A novel acoustic-vibration capacitive microelectromechanical system microphone is designed, fabricated, and implemented in this paper. The new microphone consists of a rigid diaphragm and mass blocks sensitive to low-frequency vibratory and sound signals. This sensor takes advantage of the semiconductor technology to design the capacitance sensor structure by surface micromachining technology, and the inertial mass blocks are shaped using the bulk silicon micromachining technology. The structure of the anti-stiction-dimple array is designed and deployed at the bottom of the diaphragm and the backplate to avoid the risk of sensor failure by vibration stiction. The bottom and top of the backplate are designed with an anti-humidity hydrophobic insulation protective layer, which avoids adsorption of moisture and attachment of foreign particles. The thickness of the mass blocks can be controlled by the combination of the dry and the wet micromachining method, which is sensitive to different frequency responses. This study can design and produce wafer level silicon with high consistency. The inertial mass proposed in this research can be achieved through a 6 in. wafer process with >80% consistency when the thickness of the mass is around 300 μm. The sensing frequency can be reduced to below the 4 kHz frequency bandwidth with enhanced sensitivity in the ±0.5 dB range. Typical characteristic results show that the open-circuit sensitivity of the microphone is 12.63 mV/Pa (37.97 dBV/Pa) at 1 kHz (with 94 dB as the reference sound level). The total harmonic distortion and acoustic overload point are 0.21% and 121.2 dB sound pressure level, respectively. The electronic stethoscope is a typical application of this research, which can collect the characteristics and frequency spectrum of low-frequency cardiac vibration signals.

Long-term monitoring for patients can improve patient safety, help doctors diagnose and evaluate the clinical situation. Limited manpower in hospitals makes it difficult to achieve continuous and nuanced monitoring. In this paper, we classify the patient's posture as standing, sitting, lying and falling. Using a non-intrusive, privacy-compliant lidar, a medical human pose dataset is collected in the First Affiliated Hospital, Sun Yat-Sen University, and a novel computer vision-based approach is presented to continuously detect patients pose and provide timely information to health care workers. The recognition accuracy reaches 93.46% and the recognition speed reaches 42FPS on 3080Ti. Experimental results show that the algorithm performs well on the medical human pose dataset, and can effectively solve the problem of human pose recognition in medical scenes.