Seismic Propagation from Humans in Open and Urban Terrain

Human movement produces dynamic forces that cause vibrations and seismic waves. Recent research illustrates that human movement can be recognized by measuring and analyzing mechanical responses to the movement, providing a new sensor modality to support established modes of human detection and movement tracking. Our objective is a high-performance-computing capability for simulating seismic waves propagating from walking and running humans on open terrain, on urban ground, and within buildings, to support development of human detection and signature analysis algorithms for seismic sensors. We apply viscoelastic finite-difference time-domain computations on a variable grid, using domain decomposition and MPI to operate in parallel on high performance computers. For walker/runner force inputs we develop an empirical model from published biomechanics research on human kinematics and measured ground reaction forces. Our open terrain results reveal the pulsating nature of seismic waves as they emanate from footsteps. We find that spectral characteristics of our simulated signals agree well with field signals. Our urban model results show realistic ground vibrations from runners outside of buildings, and our urban structures account for the effect of building vibrations on footstep signatures propagating from the inside to the outside of a building. We demonstrate the application of a seismic human detection algorithm using synthetic data, and conclude that the simulation method produces realistic wavefield data for virtual trials of human movement algorithms.