Multiple Subject Respiratory Pattern Recognition and Estimation of Direction of Arrival using Phase-Comparison Monopulse Radar

A scalable non-contact respiratory monitoring approach is introduced for the practical separation and recognition of multiple simultaneous human respiratory patterns, using a 24-GHz phase-comparison Monopulse radar. The proposed system demonstrates a computationally efficient phase-comparison Monopulse technique to estimate the direction of arrival (DOA) of respiratory motion signals from single and multiple subjects to determine their angular positions. Experimental results demonstrated that Monopulse radar can sense and detect three different breathing patterns (normal, fast and slow) and can estimate the angular location of a single normally breathing subject within the main beamwidth of the transceiver with an estimation accuracy of approximately 84% whereas, and almost 78% for two subjects. Based on analysis of experimental results, the estimation accuracy of phase-comparison Monopulse technique degrades with the breathing pattern deviations from nominal breathing cycle and rate (fast/slow breathing) due to reduced phase measurement accuracy for the irregular signals resulting from lower signal to noise ratios (SNR). The proposed scalable 24-GIIz radar transceiver system can successfully find the direction of the human target based on recognition of respiration signals which can have potential applications in healthcare and security surveillance systems.