High-resolution radar imaging of moving humans using doppler processing and compressed sensing

摘要

Frontal radar imaging of human activities may be useful in certain applications, such as through-wall surveillance, where cameras and x-ray sensors cannot be deployed. High-resolution radar images are currently obtained using electrically large antenna apertures operating at high frequencies. However, high frequencies are heavily attenuated by most walls. Also, the implementation of a radar with lots of array elements and associated data acquisition channels is costly and complex. In this paper, we propose methods to generate high-resolution Doppler-enhanced radar images of moving humans at low carrier frequencies with limited number of antenna elements. When a human moves, different body parts give rise to distinct Doppler returns. The key feature of our method is to dynamically resolve multiple body parts of the human across three dimensions: Doppler, azimuth, and elevation. The additional Doppler dimension allows us to relax the resolution requirements in terms of the carrier frequency and number of array elements across the other dimensions. We further reduce the number of array elements below the Nyquist limits by incorporating compressed sensing principles into two-dimensional beamforming because compressed sensing is particularly suited for solving certain types of under-resolved problems.We test our technique on simulated electromagnetic radar scattered data from a moving human for different radar configurations.We also study the robustness of the proposed technique to noise.