CASIA OpenIR  > 毕业生  > 博士学位论文
下肢康复训练机器人关键技术研究
王晓楠
学位类型工学博士
导师原魁
2017-05-23
学位授予单位中国科学院大学
学位授予地点北京
关键词下肢康复训练 运动分析 轨迹规划 人机交互 协作控制
其他摘要
        随着社会老龄化问题的日益严峻,各类残疾人和长期卧床病人大量存在并持续增长,康复医疗工作面临前所未有的压力,助老助残问题正日益成为一个重大的社会问题。减重型下肢康复训练机器人针对下肢残障患者,辅助患者进行下肢康复训练,能够为老年人和下肢残障患者提供更好的康复治疗水平,降低康复治疗人员的工作强度,对于提高老年人和下肢残障患者康复后的生活质量,减轻社会负担具有重要的现实意义。本文探讨了减重型下肢康复训练机器人所涉及的相关关键技术,并通过实验验证了所提方法的有效性和可靠性。针对减重型下肢步行康复训练机器人相关的技术领域,论文的主要工作和创新之处如下: 
        第一,基于视觉的人体下肢运动分析。以人体下肢表面布置的红外标志点作为运动分析的基础,提出了一种下肢标志点的识别算法和人体下肢在运动过程中位姿计算的方法。基于课题组开发的基于FPGA和DSP的嵌入式智能图像卡,开发了一种基于双目视觉测量步态采集系统。通过捕捉健康被试者在跑步机上的运动轨迹,建立健康个体的下肢运动轨迹库。针对人体下肢的运动特点,建立了人体下肢关节运动模型,对人体下肢运动数据进行量化分析。采集得到的健康个体下肢运动轨迹和下肢运动模型为下肢康复训练机器人轨迹规划打下了基础。 
       第二,患者被动式康复训练方法与实现。针对康复初期患者自主运动能力较弱的特点,患者被动的康复训练方法。考虑到单一步态轨迹对不同患者的适应性较差,提出了一种训练轨迹的参数化生成方法。基于健康个体的下肢运动数据库,实现了一种基于GRNN神经网络的患者训练轨迹规划方法,根据患者自身的生理学特征和运动学参数对训练轨迹进行规划,有效地提高机器人训练轨迹与患者自然步态的匹配度,增强训练轨迹对患者的适应性。通过分析机器人的运动学模型,根据当前患者的生理学参数生成的下肢运动轨迹,实现了康复训练机器人辅助患者进行被动康复训练。
        第三,患者主动式康复训练方法与实现。康复中后期患者已经具备一定程度的自主运动能力,通过机器人与患者的协同控制实现患者主动的康复训练方法。基于“机器人-患者”的人机混合动力学模型,提出了一种通过机器人与患者之间的力信号获取患者在运动过程中的主动运动意图的方法。根据患者的主动意图,实现了基于患者主动意图的控制方法,并成功应用到了康复机器人上。针对单侧肢体残障患者,提出了一种基于患者健康一侧肢体的运动轨迹的协作训练方法。患者按在训练中学习自身健康一侧肢体的运动规律,有利于改善患者康复后双侧步态不协调的问题。本文通过实验对患者主动训练方法的可行性与可靠性进行了验证。 
        最后,对论文中的工作进行了总结,并讨论了可在其基础上进行的拓展工作。
;
    With the problem of aging society becoming more and more serious, various disabled and long-term bedridden patients abound and continue to grow, the medical rehabilitation work is faced with unprecedented pressure, and to help the disabled and elderly is becoming a major social problem. Body-weight-support rehabilitation robot for lower limbs is designed for people with lower limb dysfunction, assisting patients to perform physical rehabilitation sessions. Body-weight-support rehabilitation robot can provide better rehabilitation treatment for the elderly and the disabled, reducing the work load of therapists, improving the quality of life of the elderly and the disabled after rehabilitation and has an important practical significance for reducing the burden of the society. In this paper, the key technologies of the body-weight-support rehabilitation robot are discussed, and the validity and reliability of the proposed method are verified by experiments. The main work and innovation of this paper are as follows:
       Firstly, vision-based human lower limb movement analysis is proposed. With infrared markers arranged on the surface of human lower limb, the identification algorithm of lower limb inferred marker is proposed, and the method of calculating the position and posture of human lower limb via binocular stereo vision is utilized. The proposed method is implemented on an image processing board with FPGA and DSP as main calculation units. And the trajectory of the lower limbs of healthy individuals is captured on treadmill. Through the analysis of the trajectory of the lower limbs of healthy individuals, the model of lower limb joint movement was established, and the collected data were analyzed by the lower limb movement model. The trajectories of the lower limbs of the healthy individuals were collected and applied to the trajectory planning of the lower limb rehabilitation training robot.
   Secondly, the passive training methods for patients is proposed and implemented. In view of the characteristics of the patients with lower ability of independent movement in the early stage of rehabilitation recovery, the passive training method is utilized. Given the poor adaptability of a single gait trajectory to different patients, this paper presents a parametric generation method for training trajectories. Based on the database of lower limb movement of healthy individuals, a neural network based motion trajectory generation algorithm is implemented, which can effectively improve the matching degree between robot training trajectory and natural gait and enhance the adaptability of training trajectory to patients. By analyzing the kinematic model of the lower limbs of the lower limbs of the robot, the rehabilitation training of the rehabilitation robot was carried out according to the trajectory of the lower limb generated by the physiological parameters of the current patient.
       Thirdly, the active training methods for patients is proposed and implemented. In the middle or late stages of rehabilitation, patients have a certain degree of autonomy movement ability; the active rehabilitation training method is achieved via the robot and patient collaborative control method. Through the force sensing signal between the robot and the patient, the active motion intention of the patient is obtained based on the dynamic model of robot-patient hybrid system. According to the patient's initiative intention, to achieve the initiative based on the patient's control method, and successfully applied to the rehabilitation of the robot. For the patients with unilateral limb disability, a method of training for unilateral limb disability based on the trajectory of the healthy side of the patient was proposed. During the training,the patients can learn from their own movement,improving the condition of patients after bilateral gait incongruity problems. The feasibility and reliability of the patient-in-charge method is verified by experiments.  
       Finally, the summarization of the work is given and the further work is proposed.
语种中文
文献类型学位论文
条目标识符http://ir.ia.ac.cn/handle/173211/14702
专题毕业生_博士学位论文
作者单位中国科学院自动化研究所
推荐引用方式
GB/T 7714
王晓楠. 下肢康复训练机器人关键技术研究[D]. 北京. 中国科学院大学,2017.
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
论文 最新版(最终).pdf(3359KB)学位论文 暂不开放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
[王晓楠]的文章
百度学术
百度学术中相似的文章
[王晓楠]的文章
必应学术
必应学术中相似的文章
[王晓楠]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。