Analysis of micro-Doppler signature due to indoor human motion using multilevel fast multipole algorithm on GPU cluster

Detecting and tracking human motion in indoor environments are essential for both commercial (vital sign detection of elderly) and military applications (counter terrorism). Small variations in the carrier frequency caused by motion can be detected by Doppler radar systems. The micro-Doppler frequency shift depends on the transmitted frequency and the velocity of the different body parts over time. Different types of motions can be identified and classified from micro-Doppler spectrograms. Due to its bipedal nature, human micro-Doppler signature can be differentiated from others, including those caused by four-legged animals.