SIA OpenIR  > 机器人学研究室
WT轮椅机器人楼梯攀爬过程动力学建模与控制
Alternative TitleDynamics Modeling and Control for WT Wheelchair Robot during Stair-Climbing Process
王剑1,2
Department机器人学研究室
Thesis Advisor吴成东
ClassificationTP242
Keyword轮椅机器人 楼梯攀爬 完整系统
Call NumberTP242/W33/2014
Pages107页
Degree Discipline模式识别与智能系统
Degree Name博士
2014-05-29
Degree Grantor中国科学院沈阳自动化研究所
Place of Conferral沈阳
Abstract本文的研究内容主要围绕国家自然科学基金项目“越障轮椅机器人的机构、动力学及稳定性研究”展开。该项目提出一种可变形的轮、履复合式移动机构,并研究出一款可上下楼梯的轮椅机器人—WT轮椅机器人,该机器人可通过前、后摆臂的摆动来调整攀爬楼梯过程中轮椅座椅的位姿,同时控制履带以保持一定的张紧力。 本论文的核心研究内容是轮椅机器人的控制问题,实现轮椅机器人在攀爬楼梯的过程中座椅保持水平,并且履带的张紧力能够保持恒定或某种恰当的变化,可以归结成完整系统的控制问题。由WT轮椅机器人的控制问题又引出两个问题,控制轮椅保持水平需要对轮椅机器人进行路径规划,对轮椅机器人同时进行路径控制和力控制需要建立WT轮椅机器人的动力学模型。而对轮椅机器人路径规划和动力学建模的前提是得对轮椅机器人与楼梯相互作用的机理进行分析,轮椅与楼梯环境不同的作用形式必然会造成轮椅机器人的路径和动力学不同。针对这四方面问题本论文的具体内容安排如下: (1)“系统与环境作用”方面,没有统一的方法解决这一问题,已有文献对越障机构与楼梯的相互作用关注很少。本文中提出一种界定轮椅机器人与楼梯作用性质的概念,模式。其特点是轮椅机器人与楼梯处于不同模式下会有不同的相互作用,会对轮椅机器人的动能及势能造成不同影响,从而使系统的Lagrange函数对广义坐标有不同的表达式,模式的判别在于归纳出轮椅机器人与楼梯相互作用的整个过程所有模式,可以为在不同的模式下的路径规划和动力学建模做准备。 (2)“系统动力学建模”方面,WT轮椅机器人是具有完整约束的移动机器人,而且参照攀爬楼梯定义下的假设,若轮椅机器人在铅直平面内运动的话,则蕴含着左右两个驱动轮的广义坐标恒等的条件,若假设轮椅机器人左右两侧履带周长相同,则又蕴含着机器人系统中含有两个相同的完整约束的条件。本文中提出了两个定理来化简这两种对称结构,并获得机器人的最简动力学模型。 (3)“系统路径规划”方面,比较经典的路径规划是针对底座固定,串联机构形式的机器人,应用“运动学反解”的思路,在机器人的关节空间中规划出路径。而WT轮椅机器人是复杂的混联机构移动机器人,而且带有完整约束,传统路径规划方法有所局限。本文中提出了“任务规划法”,联立“目标方程”和约束方程得方程组,并将方程组的解作为规划出的路径。 (4)“控制律设计”方面,应用复杂的数学或力学工具比如微分几何对应的几何力学来解决完整或非完整力学问题已经有很长历史了,几何力学对完整或非完整系统的物理含义已经有了很清晰和简洁的描述,对完整或非完整系统的可控性和稳定性等也有过很好的论述。但是在完整或非完整系统的控制论上还是很有局限的。传统的设计非完整系统控制律的方法最后都可以归结于欠驱动控制。即系统的控制器数等于或小于系统的广义坐标数。还有一种力/位混合的思路,这是一种分立设计的思想,其局限就在于位控制器与力控制器不能同时起作用,一种控制器只能作为另一种控制器的辅助,而且这种思路主要针对施力的对象固定的平面等,当然这是一种完整或非完整约束控制问题。只是完整或非完整系统的约束平面不固定,比较复杂。McClamroch和Wang等提出在约束平面上建立切平面和法向坐标系的方法来化简这种问题,但是这种思路把问题的处理引入了一个复杂的方向,而且他们在推导中出现了偏差,使力控制出现了余差。本文中提出一种设计完整或非完整机器人系统控制律的框架,并结合计算力矩法和动态规划法设计出WT轮椅机器人攀爬楼梯过程中的主动张紧力控制律,即在实现对各关节变量参考输入曲线跟踪的同时,实现对完整约束的约束反力参考输入曲线的跟踪。
Other AbstractThe research in this dissertation is based on a project supported by National Natural Science Foundation of China. In the project, a novel deformable wheel-track mobile mechanism is proposed, and based on such mobile mechanism a novel wheelchair robot—WT wheelchair robot, which can climb stairs, is proposed. By swing the two pairs of flippers of the proposed wheelchair robot one can adjust the position of the seat of the wheelchair as well as control the tracks to keep a certain tension during stair-climbing process. The core content of this paper is about the control problem of WT wheelchair robot, which is to control the seat of the wheelchair horizontaol as well as to control the tracks to maintain a certain tension or some appropriate changes. Such problem can be categorized into the control problem of holonomic system. However, there are also two problems raised from the control problem, dynamics modeling for the robot and path planning before designing control law to maintain seat of the wheelchair horizontal. Before solving such two problems one need also to analyze the mechanism of the interaction between the wheelchair robot and the stairs for a different interaction between the robot and the stairs should cause the defferences of dynamics modeling and path planning. According to the four aspects of the problems the content of the paper can be arranged as follows: (1) “Interaction between system and environment”: there is no uniform method to solve this problem and attention of the existing literature pay to the interaction between obstacle-crossing mechanism and stairs is limit. In this paper, a new concept namely pattern, which is used to determine the interaction between the wheelchair robot and stairs, is put forward. And by using