SIA OpenIR  > 机器人学研究室
可变形机器人构型机动性及路径规划方法研究
其他题名Research on Configuration Mobility and Path Planning Method for a Shape-shifting Robot
常健1,2
导师吴成东 ; 李斌
分类号TP242
关键词可变形机器人 滑移转向 评价 路径规划 机动性
索取号TP242/C36/2013
页数111页
学位专业模式识别与智能系统
学位名称博士
2013-05-26
学位授予单位中国科学院沈阳自动化研究所
学位授予地点沈阳
作者部门机器人学研究室
摘要本文依托的课题是国家“863”重点项目子课题“废墟洞穴搜救机器人研制”,首先针对处于直线构型下的可变形机器人在复杂非结构的搜救环境中难以转向的问题,研究可变形机器人转向机动性能,使其具有更好的环境适应性及完成任务的能力。其次针对废墟环境中,机器人在缺少操作人员指导却需要继续完成搜救任务的问题,研究机器人的变形能力与路径规划方法相结合的算法,实现动态环境中可变形机器人能够有效避障并进行路径规划。因此本论文的选题具有重要的理论意义和应用价值。可变形机器人AMOEBA-I能够利用其直线构型通过狭小的废墟空间,但是在此构型下转向性能很差。针对机器人进行转向时,连接杆受到较大的阻力矩影响且外界的复杂环境对其转向具有较大限制等问题,提出一种可变形机器人转向构型概念。通过对其研究,机器人可以有效减小转向半径、转向时间及转向阻力矩,提高可变形机器人直线构型对复杂环境的适应性。对这种转向构型进行运动学和动力学分析,证明了其能够有效的解决可变形机器人直线构型下的转向问题,而当外界环境空间对机器人转向空间影响较小时,提出了可变形机器人的滑移转向方式,研究机器人滑移转向前的变形过程并建立滑移转向的数学模型,分析机器人三个模块在转向过程中运动和力学特性,得出地面转向阻力矩及机器人所需驱动力的大小,完成机器人的滑移转向理论基础。通过实验分析,机器人在进行滑移转向时,驱动电流没有超出电动机额定值,机器人转向过程较于链式转向快速平稳,结果验证机器人滑移转向的有效性。机器人的转向性能是机器人机动特性主要研究内容之一,可变形机器人AMOEBA-I具有多种构型方式,每一种构型方式都具有不同的机动特性,而各种构型下的转向方式区别较大,因此研究机器人不同构型下的转向方式优劣具有重要的理论意义和应用价值。针对可变形机器人由于采用定性的转向评价方法而无法定量评价转向优劣的问题,提出了一种可变形机器人转向评价方法,通过对机器人转向所需的能量、角速度、时间、空间等指标的综合分析,同时利用基于五级主观赋值法及信息熵的客观赋权法相结合的主客观赋值法,确定机器人转向各因素的影响因子,得到机器人各种构型下转向优劣程度。通过对机器人在室内环境的各个转向因素进行仿真计算,得到各个构型下可变形机器人转向的评测值,采用定量的方式评测机器人转向方式的优劣,结果证明了机器人评测方法的有效性。针对复杂环境中可变形机器人需要具备的局部规划能力及提供给操作者辅助控制信息等问题,提出一种将机器人特有的变形能力融入到路径规划的方法。首先通过研究传统路径规划方法的不足及限制条件,提出了相应的改进方法,解决了障碍物不是质点、机器人局部震荡、机器人不能够到达目标点、机器人不能够变形通过狭小空间等问题。针对传统动态环境路径规划方法中,通常将障碍物信息认为已知可控,而现实条件下,障碍物的运动轨迹是未知的问题,研究了基于自回归方式的障碍物轨迹预测算法,利用前几个时刻障碍物信息对下一时刻障碍物的运动轨迹进行预测,增加整个算法使用的实际意义。针对可变形机器人所处的环境可能是动态变化的问题,提出一种基于多质心的可变形机器人的动态环境路径规划算法。通过提出一种改进的基于速度向量的势场法,将势场法中的距离信息转换成速度信息,同时对机器人各个模块分别建立势场函数,解决机器人能够利用构型变化躲避障碍物完成动态路径规划问题。实验结果表明,机器人能够对障碍物运动信息进行预测并完成动态环境的路径规划。
其他摘要The content of this doctor dissertation is around the National 863 plans sub-project of "Essential project", Research on a Relic Cave Rescue Robot. Firstly, as to the problem that the shape-shifting robot in line configuration can not take turning in non-structural complexity of the task environment, the performance of the robot mobility turning is researched in order to enhance its environmental adaptability and the ability to complete the task. Secondly, as to the problem that the robot has the ability to complete the tasks without the conduction of operator and help the operator to fulfill the rescue-searching tasks in ruins, the algorithm is researched. The ability of the robot and path planning is combinated in this algorithm and the shape-shifting robot can effectively avoid the obstalces and make the path planning successfully. Shape-shifting robot AMOEBA-I has the ability to pass the narrow and small ruins with the line configuration,but it is poor of the turning performance in this configuration. As to the problems that there are many restricts in use such as hardly turning,large turning radius,and the connecting links between modules, a concept about turning-configuration is proposed. Through the research of turning-configuration, the robot has the abilities to reduce the turning radius,the time of turning,the moment of resistance and enhance the compatibility of complex environment. The method is proved by kinematics and dynamics analysis. The method of skid-turning of robot is proposed as the environment has the little influence of the space of turning of the robot. Through the research of the process of the robot shape-shifting before it skid-turns and the establishment of mathematical model,the kinematics and mechanical properties of three modules are analyzed,and the resistance moment and the driving force is obtained. The theoretical basis of the shape-shifting robot skid-turning