基于sEMG信号的下肢康复机器人肌力预测模型与交互式自适应阻抗控制研究

项目名称： 基于sEMG信号的下肢康复机器人肌力预测模型与交互式自适应阻抗控制研究

项目编号： No.51475342

项目类型： 面上项目

立项/批准年度： 2015

项目学科： 机械、仪表工业

项目作者： 艾青松

作者单位： 武汉理工大学

项目金额： 83万元

中文摘要： 老年人、残疾人和其他具有肢体运动障碍人群的康复训练已成为亟待解决的社会问题。本项目重点研究人体下肢表面肌电信号(sEMG)的分析处理及其肌力预测模型，以及下肢康复机器人交互式自适应阻抗控制方法。主要包括康复机器人的动力学模型与运动控制技术，提出一种人机交互式下肢康复机器人阻抗控制模型；针对康复训练过程中肢体运动受限且连续动作切换的特点，设计一种高分离度的sEMG多尺度特征提取和模式识别方法；通过sEMG信号和力、角度、加速度等传感器对下肢肌肉活动状态进行监测与评估，建立一种定量准确的sEMG信号与肌力之间的映射模型；结合模糊神经网络等推理方法，优化康复机器人在人机交互时的阻抗控制模型，提出一种新的匹配患者主动运动意图和肌体活动能力的交互式自适应阻抗控制方法，构建一个科学高效的康复机器人运动控制系统，为运动障碍人群的下肢康复训练和辅助运动提供模式多样、性能稳定、效果显著的康复训练策略与系统。

中文关键词： 下肢康复机器人；表面肌电信号；肌力预测模型；自适应阻抗控制；交互式控制

英文摘要： It has become an important social issue to meet the increasing requirements of rehabilitation and training for a growing population of elderly, disabled people and those with limb movement disorders. This project mainly focuses on the analyzing and processing methods of surface electromyography (sEMG), and the mapping model of the relationship between sEMG and humans' muscle strength prediction, as well as the interactive and adaptive impedance control methods for lower limb rehabilitation robots. The research contents include: study of the dynamic model and position control technology for parallel rehabilitation robots; proposal of a human-robot interaction impedance control model for lower limb rehabilitation robots; design of a high precision sEMG feature extraction and pattern recognition algorithm aiming at solving the problem of the lower limb movements during rehabilitation procedure often being restricted and tending to be continuously switched; establishment of a quantitative predictive model of the relationship between sEMG signals and the muscle strength by integrating the the sEMG signals and the force, angle, and acceleration sensors to monitor and evaluate the muscle activity status;proposal of a novel adaptive impedance control approach based on patient's movement abilities and recovery conditions, and construction of a scientific and efficient control system by introducing fuzzy neural networks and other inference methods to optimize the impedance control model of rehabilitation robots, so as to provide people with movement disorders with diverse, stable, and effective training strategies and systems for lower limb rehabilitation.

英文关键词： Lower limb rehabilitation robot;surface electromyography;muscular force prediction model;adaptive impedance control;interactive control