基于多传感器信息融合的步态识别方法研究  

Research on Gait Recognition Based on Multi-Sensor Fusion

作　　者：周洁[1] 

机构地区：[1]西南交通大学电气工程学院,成都262500

出　　处：《现代计算机（中旬刊）》2016年第3期40-45,共6页Modern Computer

摘　　要：准确的步态相位识别是下肢外骨骼机器人运动意图判断与控制策略制定的前提与基础。通过人体下肢运动信息检测系统,获取下肢各关节运动角度及足底分布式压力,在运动学步态分析的基础上,提取信号有效特征。针对正常步行运动中各步态存在次序及界限不精确的特性,提出加入模糊逻辑概念的有限状态机方法,对下肢传感器系统提供的运动信号进行多传感器信息融合,实现步态识别。实验结果证明该方法在2.4km/h、3.6km/h、4.8km/h和6.0km/h四种速率的步行模式下都可以进行有效的步态识别,具有极高的准确率。Accurate gait phase recognition is the premise and foundation for lower extremity exoskeleton robot motion intention judging and control strategy formulating. Through the human lower limb motion information detection system, gets the lower limb joint motion angle and plan- tar distributed pressure, and extracts effective signal features based on the analysis of kinematic gait. According to the characteristics of gait＇s inaccurate order and limit by normal walking, puts forward a finite state machine method combined of fuzzy logic to realize gait recognition, which fusing multi-sensor information for the motion signals provided by the lower limb sensor system. Experimental results proved this algorithm can achieve effective and accurate gait recognition on the four walking rates of 2.4km/h, 3.6kin/h, 4.8km/h and 6.0km/h.

关 键 词：步态识别 关节运动角度 足底压力 模糊逻辑 有限状态机 传感器融合 

分 类 号：TP274.2[自动化与计算机技术—检测技术与自动化装置]