基于脚部惯性传感数据的人员运动速度识别方法

现有的基于脚部惯性传感数据的人员运动速度估计方法只能对人员低速行走时的速度进行有效的估计。为了采用脚步惯性传感数据识别人员快速行走以及跑步时的速度，该文提出了一种利用单步统计特征进行速度识别的方法。该方法利用脚部惯性传感器对人员在不同速度下运动的惯性数据进行采集，采用峰值检测的方法对数据进行单步划分，最后从单步数据中提取65维统计特征分别采用最小二乘法（LS）、支持向量机（SVM）、K近邻（KNN）、线型贝叶斯正态分类器（LDC）4种常见的机器学习分类方法对人员运动速度进行识别。经实验验证，所建议的方法中采用SVM分类器的识别率高达96.3%，所以采用该方法可以有效的识别人员的运动速度。

Speed Recognition Using Foot-mounted Inertial Sensing Data

The existing pedestrian speed estimation method based on foot-mounted inertial sensing data could only be used for slow walking effectively. In order to obtain the velocity of pedestrian while walking fast and running, the paper proposed an approach utilizing statistical characteristics of each step for speed recognition. The method acquired inertial sensing data at different speeds with a foot-mounted inertial sensor firstly. Then, the approach extracted 65 different statistical characteristics from inertial data of each step which acquired by peak detection method. Finally, the proposal making use of several machine learning classifiers, such as Least Square (LS) classifier, Support Vector Machine (SVM) classifier, K-NearestNeighbor (KNN) classifier, Linear Bayes Normal Classifier (LDC) for speed recognition. Numerous experiments prove that proposed method could recognize velocity preciously while the accuracy of SVM classifier achieved 96.3%. As a result, the presented approach can recognize speed effectively.