is computed and the evaluation criterion is proposed for the performance of robot turning. By the analysis of experiment,the drive current of the motor does not exceed the limit load of the motor and it is more quick and smooth comparing to the link-turning. The validity and enforceability of the turning is proved by simulations and experiments. The performance of turning of the robot is one of the main research content of mobility characteristics .The shape-shifting robot AMOEBA-I has many configurations and each configuration has the different characteristics of mobility, so there are many important theoretical significance and application values to evaluate the advantages and disadvantages of robot turning in different configurations. As the the problem that the advantages and disadvantages of robot can not be evaluated by qualitative manner withnot quantitative manner, an evaluation method of robot turning is proposed. The influenced factors are determined by steering comprehensive analysis of the energy, angular velocity, time, space, and other indicators of the robot turning and then taking advantage of the subjective and objective weighting method to decide the extent of robot turning. The subjective weighting method is based on five-grade scale and the objective weighting method is based on entropy of information to detetmine the factors. Through the theory analysis, the variouses of turning factors of robot are collected and computed in the ideal environment indoor. The evaluation of robot turning is gotten by quantitative manner and the effectiveness of this mehod of evaluation is proved. As to the problems that the robot should have the ability to take the path planning in local environment and help the operator to fulfill the tasks, an algorithm of path planning is proposed with the combination of the shape-shifting ability of the robot. Firstly, as to the problems that traditional algorithm of potential field have, it is solved that obstacle is not a particle, robot shocks in local and can not reach the target point, and the robot can get through the narrow space by shape-shifting. Secondly, as to the problem that it is considered that the information of obstacle to be known in traditional methods and in fact the trajectory of the obstacle is unknown, the method is proposed. In order to satisfy the actual significance of whole algorithm, the algorithm of prediction of obstacle is studied based on autoregressive way algorithm through the information of obstacle in first few moments to predict the information of obstacles in next times. At last, as to the problem that the environment of shape-shifting is dynamic, the algorithm of path planning is proposed. The algorithm is based on the multiple centroids of the modules of the robot. It converts the distance information between robot and obstacle into velocity information and establishes the potential field of each module of the robot. It can effectively solve the dynamic path planning problem. Experimental results show that the algorithm can predict obstacle motion information and can be simple to complete dynamic environment path planning.
语种中文
产权排序1
文献类型学位论文
条目标识符http://ir.sia.cn/handle/173321/10782
专题机器人学研究室
作者单位1.中国科学院沈阳自动化研究所
2.中国科学院大学
推荐引用方式
GB/T 7714
常健. 可变形机器人构型机动性及路径规划方法研究[D]. 沈阳. 中国科学院沈阳自动化研究所,2013.
条目包含的文件
文件名称/大小 文献类型 版本类型 开放类型 使用许可
可变形机器人构型机动性及路径规划方法研究(2622KB) 开放获取CC BY-NC-SA请求全文
个性服务
推荐该条目
保存到收藏夹
查看访问统计
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[常健]的文章
百度学术
百度学术中相似的文章
[常健]的文章
必应学术
必应学术中相似的文章
[常健]的文章
相关权益政策
暂无数据
收藏/分享
所有评论 (0)
暂无评论
 

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。