Short-Time State-Space Method for Micro-Doppler Identification of Walking Subject Using UWB Impulse Doppler Radar

• Published in: IEEE transactions on microwave theory and techniques Vol. 66; no. 7; pp. 3521 - 3534
• Main Authors: Ren, Lingyun, Tran, Nghia, Foroughian, Farnaz, Naishadham, Krishna, Piou, Jean E., Kilic, Ozlem, Fathy, Aly E.
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.07.2018; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Body parts; Broadband; Computer simulation; Doppler effect; Doppler radar; Electromagnetic (EM) scattering model; Feature extraction; Fourier transforms; Gait; Kinematics; Legged locomotion; Mathematical model; micro-Doppler (<italic xmlns:ali="http://www.niso.org/schemas/ali/1.0/" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">μ -D) signatures; Radar echoes; Radar signatures; Radar targets; short-time Fourier transform (STFT); short-time state-space method (STSSM); Sliding; Torso; Ultra wideband radar; ultra-wideband (UWB) radar; Ultrawideband radar; Walking; Wideband communications; Windows (intervals)
• ISSN: 0018-9480; 1557-9670
• DOI: 10.1109/TMTT.2018.2829523

In this paper, ultra-wideband (UWB) Doppler radar signatures from walking human subjects are processed with the state-space method (SSM) to analyze compressed radar echoes for the first time. A new mathematical model is proposed to detect the human walking strides using UWB radar, while micro-Doppler (<inline-formula> <tex-math notation="LaTeX">\mu </tex-math></inline-formula>-D) features are extracted for gait analysis using short-time SSM (STSSM). To distinguish <inline-formula> <tex-math notation="LaTeX">\mu </tex-math></inline-formula>-D signatures of different subjects' body parts, the SSM used for the characterization of radar target signatures and sensor fusion is applied on a sliding short-time window to enhance the resolution of feature extraction from data collected on a dismount. This method of the application of SSM to sliding short-time data is validated with a full-wave electromagnetic (EM) simulation of a walking subject using the Boulic model that represents the human kinematics. An EM scattering model is then utilized to compare the performance of a short-time Fourier transform with STSSM. Experimental results show that STSSM can be successfully applied to identify multiple <inline-formula> <tex-math notation="LaTeX">\mu </tex-math></inline-formula>-D trajectories in real experimental data, thus demonstrating the capability to positively identify human motions (right foot, left foot, and torso) even in a low signal-to-noise ratio environment.