液压驱动下肢外骨骼机器人摆动相控制系统研究

发布时间：2018-02-24 00:22

本文关键词： 液压驱动下肢外骨骼结构优化逆雅克比矩阵人机交互系统多传感器系统滑模变结构控制模糊控制自适应控制　出处：《浙江大学》2017年博士论文　论文类型：学位论文

【摘要】：下肢外骨骼机器人的样机在最近几年不断涌现,目前主要应用在助力及康复训练等领域,其作为典型的人机交互型机器人,涉及机构学、电子技术、计算机技术、控制技术、信息技术、传感技术、人工智能和仿生学等多学科知识。液压驱动下肢外骨骼机器人是一个典型的非线性系统,因此如何实现与穿戴者进行友好的人机交互,充分发挥穿戴者运动的灵活性以及外骨骼设备的耐力将是一个系统研究的课题。第一章通过广泛的国内外调研发现,阐述了课题研究的背景和意义。第二章,首先借助于CAD辅助设计软件、数值计算方法等工具实现参数的优化选择,并使用SolidWorks软件进行了三维绘图;其次设计了液压系统和电控系统,完成了液压驱动下肢外骨骼平台的搭建。第三章,对摆动腿的运动学、动力学建模,便于设计基于模型的上层控制器;通过建立单关节液压缸的非线性模型,设计了三种底层力跟踪控制器;针对传统DLS方法在解决下肢外骨骼摆动腿雅可比矩阵奇异性问题时所引发的新问题,提出了三种改进方法,并通过MATLAB进行了仿真验证。第四章,首次提出将获取人体运动意图的方法分为两层:第一层主要实现对步态的判别,因此设计了多传感器鞋底;第二层实现对人体运动意图物理信息的获取,在人机之间安装多维力传感器,并且引入导纳模型作为推导人体运动意图的方法,实现人机交互力与人体运动意图物理信号的建模。最后,搭建了人机交互系统摆动相的控制系统,明确了上层控制算法与底层单关节液压缸力跟踪控制器的不同应用点。第五章,针对下肢外骨骼机器人系统的非线性和系统中不确定因素的干扰,引入了滑模变结构控制器。为了减小系统稳态跟踪误差,引入了积分滑模面。针对传统的积分滑模面容易引起Windup效应,出现超调或执行器饱和,提出了两种改进方法。并且采用趋近律的以改善趋近运动的动态品质,最后通过MATLAB仿真验证了系统性能。为了改善系统的平滑性,提高系统的鲁棒性,引入模糊系统以逼近干扰力矩,避免了因随意对干扰力进行估值所造成的系统波动。第六章,考虑到对于复杂的非线性系统,难以获取准确的动力学模型。提出了一种基于单输入的模糊自适应滑模变结构控制器,通过利用模糊系统的逼近特性和自适应控制的强鲁棒性来设计滑模变结构控制器,从而降低控制器对模型的依赖性。第七章,对论文的主要研究工作进行了总结,并且描述了相关创新点。
[Abstract]:The prototype of exoskeleton robot of lower limb has been emerging in recent years. At present, it is mainly used in the fields of assistance and rehabilitation training. As a typical human-computer interactive robot, it involves mechanism, electronic technology, computer technology, control technology, etc. Information technology, sensing technology, artificial intelligence, bionics and other multidisciplinary knowledge. Hydraulic drive exoskeleton robot is a typical nonlinear system, so how to achieve friendly man-machine interaction with the wearer, It will be a systematic research topic to give full play to the exercise flexibility of the wearer and the endurance of exoskeleton equipment. Chapter one, through extensive domestic and international investigation and discovery, expounds the background and significance of the research. First of all, the parameters are optimized by means of CAD aided design software and numerical calculation method, and the 3D drawing is carried out by using SolidWorks software. Secondly, the hydraulic system and electric control system are designed. In the third chapter, the kinematics and dynamics of the swing leg are modeled to facilitate the design of the upper controller based on the model, and the nonlinear model of the single joint hydraulic cylinder is established. In order to solve the problem of Jacobian matrix singularity of exoskeleton swinging leg in lower extremity, three improved methods are proposed and verified by MATLAB. Chapter 4th. For the first time, the method of obtaining human motion intention is divided into two layers: the first layer mainly realizes the discrimination of gait, so a multi-sensor sole is designed, and the second layer realizes the acquisition of physical information of human motion intention. A multi-dimensional force sensor is installed between man and machine, and the admittance model is introduced as the method to deduce the motion intention of human body. Finally, the physical signal of human-computer interaction and human motion intention is modeled. The control system of the oscillating phase of the man-machine interactive system is built, and the different application points of the upper control algorithm and the bottom single joint hydraulic cylinder force tracking controller are clarified. Chapter 5th, Aiming at the nonlinearity of exoskeleton robot system and the disturbance of uncertain factors in the system, a sliding mode variable structure controller is introduced. In order to reduce the steady-state tracking error of the system, a sliding mode variable structure controller is introduced. In this paper, the integral sliding mode surface is introduced. In view of the Windup effect caused by the traditional integral sliding mode surface, overshoot or actuator saturation, two improved methods are proposed, and the approach law is adopted to improve the dynamic quality of the approaching motion. Finally, the system performance is verified by MATLAB simulation. In order to improve the smoothness and robustness of the system, the fuzzy system is introduced to approximate the disturbance torque, and the system fluctuation caused by the random estimation of the disturbance force is avoided. Chapter 6th, Considering that it is difficult to obtain accurate dynamic model for complex nonlinear systems, a fuzzy adaptive sliding mode variable structure controller based on single input is proposed. By using the approximation property of fuzzy system and the strong robustness of adaptive control, a sliding mode variable structure controller is designed to reduce the dependence of the controller on the model. Chapter 7th, the main research work of this paper is summarized. And describes the relevant innovation points.
【学位授予单位】：浙江大学
【学位级别】：博士
【学位授予年份】：2017
【分类号】：TP242

【相似文献】