Estimation of autonomic nervous activity toward affective human-robot interaction

The purpose of this study is to enable a communication robot to estimate emotion state of a person based on autonomic nervous activity toward human-robot `emotional' interaction. In this paper, a microwave Doppler sensor is used as a non-contact technique in order to measure heart rate variability (HRV) which reflects the activity of the autonomic nervous system. The preliminary experiment was conducted to confirm whether HRV of subjects can be precisely measured by analyzing heartbeat interval (R-R interval) from the signal obtained by the microwave Doppler sensor. The result showed that HRV was fluctuated according to subject's internal state during watching different genre of videos. The subsequent experiment was carried out to investigate the influence of subject's speaking on the output signal of the sensor, and we attempted to improve the extracting accuracy of heartbeat signal even when the subject was speaking. As the result, it was confirmed that the measurement of HRV using the microwave Doppler sensor has a potential to be applicable for estimating user's internal state in human-robot interaction.