基于超宽带雷达的人体动作识别技术研究

Research on Human Motion Recognition Technology Based on UWB Radar

论文就基于超宽带雷达的人体动作识别技术研究这一课题,从实验数据采集、回波信号处理、特征参数提取、机器学习分类四个方面展开了研究工作,针对四种仅凭时频结果较难区分的动作,提出了一种多域峰值点跟踪算法,使用该算法可以对动作的时间、频率、距离特征进行分析;根据定义的特征参数对特征分析结果进行特征提取,构造人体动作特征数据集,最后使用机器学习分类算法对人体动作进行识别,测试样本识别结果验证了该分类算法的有效性。

In this paper,the human motion recognition technology based on UWB radar is studied from four aspects:experi⁃mental data collection,echo signal processing,feature parameter extraction and machine learning classification.A multi-domain peak point tracking algorithm is proposed for four kinds of actions which are difficult to be distinguished based on time-frequency re⁃sults alone.According to the defined characteristic parameters,the feature analysis results are extracted and the human motion fea⁃ture data set is constructed.The results show that the recognition accuracy of the SVM algorithm is high,and the recognition results of test samples verify the effectiveness of the classification algorithm.