论文部分内容阅读
家居环境的活动监测备受人们关注.传统的基于视频技术和可穿戴传感器的方法,都受到硬件成本的限制.提出了一个利用普通的商用WiFi设备的人体动作识别系统,通过分析WiFi信号的信道状态信息(CSI)识别家居环境中8个常见的人体动作.为了获取相同时刻的CSI测量值,提出针对不同时间间隔的CSI序列进行插值处理的方法.通过分析不同的子载波和人体动作的相关性,提取不同动作对应的子载波特征方差,进而采用基于稀疏表示分类(SRC)的算法进行分类.在真实的家居环境中对该系统进行实验,平均识别率可达到96.4%.
The monitoring of activities in home environment attracts much attention.Traditional methods based on video technology and wearable sensors are limited by the cost of the hardware.A human motion recognition system using common commercial WiFi devices is proposed, The state information (CSI) identifies eight common human actions in the home environment.In order to obtain CSI measurements at the same time, a method of interpolating CSI sequences with different time intervals is proposed. By analyzing the correlation between different subcarriers and human actions , And extract the subcarrier feature variance corresponding to different actions, and then use the algorithm based on sparse representation classification (SRC) to classify.In the real home environment of the system experiments, the average recognition rate of up to 96.4%.