Abstract: Human pulse records contain rich physiology and pathology information，which are emphasized in traditional Chinese medical science.A body sensor network based pulse measurement system is developed.The wireless sensor integrates the high sensitive PVDF pulse sensor and near field communication module.It’s proved to be convenient and effective without too much inference to the subject’s daily activities.Based on the pulse records collected from health subjects，multi time spatial features are compared in the gender distinguishment problem.SVM（Support Vector Machine） is used for the classification.The results show that the frequency domain feature leads to 56% correctness rate，which gives positive indications，and the time domain feature isn’t effective enough in the experiment.

Key words: body sensor networks, traditional Chinese medical informatization, pulse measurement, time-spatial feature extraction, gender distinguishment

摘要： 人体脉搏波中包含着丰富的生理病理信息，在中国传统医学中深受重视。基于脉搏波测量传感器和在体传感器网络通信技术，设计并实现了一种无线脉搏波测量装置，使测量过程更加方便灵活，并能远程进行。在实测数据的基础上，引入多种时频特征提取方法和支持向量机进行了男女性别区分的探索，频域特征准确率为56%，时域特征作用不明显。实验结果支持了中医界对切脉判定性别的说法，并说明频域特征对性别区分是有一定启示作用的。

关键词: 在体传感网, 中医信息化, 脉搏波测量, 时频特征提取, 性别区分