拟人机器人手臂鲁棒非线性控制研究

发布时间：2017-12-30 19:40

本文关键词：拟人机器人手臂鲁棒非线性控制研究　出处：《郑州大学》2016年博士论文　论文类型：学位论文

【摘要】：移动速度快、灵巧、平滑的拟人机器人手臂在医疗康复、航空航天、社会服务等领域越来越受到重视。在这些特殊领域,拟人机器人手臂的运动不单单是以完成具体的动作和功能为主要目标,还要求在运动过程中能根据具体环境需求快速、平滑、精确地进行运动和执行动作,甚至能像多关节人臂一样熟练和灵巧。另外,应用拟人机器人手臂的目的是能够替代或辅助人们完成一些重复劳动或特殊工作,人们不仅希望这些拟人机器人手臂能够根据周围环境灵活地改变自身形态来完成期望的任务,在执行任务中还希望能够减少其对自身或被接触人类的伤害,这就要求一些能够根据不同需求自动变换形态的智能软体型机器人代替金属骨架型机器人来完成。因此,如何从多关节人臂的运动机理特性出发,结合机器人手臂的机械特性,构建基于特殊仿人智能软体驱动器的拟人机器人手臂,实现其结构和功能仿生,是制造出功能集中、效率高、像多关节人臂一样灵巧和熟练的拟人机器人手臂的有效途径之一。基于现阶段在构建拟人机器人手臂中存在的软体驱动器建模与控制、先进控制等技术难题,本论文通过对拟人机器人手臂鲁棒非线性控制的研究,深入探讨了软体驱动器的机理特性与建模、仿多关节人臂运动机理的拟人控制器设计、多输入多输出机器人手臂的先进鲁棒控制等几个关键技术。其主要研究内容和成果包括:1.软体智能驱动器的特性分析与建模。软体机器人能在非结构化的环境中自如运动,具有优良的安全性和灵活性。虽然软体机器人有着传统刚性机器人无法比拟的优势,但材料的柔软性使得软体驱动器的运动机理变得极其复杂,给软体驱动器或机器人动力学建模、控制器设计等带来了极大困难。为解决制约软体机器人的这些实际应用问题,本论文以一种单腔双向弯曲软体驱动器为研究对象,讨论了基本结构、设计原理、强弯曲特性等问题。分析和研究了基于物理特性的压力输入和弯曲角度输出关系的位置控制模型,并通过实验数据辨识出一种实用的压力输入与力输出关系的力控制模型。2.拟人机器人手臂鲁棒非线性控制系统设计。通过对拟人机器人手臂的研究背景、研究意义及其发展现状的分析,提出了一种刚性和软体混合结构的拟人手臂系统总体结构框架,进行了拟人机器人手臂鲁棒非线性控制系统设计研究。在其控制系统设计中,首先,讨论了拟人机器人手臂非线性鲁棒控制系统总体架构,提出了以构成拟人机器人手臂的刚性机器人臂和软体纯软手指为主要控制和执行单元的两个子系统的控制性能指标。其次,分别设计了基于算子理论的鲁棒右互质分解技术的机器人臂与纯软微手鲁棒非线性控制子系统,完成了对机器人臂末端位置和软手指位置和力的控制研究。推导出了基于演算子的控制器和跟踪控制器,并对机器人臂鲁棒控制子系统进行了鲁棒稳定与跟踪特性分析与评价,讨论了鲁棒稳定与跟踪条件。同时,也对机器人软手鲁棒非线性控制子系统进行了鲁棒稳定与跟踪特性分析与评价,讨论了鲁棒稳定与跟踪条件。最后,对整个拟人机器人手臂系统的非线性鲁棒稳定与跟踪特性进行了分析讨论。3.基于多关节人臂结构与粘弹性特性的结构和功能拟人化方法。通过对如何将多关节人臂的粘弹性特性有效地应用在拟人机器人手臂系统设计与控制中的研究,探索出一种实现拟人机器人手臂的结构和功能拟人化方法,也就是通过对运动中多关节人臂粘弹性机理和机器人手臂的动态机械特性的分析与研究,设计出一种基于多关节人臂粘弹性特性的仿人臂控制器。4.基于演算子理论的不确定性补偿控制问题。模型不确定性会带来跟踪控制器设计困难问题,针对本问题,提出了一种基于演算子理论的不确定性补偿控制框架,通过设计基于标称模型特性的演算子观测器完成了对模型不确定的补偿,消除了不确定对设计精确跟踪控制器的影响。
[Abstract]:Moving fast, smart, smooth arm of humanoid robot in medical rehabilitation, aerospace, more and more attention in the field of social services. In this special field, humanoid robot arm movement is not only to complete the action and function of concrete as the main target, according to the needs of the specific environment also requires fast, smooth movement in the process in the movement and action accurately, even as many people as skilled and dexterous arm joint. In addition, the application of humanoid robot is to can replace or assist people to complete some repetitive labor or special work, people not only hope that the humanoid robot can be expected to complete according to the surrounding environment the flexibility to change its form the task in the execution of a task is to reduce its contact or human injury, which requires some can according to different requirements Intelligent soft body automatic robot shape shifting robots instead of the metal skeleton. Therefore, how to proceed from the mechanism of multi joint arm, combined with the mechanical characteristics of the robot arm, build humanoid robot special intelligent software driver based on arm, the structure and function of bionic, creating function, high efficiency, as one of the effective ways of humanoid robot arm and dexterous skilled people like multi joint arm. At this stage in the construction of software modeling and control of actuator are humanoid robot based on the technical problems of advanced control, this paper through the research of robust nonlinear control of humanoid robot arm, discusses the characteristics and modeling mechanism software driver, controller design humanoid multi joint arm movement imitation mechanism, multi input and multi output of the robot arm in robust control The key technology of system. The main research contents and achievements include: 1. characteristics analysis and modeling software. The software can drive intelligent