such definition one can tell out the different interaction in different pattern, which can make different effects on the kinetic energy and potential energy. (2) “Dynamics modeling of the system”: WT wheelchair robot is a mobile robot with holonomic constraints, and when referencing to the assumption of the definition of climbing stairs, if the wheelchair robot moves in the vertical plane, there contains a condition that the generalized coordinates of two driving wheels of the robot are identical to each other all the time. Also when assuming the perimeters of two tracks are the same, there contains another condition that there are two equal holonomic constraints in the system. In the paper, two theorems are proposed to simplify the two symmetrical structures, and then the most simple dynamics model of robot is obtained. (3) “Path planning of the system”: classical path planning aims at the robot with fixed base and serial mechanisms, which employs “inverse kinematics” idea to plan a path in the joint space of the robot. But such classical method cannot be applicable to WT wheelchair robot which contains complex mechanisms and holonomic constraints. In the paper, a novel method namely “stack planning” is proposed, by which one establishes group of a simultaneous equations of “objective equations” and constrained equations and make the solutions as a path. (4) “Control law”: there is a long history for researchers to apply complex mathematical or mechanical methods such as differential geometry or geometric mechanics to solve holonomic and nonholonomic mechanical problems. And there is also a clear and concise description about the physical meaning of holonomic and nonholonomic systems based on geometric mechanics as well as about the controllability and stability. However it is still limited in the control theory of holonomic and nonholonomic systems. The traditional method can be attributed to designing control law for so called underactuated system with the number the controlled input is equal to or less than the number of generalized coordinates of the system. Although this method can control the pose of system, it is incapable of control for the constrained forces. There is also another method namely force/position hybrid control, which is a discrete design idea with force controller and position controller working not at the same time. And such idea mainly aims at the environment with a fixed plane. McClamroch and Wang propose a method to establish the tangent plane and normal plane in the constrained plane to realize force/position hybrid control, which points to a complicated situation. And in their deduction a residual error of the force is introduced. In the paper, a novel control framework for holonomic and nonholonomic systems is proposed, based on which combing with computed torque method or dynamic programming method one can design control law to control the robot to track the reference input of each joint as well as to control the robot to track the reference input of constrained forces.
Language中文
Contribution Rank1
Document Type学位论文
Identifierhttp://ir.sia.cn/handle/173321/14829
Collection机器人学研究室
Affiliation1.中国科学院沈阳自动化研究所
2.中国科学院大学
Recommended Citation
GB/T 7714
王剑. WT轮椅机器人楼梯攀爬过程动力学建模与控制[D]. 沈阳. 中国科学院沈阳自动化研究所,2014.
Files in This Item:
File Name/Size DocType Version Access License
WT轮椅机器人楼梯攀爬过程动力学建模与控(8750KB) 开放获取CC BY-NC-SAApplication Full Text
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[王剑]'s Articles
Baidu academic
Similar articles in Baidu academic
[王剑]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[王剑]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.