Available Power Analysis for Background Tasks on Ubiquitous Sensor and Actuator Networks

Abstract

This paper first presents the sensor data processing framework integrating a power consumption scheduler for control actions, and then analyzes the available power for background tasks which are continuously performed using the remaining power. As the scheduler is designed to reduce the peak load for the given task set, it can make flatter the consumption level of dynamic tasks activated by the event detection. The experiment result obtained from the prototype implementation reveals that the proposed scheme can improve the minimum available power by up to 17.9 %, compared with the Earliest scheduling scheme, having a smaller standard deviation over the slots included in a scheduling window. Hence, our framework can achieve cost savings as the sensor network can purchase a smaller amount of energy from a utility company and can be built with lower power transmission capability.