm, combined with human anatomy, respectively on the shoulder, elbow and forearm and wrist were analyzed in detail and kinematics, construct the human arm movement chain shoulder elbow wrist composition, and its kinematics was modeled. Analysis of trajectory of the arm joint angle, angle and force characteristics of muscle, solve the arm calculation is presented based on human arm motion capture dual four element space vector method of joint angle data.2. method in the calculation problem of rotation over the shoulder and elbow joint torque in the process. According to the shortcomings of the traditional acquisition methods of human arm movement, describes the advantages and optical motion capture system The motion capture methods of data analysis and processing of the problem to be solved. The application of space vector method to calculate the human shoulder and elbow joint angle, through the dual four element inverse kinematics of the markers of the prototype are solved to avoid the singularity of Euler angles using Euclidean.3. control method of inverse solution the combination of group structure and the optimal matching model, established the human-machine robot arms around. Redisplay human moving target, through the capture of human arms on the trajectory velocity and joint angle change data are corrected, and the model data and the modified robot kinematics, solves the optimal matching model of human motion data to arms the problem of robot. The position of the anthropomorphic arms movement.4. humanoid robot dual arm robot tracking accuracy comparison of different control strategies Study. Based on the classical backstepping controller design, the establishment of an improved arms coordination control strategy. The inversion strategy has three characteristics: first, using the fuzzy control algorithm of nonlinear inversion approximation, the uncertainty of estimates, to further improve the end position tracking precision. Hybrid control system based on force and position at the end, the fuzzy approximation of the desired trajectory, completed the task coordination capture arms. Second, using the extended adaptive backstepping control method to control arms stability, taking full advantage of the complexity of the inversion method, using a reduced order filter controller. Fuzzy adaptive backstepping control method based on torque compensation, to achieve accurate trajectory the tracking robot arms. Third, using sliding mode adaptive backstepping control method to reduce the total interference system, combined with the characteristics of nonlinear disturbance observer, presents a Sliding mode adaptive inverse control method of nonlinear disturbance observation. This method reduces chattering in sliding mode control, improve the control performance of humanoid arm robot system, to ensure the stable operation of the.5. system design of the dual arm robot arms control and stable operation of anthropomorphic arms coordination scheme. According to the robot arms to achieve complete anthropomorphic operation problems and application at present the general simulation platform V-REP (Virtual Robot Experimentation Platform) from the perspective of simulation, experimental validation of the proposed method, it provided a new direction and method for humanoid robot arm coordinate stability control and anthropomorphic arms reproduction operation, expand the application of the arms of the robot, it has practical significance to the academic guiding value and important project.

【学位授予单位】：中国科学技术大学
【学位级别】：博士
【学位授予年份】：2017
【分类号】：TP242

【相似文献】