Abstract
Autonomic nervous system is important to maintain homeostasis, and respiratory sinus arrhythmia (RSA) is known as a selective index of cardiac vagal activity. In this paper, we evaluated errors in the amplitude of RSA under the condition of body motion (keyboard typing and mental arithmetic with touching panel) and proposed a method for reducing them. It was found that elastic chest band is suitable under the quiet condition, while thermistor is suitable under the condition of body motion. It was also found that Berger's interpolation method was the best for detecting instantaneous heartbeat intervals in real-time signal processing. Furthermore, we proposed a error reduction algorithm by mixing the data of thermistor and elastic chest band , and applied it for interactive CG (Computer Graphics) system that reflected the amplitude of RSA estimated in real-time.
