An Adaptive Filtering Technique for Driver's Heart Rate Monitoring through Vibration Signal by Seat-Embedded Piezoelectric Sensors

Abstract

There are many situations in which a biophysical measurement brings useful information for the safety of human-machine systems. However, in several cases, the measurement signal can become contaminated with environmental noise. When this noise is severe, the signal processing algorithm is needed to extract the important information from the noisy signal. In this paper we consider the driver's heart rate monitoring problem to decide the drowsiness level for safety driving. To detect the heart rate of the driver, we adopt seat-embedded piezoelectric sensors. This sensor measure the body vibration caused by the heart beat but the signal also contains the large amount of car body vibration according to the road conditions. For this problem, we propose a heart rate detection system based on the model based on-line adaptive filtering technique. The proposed system has a simple structure compared with offline time-series model (ARX model) based system and similar performance.