可变形六足仿生机器人运动分析与控制研究
发布时间：2019-02-19 19:35
【摘要】：近些年,在星际探测、复杂地形的抢险救灾、海底作业、野外监测、军事侦查等特殊危险作业行业中,任务难度越来越高,不适合人类去完成,这促使人们不断研究机器人技术,用机器人代替人类去完成艰巨的任务。本文针对带有变形关节的新型六足仿生机器人,深入研究了其单腿运动学、动力学和运动控制及并联机身运动学。主要工作如下:(1)利用旋量理论与指数积方法建立腿部运动学模型并给出了运动学显式逆解,针对腿部后两关节轴线交于一点的特点,结合经典消元理论和Paden-Kahan子问题方法计算其运动学显式逆解;基于旋量法计算摆动腿的雅可比矩阵,仿真结果验证了其正确性。(2)为了提高六足机器人腿部控制的平稳性、动作快速性,利用五次B样条曲线进行了关节空间轨迹规划,从而使关节运动的速度、加速度和脉动连续。利用B样条曲线的凸包性把腿部关节速度与加速度的曲线约束转化为对曲线控制顶点的约束。(3)当六足机器人处于三足支撑阶段时机体为3-URR并联机构,利用旋量理论和指数积方法建立其正运动学模型,利用消元理论和Paden-Kahan子问题方法建立其逆运动学模型,为进行六足机器人并联动力学分析和控制打下基础。(4)利用Kane动力学分析方法,建立了机器人单个可变形腿的逆动力学模型;可用于单个可变形腿运动控制器的设计;分析了关节转动速度和转动加速度对驱动力矩产生的影响,并分析了各关节驱动力矩中惯性力矩、向心力矩、哥氏力矩及重力矩所占的比例,分析结果有助于进一步规划机构的运动,优化机构设计和选择适当功率的电机。(5)设计了计算力矩加RBF神经网络复合控制器,其中计算力矩是已建模的部分,神经网络补偿动力学模型中未建模部分、结构参数的测量误差以及外部扰动。采用仿真和实验对腿部的轨迹跟踪控制进行了验证,结果表明了该控制器跟踪精度高,控制性能优良。
[Abstract]:In recent years, in the fields of interstellar exploration, rescue and relief of complex terrain, submarine operations, field monitoring, military reconnaissance and other special dangerous operations, the tasks are becoming more and more difficult and are not suitable for human beings to complete. This has prompted people to continue to study robot technology, using robots to replace human to complete the arduous task. In this paper, the kinematics, dynamics, motion control and kinematics of a new six-legged bionic robot with deformed joints are studied. The main work is as follows: (1) the kinematics model of leg is established by spinor theory and exponential product method, and the explicit inverse kinematics solution is given. The explicit inverse kinematics solution is calculated by combining classical elimination theory and Paden-Kahan subproblem method. The Jacobian matrix of swinging leg is calculated based on spinor method, and the simulation results verify its correctness. (2) in order to improve the stability of leg control of six-legged robot and the speed of action, the joint space trajectory planning is carried out by using the five-degree B-spline curve. Thus, the velocity, acceleration and pulsation of joint motion are continuous. By using the convexity of B-spline curve, the curve constraint of the velocity and acceleration of the leg joint is transformed into the constraint on the control point of the curve. (3) when the six-legged robot is in the three-legged support stage, the body is a parallel mechanism of 3-URR. The positive kinematics model is established by spinor theory and exponential product method, and the inverse kinematics model is established by means of elimination theory and Paden-Kahan subproblem method. It lays a foundation for parallel dynamics analysis and control of hexapod robot. (4) the inverse dynamic model of a single deformable leg of a robot is established by using Kane dynamic analysis method. It can be used in the design of single deformable leg motion controller. The effect of rotation speed and acceleration on the driving torque is analyzed, and the proportion of inertia moment, centripetal force moment, Coriolis moment and weight moment in the drive moment of each joint is analyzed. The analysis results are helpful to further plan the movement of the mechanism, optimize the mechanism design and select the motor of appropriate power. (5) the compound controller of calculating torque and RBF neural network is designed, in which the calculated torque is the part of the model. The unmodeled part of the neural network compensates the dynamic model, the measurement error of the structural parameters and the external disturbance. The trajectory tracking control of the legs is verified by simulation and experiment. The results show that the tracking accuracy of the controller is high and the control performance is good.
【学位授予单位】：河北工业大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TP242
[Abstract]:In recent years, in the fields of interstellar exploration, rescue and relief of complex terrain, submarine operations, field monitoring, military reconnaissance and other special dangerous operations, the tasks are becoming more and more difficult and are not suitable for human beings to complete. This has prompted people to continue to study robot technology, using robots to replace human to complete the arduous task. In this paper, the kinematics, dynamics, motion control and kinematics of a new six-legged bionic robot with deformed joints are studied. The main work is as follows: (1) the kinematics model of leg is established by spinor theory and exponential product method, and the explicit inverse kinematics solution is given. The explicit inverse kinematics solution is calculated by combining classical elimination theory and Paden-Kahan subproblem method. The Jacobian matrix of swinging leg is calculated based on spinor method, and the simulation results verify its correctness. (2) in order to improve the stability of leg control of six-legged robot and the speed of action, the joint space trajectory planning is carried out by using the five-degree B-spline curve. Thus, the velocity, acceleration and pulsation of joint motion are continuous. By using the convexity of B-spline curve, the curve constraint of the velocity and acceleration of the leg joint is transformed into the constraint on the control point of the curve. (3) when the six-legged robot is in the three-legged support stage, the body is a parallel mechanism of 3-URR. The positive kinematics model is established by spinor theory and exponential product method, and the inverse kinematics model is established by means of elimination theory and Paden-Kahan subproblem method. It lays a foundation for parallel dynamics analysis and control of hexapod robot. (4) the inverse dynamic model of a single deformable leg of a robot is established by using Kane dynamic analysis method. It can be used in the design of single deformable leg motion controller. The effect of rotation speed and acceleration on the driving torque is analyzed, and the proportion of inertia moment, centripetal force moment, Coriolis moment and weight moment in the drive moment of each joint is analyzed. The analysis results are helpful to further plan the movement of the mechanism, optimize the mechanism design and select the motor of appropriate power. (5) the compound controller of calculating torque and RBF neural network is designed, in which the calculated torque is the part of the model. The unmodeled part of the neural network compensates the dynamic model, the measurement error of the structural parameters and the external disturbance. The trajectory tracking control of the legs is verified by simulation and experiment. The results show that the tracking accuracy of the controller is high and the control performance is good.
【学位授予单位】：河北工业大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TP242
【相似文献】
相关期刊论文 前10条
1 肖红;;基于慧鱼技术的六足昆虫仿生机器人的设计[J];荆门职业技术学院学报;2006年03期
2 王丽慧;周华;;仿生机器人的研究现状及其发展方向[J];上海师范大学学报(自然科学版);2007年06期
3 ;我国研制成功水下微型仿生机器人[J];机械;2010年07期
4 殷悦;马丽;;仿生机器人及其应用[J];黑龙江科技信息;2011年32期
5 宋红生;王东署;;仿生机器人研究进展综述[J];机床与液压;2012年13期
6 ;仿生机器人比赛:贴近生活 创意无限[J];硅谷;2012年14期
7 简小刚;王叶锋;杨鹏春;;基于蚯蚓蠕动机理的仿生机器人研究进展[J];中国工程机械学报;2012年03期
8 林雨W，
本文编号：2426803
本文链接：https://www.wllwen.com/shoufeilunwen/xxkjbs/2426803.html
