基于GSO-RF意图识别算法的全身助力外骨骼控制方法研究  

Research on a Full-body Power-assisted Exoskeleton Control Method Based on GSO-RF Intent Recognition Algorithm

作　　者：袁小庆[1] 吴涛 原勋 王文东[1] YUAN Xiaoqing;WU Tao;YUAN Xun;WANG Wendong(School of Mechanical Engineering,Northwestern Polytechnical University,Xi'an 710072)

机构地区：[1]西北工业大学机电学院,西安710072

出　　处：《机械工程学报》2024年第17期91-101,共11页Journal of Mechanical Engineering

基　　金：陕西省自然科学基础研究资助项目(2018JM5107,2020JM-131)。

摘　　要：针对现有助力外骨骼结构复杂、功能单一的问题,基于人体工学理论,设计一套上下肢一体化的全身助力外骨骼,提升助力效果。针对助力外骨骼运动意图识别准确率不高的问题,提出一种基于位姿信号、肌电信号的运动意图识别方法,引入有限状态机(Finite state machine, FSM),设定三种穿戴者运动状态;使用随机森林(Random forest, RF)算法对人机交互信号进行分类,确定穿戴者运动意图;采用人工萤火虫优化算法(Glowworm swarm optimization, GSO)优化随机森林,提高分类准确率,减少分类时间。为提高外骨骼控制系统轨迹跟踪精确性,保证外骨骼运动柔顺性,提出一种基于导纳自适应模糊反演算法的运动控制策略。搭建外骨骼实验平台,进行运动意图识别和轨迹跟踪实验,结果表明运动意图识别准确率可达96.6%,且实时性较好,提出的控制策略有效可行,轨迹跟踪效果较好。通过助力效能实验证明外骨骼的助力性能良好。To address the complex structure and single function of existing power-assisted exoskeletons,a full-body power-assisted exoskeleton is designed based on the ergonomic theory,and the upper and lower limbs are integrated to improve the assisted effect.To address the problem of low accuracy of motion intention recognition of the power-assisted exoskeleton,a motion intention recognition method based on posture signals and electromyography signals is proposed.The random forest(RF)algorithm is used to classify the human-robot interaction signals and determine the wearer's movement intention;the Glowworm swarm optimization(GSO)algorithm is used to optimize the random forest to improve the classification accuracy and reduce the classification time.To improve the trajectory tracking accuracy of the exoskeleton control system and ensure the suppleness of the exoskeleton movement,a motion control strategy based on the guided adaptive fuzzy inversion algorithm is proposed.The exoskeleton experimental platform is built to carry out motion intention recognition and trajectory tracking experiments.The results show that the accuracy of motion intention recognition can reach 96.6%,and the real-time performance is good,the proposed control strategy is effective and feasible,and the trajectory tracking effect is good.The booster performance of the exoskeleton has been proven through booster effectiveness tests.

关 键 词：全身助力外骨骼 运动意图识别 人机交互 运动控制 

分 类 号：TG156[金属学及工艺—热处理]