基于PCA和DCT的雷达人体动作识别

无载波超宽带雷达人体动作识别系统的关键优势在于无载波超宽带雷达具有极高的分辨率，能够捕获人体的细微动作变化，并且对于室内复杂环境具有很强的抗干扰能力。但是由于无载波超宽带雷达信号不含载波信息，本身能量集中于极窄的波形内，并且发射信号与回波相关性弱，因此传统的提取信号特征的方法不再适用。针对这一问题，首次搭建无载波超宽带雷达人体动作识别系统，并提出一种新颖的基于主成分分析法（PCA）和离散余弦变换（DCT）相结合的无载波超宽带雷达人体动作识别方法，同时利用改进的网格搜索算法优化支持向量机的参数并验证该方法的优越性。最后，基于实测数据在Matlab平台上进行仿真，对实测的10种不同类型的人体动作进行分类识别，实验结果显示，该方法具有很高的识别率，针对不同的方案识别率均能达到99%以上，对小训练样本具有很强的鲁棒性。

A Novel Human Action Recognition System Based on PCA and DCT for Carrier-Free UWB Radar

The key advantage of human action recognition system based on the carrier-free UWB radar is that the carrier-free UWB radar has a very high resolution which can capture the slight movements of the human action and has a strong anti-jamming capability in indoor complex environments. However, the carrier-free UWB radar signal does not contain carrier information and its energy is concentrated on a very narrow waveform, so the correlation between the transmitted signal and the echo signal is weakly. Thus, the traditional method of extracting the signal feature is no longer suitable. For this problem, a carrier-free UWB radar human action recognition system is first established, and a novel carrier-free UWB radar human action recognition method based on principal component analysis (PCA) and discrete cosine transform (DCT) is proposed. The improved grid search algorithm is used to optimize the parameters of SVM and verify the superiority of the proposed method. Finally, based on the measured data simulation in the Matlab platform, the actual measurement of the ten different types of human actions are classified and identified. The experimental results show that the proposed method has a high recognition rate (above 99% for different schemes), and is very robust under small training samples.