Title :
Near-field interactions for micro-Doppler signature of human motion in forest using FMM on hybrid platforms
Author :
Tran, Nghia ; Dang, Vinh ; Kilic, Ozlem
Author_Institution :
Dept. of Electr. Eng. & Comput. Sci., Catholic Univ. of America, Washington, DC, USA
Abstract :
This paper utilizes the parallelization of Fast Multipole Method (FMM) on hybrid CPU-GPU high-performance computing platforms to analyze the near-field human-forest interactions for micro-Doppler signature of human walking in a forest environment. With the aid of the hybrid platform, larger problem sizes that can accommodate higher order of mutual coupling effects can be solved for.
Keywords :
Doppler radar; electrical engineering computing; graphics processing units; Doppler radar; FMM; fast multipole method; human motion; human walking; hybrid CPU-GPU high-performance computing platform; microDoppler signature; mutual coupling effect; near-field human-forest interaction analysis; Animals; Antennas; Dielectrics; Doppler effect; Graphics processing units; Legged locomotion; Mutual coupling; Fast Multipole Method (FMM); forest scattering; hybrid CPU-GPU systems; micro-Doppler signature;
Conference_Titel :
Applied Computational Electromagnetics (ACES), 2015 31st International Review of Progress in
Conference_Location :
Williamsburg, VA
