【摘要】 近年来,随着智能物联网应用的快速发展,雷达传感器由于具有保护隐私、全天候全天时工作、不受光线和遮挡影响等优点,在人体目标日常行为活动识别方面受到了学术界和产业界的极大重视。针对一种超低辐射的超宽带雷达(Impulse Radio Ultra-Wideband,IR-UWB),提出了一种室内人员日常活动(包含静止、坐下、走路、起立)分类方法。该方法首先利用目标检测方法检测出目标有效距离单元;其次,提出了基于平均多普勒频率、信息量和多普勒能量的3种微多普勒特征进行动静目标粗分类;最后,采用长短期记忆神经网络(Long Short-Term Memory,LSTM)对人体活动进行细分类。实验结果表明,人体活动细分类的平均准确率能达到92.52%。更多还原

【Abstract】 Recent years have witnessed great development in intelligent Internet of Things.In the field of human daily motion recognition,radar sensor has been attached great importance by the academic and industrial circles with its unique advantages of privacy protection,all-weather working mode,operating free from light and occlusion.In this paper,a method is proposed for indoor daily motion classification including motionless,sitting down,walking and standing up based on an ultra-low impulse radio ultra-wideband(IR-UWB)radar.Firstly,a target detection method is used to detect the effective range bins.Secondly,the moving and static targets are classified based on three micro-Doppler features of average Doppler frequency,information entropy and energy of Doppler frequency.Finally,a long short-term memory(LSTM)neural network is applied to further classify the human activities.Experimental results show that the average accuracy of human motion classification will reach 92.52%.更多还原