Disaster recovery and through rubble detection by means of quasi-real time UWB multilateration reconstruction

This paper presents a quasi-real time UWB image reconstruction method capable of detecting minute movements of victims trapped under rubble, including respiration movements. The approach is based on multi-static measurements of a C-shaped sensor array scanner arranged for a limited view imaging geometry where data is repetitively and continuously collected over small time intervals. Reconstruction is based on a multi-static multilateration method which correlates the movements captured by subtracting successive signals obtained from each transmitter-receiver pair into a point on the image, without the need of wall parameter estimations or priori information. Images processing requires 0.36 seconds on a standard desktop PC, offering nearly 3 frames/sec. Results based on several numerical experiments of varying complexity are presented and discussed.