A Multi-target In-bed State Recognition Method Based on mmWave Radar

In this paper, in order to recognize the human activity relative to the bed in the bedroom, a multi-target in-bed state recognition method based on mmWave radar is proposed. Specifically, this paper first introduces the signal model of mmWave radar. Secondly, the radar echoes are processed to extract the distance, angle and point cloud information of the target. Then, under the strategy of cooperative measurement, the center point position of point cloud and the bed center position will be calculated through mean operation and statistical processing of 2-D histogram. Further, numerical features will be extracted, counted and provided to the SVM classifier for training. Then, the trained model is transferred to the in-bed state recognition task under multiple human targets for real-time recognition. Finally, experimental results verify the effectiveness of the proposed method.