| Summary: | 碩士 === 大同大學 === 機械工程學系(所) === 107 === Fingerprint identification technology is one kind of biometric identification technology and has been widely used for personal identification in mobile devices to ensure personal privacy and quickly access personal data. Compared to optical and capacitive fingerprint sensors, ultrasonic fingerprint sensor has higher resolution and less sensitive to stain. Therefore, this thesis aims to study the characteristics of ultrasonic fingerprint sensors using finite element analysis (FEA) for reducing experimental time and development cost.Firstly, the FEA software, COMSOL Multiphysics, is used to calculate the voltage difference of the sensor between finger valley and ridge. The optimal operating frequency is analyzed by the periodic structure setting, and the feasibility and correctness of the simulation tool is verified by comparing with the experimental results. The influence of the adhesive film thickness on the recognition performance is also studied. Then, the full-scale sensor is simulated for exploring the identification correctness on the center position and width of fingerprints. The results show that the identifiable range of the sensor is 5 mm in the middle region, and the error of the fingerprint center position and width is less than 50 µm. Finally, the acoustic impedance matching layer is conducted to improve the recognition performance of an ultrasonic fingerprint sensor. The optimal acoustic impedance of the matching layer is calculated based on the one dimension wave propagation theory and simplified FEA model. In addition, ultrasonic fingerprint sensors with several kinds of acoustic impedance matching layers are also analyzed. The results show that the acoustic impedance matching layer can indeed improve the recognition performance of an ultrasonic fingerprint sensor.
|
|---|
