Non-contact biometric identification and authentication using microwave Doppler sensor

As described in this paper, we propose a non-contact biometrie authentication method with heartbeat features measured using a microwave Doppler sensor. The heartbeat component is measured as personal characteristic information attributable to individual differences in the myocardium and blood vessels. Biometric authentication using electrocardiogram (ECG) or pulse wave has been proposed in reports of earlier studies, but such methods require direct contact of the sensor with the human skin. However, heartbeat information can be measured and authenticated without contact to the skin when using the proposed method. The microwave Doppler sensor can detect minute vibrations of the body surface caused by heartbeat. The salient difficulty of the microwave Doppler sensor is noise contamination such as that caused by body motion. This study introduces the use of time-frequency analysis with an autoregressive model to reduce the noise influence and emphasize heartbeat features. An ID generation and authentication algorithm using a frequency feature of the heartbeat component is proposed. The proposed method was evaluated using measurements taken of 11 participants. Measurement results show a 92.8% true acceptance rate and a 3.9% equal error rate.