Smart sensing is used to provide specific information by using physical quantities measured by multiple sensors or multi-modal sensor modules. Nowadays, the smart sensor is configured as a compact system using very small transducers implemented by MEMS technology. Accordingly, the multi-modal sensing device is getting miniaturized. A microphone array can be classified as a smart sensing device that estimates the sound field information. However, it is difficult to implement a compact system because the measurable frequency range is limited by the spacing between adjacent microphones. Also, the calibration is difficult due to the irregularly shaped sensor element; therefore, the traceability is not yet established. In this study, sound source localization algorithms are applied for implementing the compact microphone array. The inherent error due to the sensor position tolerance, magnitude difference, and phase mismatch is analyzed. Because the coherent noise increases uncertainty in the actual test environment, the estimation error depending on the critical distance is analyzed. Finally, a compact webcam-based multi-modal sensor is implemented and the localization test result is reviewed.
Read full abstract