Event-Driven Sensing System with PID-Type Data Volume Control Function

We have been studying a wearable sensor network composed of an original event-driven sensing system and multiple sensors for human motion detection. To detect human behaviors in detail, multi-point detection with high time resolution and suppression of data congestion is required. For achieving such a sensing system, this paper proposes a PID-type (Proportional-Integral-Derivative) data volume control function that variably increases and decreases data volume of each sensor node in response to large motions, continuous motions, and sudden motions of human. Basic verification using a loading stage and application verification using a wearable sensor glove showed that the sum of proportional, integral, and derivative values of sensor data changes in response to large motion, continuous motion, and sudden motion, and that the date volume changes during sensing. The effectiveness of the PID-type data volume control function was demonstrated.