robot moving freely in the unstructured environment, with excellent safety and flexibility. Although the software robot has a traditional rigid robot incomparable advantages, but the soft material making the movement mechanism of software driver has become extremely complex, to the software driver or robot dynamic modeling, controller design has brought great difficulties. To solve these practical problems which restrict the robot, this paper takes a single cavity bidirectional bending drive software as the research object, discusses the basic structure, design principle, strong bending characteristics etc.. The analysis and study of the relationship between the pressure and the bending angle of the physical characteristics of the input output model based on position control, and according to the experimental data identification A kind of practical pressure force input and output relations of the force control arm model robust nonlinear control system design of the.2. humanoid robot. Through the research background of humanoid robot, analysis the significance of the research and development status, put forward a kind of rigid and soft mixed node framework structure of humanoid arm system of anthropomorphic system the design of nonlinear robust control of robot arm. In the design of the control system, firstly, discusses the humanoid robot nonlinear robust control system architecture, put forward the control performance of the two subsystems to form a quasi man arm robot with rigid robot arm and the software of pure soft finger as the main control and execution units. Secondly, the design of the robot arm are robust right coprime decomposition technique of operator theory and pure soft micro hand robust nonlinear control system based on, Study on the control of the robot arm end position and soft finger position and force complete. Deduced controller operator and tracking controller based on arm and robust control subsystem of robot are analyzed and evaluated the robust stability and tracking characteristics, discusses the robust stability and robust tracking conditions. At the same time, also on the soft hand robust robot the nonlinear control system of analysis and evaluation of robust stability and robust tracking characteristics, discusses the robust stability and robust tracking conditions. Finally, nonlinear robust stability and tracking properties of the humanoid robot system are analyzed and discussed the structure and function of.3. anthropomorphic arm joint structure and method of viscoelastic properties based on how best to. The study on the viscoelastic properties of multi joint arm effectively applied in design and control system of humanoid robot arm, explore a quasi human The structure and function of the anthropomorphic robot arm method, is through the dynamic mechanical properties of multi joint arm movement viscoelastic mechanism and robot arm of analysis and research, design a.4. controller of humanoid arm arm viscoelastic properties of multi joint people based on operator theory of uncertain compensation control problem the model uncertainty will bring about tracking controller design problem, aiming at this problem, put forward a kind of operator theory of uncertain compensation control based on the framework, through the design of the nominal model characteristics of operator observer completed the uncertainty of model based compensation, eliminate the uncertain influence on the design of accurate tracking controller.

【学位授予单位】：郑州大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TP242

【相似文献】