Research on extraction of weak life signals based on multi-channel models

Through-wall radar detection technology is mainly oriented to targets moving at a speed of up to 3 m/s, where, the Doppler frequency shift resulting from the moving of the target usually falls within 20 Hz, and due to existence of walls and other barriers, the electromagnetic energy can be absorbed by the radar only after two-way attenuation in penetration of walls. Therefore, the echo features low energy and frequency. Due to presence of direct waves, there is a strong interference to the radar in DC fields. How to accurately extract and detect target echo in the presence of strong interference is the key to the project. This paper mainly introduces the composition of radar echo signals modulated by human movements (including movement of body, breathing, heartbeat, and other vital signs), proposes an algorithm for extraction of radar life signals based on multi-characteristic models according to the time-domain and frequency-domain difference between the target echo and interference noise, and verifies the effectiveness of the algorithm through tests on the sample machine.