摘要
人体行走步态是一类重要的生物特征,通过分析连续波雷达采集的人体反射回波信号可以获取行走步态的关键特征。提出了人体行走步态特征信号模型,在构建连续波步态试验雷达的基础上,应用时频分析方法对采集的步态样本信号进行了分析,重点讨论了多分量非稳态信号的时频分析方法,对比了短时傅里叶分析、WIGNER-VILLE分布、伪平滑WVD分布及其进一步改进方法在分析步态回波信号中的应用,通过仿真和对实际数据进行分析,可以看出基于伪平滑再分配的WVD方法具有较好的步态信号分辨率,为进一步的特征提取打下基础。
Human walking gait is one of the biological feature for recognition.The human echo signal to a CW radar contains the human gait feature.A human gait signal model is formed and based on the constructed test CW gait radar,the gait sampled signal is analyzed by time frequency method,and the methods for analyzing multi component non-stationary signal are discussed in detail.The comparison among the algorithm of STFT,WVD,Pseudo-smoothed WVD on gait sig-nal and its improvements is given,the basic method for gait feature extraction based on time-frequency analysis is pro-posed,and the results in this paper will be a well support for further research.
出处
《电子测量与仪器学报》
CSCD
2011年第6期500-505,共6页
Journal of Electronic Measurement and Instrumentation
基金
天津自然科学基金(编号:09JCYBJC27300)资助项目
