基于改进粒子群算法的UWB雷达人体动作识别研究OACSTPCD
Research on UWB radar human motion recognition based on improved particle swarm optimization algorithm
针对雷达信号中的杂波干扰及样本数量对人体动作识别精度的限制,提出一种基于改进粒子群算法(parti-cle swarm optimization,PSO)优化支持向量机(support vector machine,SVM)模型的超宽带(ultra-wideband,UWB)雷达人体动作识别算法.利用动态目标指示(moving target indication,MTI)与小波阈值滤波对接收到的UWB回波信号进行预处理,消除回波信号中的杂波和噪声对人体动作识别的影响;结合二维离散小波包分解(two dimensional discrete wavelet packet decomposition,2D-DWPD)与奇异值分解(singular value decomposition,SVD),对预处理后的雷达信号进行特征提取和降维;提出一种改进粒子群算法,优化SVM模型的相关参数进行识别和分类.实验结果表明,提出的算法准确率可达到 96.25%,具有良好的识别性能.
For the clutter interference in radar signals and the limitation of the number of samples on the accuracy of human motion recognition,this paper proposes an ultra-wideband(UWB)radar human motion recognition algorithm based on im-proved particle swarm optimization(PSO)to optimize the support vector machine(SVM)model.Moving target indication(MTI)and wavelet threshold filtering are used to preprocess the received UWB echo signals to eliminate the influence of clutter and noise in the echo signals on human motion recognition.Two-dimensional discrete wavelet packet decomposition(2D-DWPD)and singular value decomposition(SVD)are combined to extract features and reduce dimensions of the pre-processed radar signals.An improved particle swarm optimization algorithm is proposed to optimize relevant parameters of the SVM model for recognition and classification.Experimental results show that the accuracy of the proposed algorithm can reach 96.25%,and it has good recognition performance.
李新春;曾仕豪
辽宁工程技术大学 电子与信息工程学院,辽宁 葫芦岛,125105辽宁工程技术大学 研究生院,辽宁 葫芦岛,125105
电子信息工程
超宽带雷达人体动作识别小波阈值滤波改进粒子群算法
ultra-wideband radarhuman motion recognitionwavelet threshold filteringimproved particle swarm optimiza-tion algorithm
《重庆邮电大学学报（自然科学版）》 2024 (002)
268-276 / 9
国家自然科学基金项目(61971210) The National Natural Science Foundation of China(61971210)
评论