Signal Processing Wall Animate Body Detection

Through-the-wall vital sign detection is a technique that takes advantage of the penetrability of radar to detect vital signs of humans or animals, such as heartbeat and respiration, on the other side of walls or other barriers. This technique can be widely applied in emergency rescue, health monitoring, military, antiterrorist mission and customs inspection. Through-the-wall vital sign detection radar always has to operate in complex environments full of various kinds of interference. Therefore, precisely extracting the vital signal using advanced signal processing methods becomes the core problem of this technique.Continuous wave Doppler radar for through-the-wall vital sign detection suffers from null point problem and distortion problem induced by single channel selection linear approximation demodulation method. Therefore, AT (Arctangent) demodulation method is discussed in this thesis. The accuracy of this method can be affected by DC offset resulted from stationary objects clutter and circuit. To solve this problem, AT demodulation method proposed in this thesis firstly filters all the DC components in the baseband signal, and then recovers the expected DC components using a circle centre estimation method, which combines LM algorithm with Pratt algorithm. Simulation result demonstrates the effectiveness of the AT demodulation method.Vital signal is heavily interfered by random physical movements of human and radar sensor motion. To this end, this thesis applies CEEMD (Complete Ensemble Empirical Mode Decomposition) method to extract heart signal from fidgeting interference. Then Ililbert transform is applied to analyse the instantaneous frequency of heartbeat. Simulation results illustrate that CEEMD method can remove fidgeting interference with frequency close to that of heartbeat, and Hilbert transform can obtain the instantaneous frequency of heartbeat precisely.At last, experiment is carried out using a35GHz millimetre wave continuous wave Doppler radar system to detect human behind a hard paper board. The interfered heartbeat signal is processed by Hilbert transform based on CEEMD. The heartbeat signal is successfully extracted from the interfered signal resulted from human body movements, and the instantaneous frequency of heartbeat can be monitored as well.