轮椅机器人履带张紧力估计与控制方法研究
文献类型：学位论文
| 作者 | 滕赟 |
| 学位类别 | 硕士 |
| 答辩日期 | 2010-06-03 |
| 授予单位 | 中国科学院沈阳自动化研究所 |
| 授予地点 | 中国科学院沈阳自动化研究所 |
| 导师 | 姚辰 |
| 关键词 | 变形履带轮椅机器人 主动张紧力 静力学 运动控制 |
| 其他题名 | Wheel Chair Robot Track Tension Estimation and Control Method Research |
| 学位专业 | 模式识别与智能系统 |
| 中文摘要 | 目前许多国家已进入老龄化社会，行动不便的人口数量越来越多，随着社会文明程度的提高，人们对残疾人的关心程度也越来越高。传统的轮椅依靠人力运动，且不具备上下楼梯的功能。科学技术的蓬勃发展使得开发一种能够爬楼梯的轮椅成为可能。爬楼梯轮椅移动机器人期望于能够安全平稳的使得乘坐者在无外力帮助的情况下通过常见障碍物，要求其具有较高安全性、较强的环境适应能力和有效的操控性。 本论文以中国科学院沈阳自动化研究所正在研发的一种变形履带式爬楼梯轮椅机器人为平台，对其履带张紧力的动态决策、动态性能及控制方法进行了较为系统的研究。 该种轮椅机器人的移动机构为履带式移动机构，由可变形履带机构，支撑架，座椅等部分组成。两个前主动轮单独驱动，差速完成转向。后轮为倾角从动轮，用以当车体倾角发生改变时调整轮椅姿态，为了保证两侧姿态的一致性，由一台电机驱动。前方配有摆臂用以调整履带张紧力，两侧各由一台电机驱动。以嵌入式控制器为核心进行控制系统模块化设计，保证了系统重量轻、体积小、实时性好和可靠性高的应用要求。 作为智能方法的一种，模糊技术可以对那些数学模型难以求取或无法求取的对象进行有效控制或决策。考虑到获得柔索张紧力的精确模型通常都较为困难，但根据经验可知约束履带张紧力大小的因素。故可以认为应施加张紧力的大小是各个因素贡献的叠加，本文中各个因素的贡献由模糊决策确定，然后由一组样本输入与期望输出通过最小二乘法得到各因素权重，从而估计出最优张紧力。相对于直接的曲线拟合，该方法融合了直观经验，物理意义鲜明，具有一定的决策能力，且需要的样本较少。传统的模糊方法规则数随输入变量的增加呈指数增长，这里将各种因素分别考虑，大大减少了规则数量，是一种动态主动张紧力决策方法。 轮椅机器人由于乘坐者的特殊性，对其行驶过程中的平稳性要求较高。由于轮椅的运动过程为高度非线性，多变量耦合系统，故建立动态模型不仅十分困难同时也难以设计出有效地控制器。结合动态张紧力的结果，对慢速过程进行静力学分析。这里电机力矩被分为两部分，一部分用以平衡静力学力矩，一部分用控制器进行调整。分析结果表明上述方法是有效的，能够很大程度上提高轮椅的平稳性。 轮椅的硬件控制部分采用研华的5510系统，该系统是研华公司生产的基于PC的、能够独立完成数据采集与控制、具备RS485/RS232通讯功能的“微缩型”可编程微控制器。ADAM5510控制器的二次开发编程环境是Borland Turbo C++3.0，这使得开发控制程序的周期大大缩短。ADAM5510开发软件包中提供了诸如查询式RS232/485通讯，存取、控制I/O模块等功能的C语言库函数，在自己的应用程序中可调用这些函数。 |
| 索取号 | TP242/T49/2010 |
| 英文摘要 | Many countries have been into ageing society. The number of people who have difficulty of movement has become larger and larger. As the development of social society, people take care of the disabled more than ever. Traditional wheelchairs depend on man’s strength to move and do not have the ability to climb stairs. With the high speed development of science, it is possible to develop a kind of wheelchair which can climb stairs. The wheelchair robot is expected to carry passengers to across usual obstacle without extra help. It should have the properties of high security , good adoption of environment and effective manipulation. According to a kind of transmuting track wheelchair robot which is under research in Shenyang institute of automation Chinese academy of sciences, the paper systemically investigate the dynamical determination of track tension, dynamical property and control method. The frame of wheelchair robot consists of fixed shelf, seat and track transmutation mechanism. The former two wheels are driven respectively and turn by the velocity difference. The back two wheels rotate according to pitch to adjust wheelchair gesture. They are driven by one electric motor to keep consistency of two sides. The former two swing arms driven respectively to adjust track tension. With embed controller as core to design the control system modularly, we can guarantee the whole system has low weight, good real time reaction and high reliability. As an intelligent method, fuzzy technology can control and make decision to the object whose math module can’t be gotten. Considering the difficulty to get the accurate module of soft rope, we know the factors that restrict track tension. So it can be deemed that value of anticipated tension is the superposition by the effect of each fact. Then by a group of input and supposed output sample, we decide the weight of each fact through least square method and get the optimally estimation of track tension. Compared to direct curve fitting, this method combines intuition experience and has obvious physics meaning. Also it has decision making ability and needs less training samples. The rules number of traditional fuzzy method increases by exponential along with the number of input variables. It can decrease the number of rules much to consider each fact respectively. It is a dynamical track tension determination method. As the specialties of passengers, the wheelchair is required to have high stability. Because the movement process is nonlinear, multi-variables, it is difficult to get dynamical module and design effective controller. Combined with track tension, we analyze statics with low speed process. The torque is divided into two parts, one is to balance statics torque, the other is adjusted by controller. Analysis result shows the above method is effective and can improve stability quite well. The control hardware part adopts 5510 system which is produced by yanhua corporation. It is based on PC, can gather data independently and has the RS485/RS232 communication ability. Its second development environment is Borland Turbo C++3.0, which shorten work time. ADAM5510 afford C library functions such as RS232/485 communication, data withdraw, controlling I/O module etc. We can call those during application. |
| 语种 | 中文 |
| 公开日期 | 2012-07-27 |
| 产权排序 | 1 |
| 分类号 | TP242 |
| 源URL | [http://ir.sia.ac.cn/handle/173321/9394]  |
| 专题 | 沈阳自动化研究所_机器人学研究室 |
| 推荐引用方式 GB/T 7714 | 滕赟. 轮椅机器人履带张紧力估计与控制方法研究[D]. 中国科学院沈阳自动化研究所. 中国科学院沈阳自动化研究所. 2010. |
入库方式： OAI收割
来源：沈阳自动化研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。