膝上型假肢的运动意图识别与控制
发布时间：2019-03-20 14:31
【摘要】：根据不同路况条件和典型步速的笛卡尔积组合,利用装配在残肢侧的陀螺仪、加速度计和足底前后的压力传感器的信息,通过相关性系数分析、传感器融合、隐马尔可夫模型的方法,判断假肢使用者的运动意图.以健肢运动状态为参考值,利用迭代学习控制分别建立不同路况和步速情况下的控制知识数据库.通过传感器的关键状态变化信号驱动有限状态机状态转换,输出控制知识库中的控制量,实现假肢膝关节在不同路况、步速条件下对步态相位的控制.针对控制过程中出现的输出量实时偏差,采取了在线校正措施.对于有限状态机输出控制数据序列在时间同步上的超前和滞后问题,采取了相应的保持和补偿措施.结果表明,经隐马尔可夫模型处理后路况判断准确率可提升到91.7%,基于数据驱动的无模型控制方法能够实现对不同路况、步速下假肢步态的有效控制.
[Abstract]:According to different road conditions and typical Cartesian product combination, using the information of gyroscope, accelerometer and pressure sensor in front and back of foot, through correlation coefficient analysis, the sensor is fused. The Hidden Markov Model (hmm) is used to judge the motion intention of the prosthetic user. Taking the motion state of the healthy limb as the reference value, the control knowledge database under different road conditions and walking speed was established by using iterative learning control (ILC). The key state change signal of the sensor drives the state transition of the finite state machine and outputs the control quantity in the control knowledge base to realize the control of the gait phase of the knee joint under different road conditions and gait speed conditions. Aiming at the real-time deviation of output in the process of control, on-line correction measures are adopted. For the time synchronization problem of the output control data series of the finite state machine, the corresponding measures are taken to maintain and compensate the time synchronization. The results show that the accuracy of road condition judgment can be improved to 91.7% after the processing of Hidden Markov Model. The model-free control method based on data-driven can effectively control the gait of artificial limbs under different road conditions and gait speeds.
【作者单位】： 河北工业大学控制科学与工程学院;河北工业大学智能康复装置与检测技术教育部工程研究中心;
【基金】：国家自然科学基金资助项目(61703135,61773151) 河北省自然科学基金青年基金资助项目(F2016202327)
【分类号】：R496;TP212.9
本文编号：2444304
[Abstract]:According to different road conditions and typical Cartesian product combination, using the information of gyroscope, accelerometer and pressure sensor in front and back of foot, through correlation coefficient analysis, the sensor is fused. The Hidden Markov Model (hmm) is used to judge the motion intention of the prosthetic user. Taking the motion state of the healthy limb as the reference value, the control knowledge database under different road conditions and walking speed was established by using iterative learning control (ILC). The key state change signal of the sensor drives the state transition of the finite state machine and outputs the control quantity in the control knowledge base to realize the control of the gait phase of the knee joint under different road conditions and gait speed conditions. Aiming at the real-time deviation of output in the process of control, on-line correction measures are adopted. For the time synchronization problem of the output control data series of the finite state machine, the corresponding measures are taken to maintain and compensate the time synchronization. The results show that the accuracy of road condition judgment can be improved to 91.7% after the processing of Hidden Markov Model. The model-free control method based on data-driven can effectively control the gait of artificial limbs under different road conditions and gait speeds.
【作者单位】： 河北工业大学控制科学与工程学院;河北工业大学智能康复装置与检测技术教育部工程研究中心;
【基金】：国家自然科学基金资助项目(61703135,61773151) 河北省自然科学基金青年基金资助项目(F2016202327)
【分类号】：R496;TP212.9
【相似文献】
相关期刊论文 前1条
1 毕胜,吴怀宇,章刚华,鄢达来,熊沈蜀,王福根,周兆英;功能性电刺激中应用P型迭代学习控制在上肢多关节运动控制的试验研究[J];中华理疗杂志;2001年02期
相关会议论文 前1条
1 毕胜;王福根;鄢达来;吴怀宇;章刚华;熊沈蜀;周兆英;;功能性电刺激P型迭代学习控制方法的研究[A];中国康复医学会第四届会员代表大会暨第三届中国康复医学学术大会论文汇编[C];2001年
相关博士学位论文 前1条
1 Oluwarotimi Williams Samuel;提高肌电上肢假肢控制的稳定性与可靠性研究[D];中国科学院大学(中国科学院深圳先进技术研究院);2017年
相关硕士学位论文 前2条
1 王朋;基于MYO臂环的假肢手控制技术研究[D];上海师范大学;2017年
2 赵丽娜;基于迭代学习控制的有限状态机在动力型假肢中的应用[D];河北工业大学;2015年
，本文编号：2444304
本文链接：https://www.wllwen.com/kejilunwen/zidonghuakongzhilunwen/2444304.html
