Real-Time Scheduling of Sensor and Actuator Networks

Abstract

Real-time scheduling issues in sensor and actuator networks (SANs) was investigated. The research studies the real-time issues related to the interplay between acquisition of data from sensors, its use in controller development, and optimal, real-time task scheduling based on available computing resources. Real-time task scheduling with the goal of achieving optimal system performance was the focus. Task schedulability with higher task rates and resource efficiency was investigated. Consideration of task execution times less than the worst case execution times results in achieving higher task frequencies. The problem of handling overruns using local (a task overrun reduces only that particular task frequency) and global (a task overrun initiates reduction of all task frequencies) approaches was investigated. A bisection method is proposed that has the potential to achieve higher task rates and resource efficiency. A number of examples are used to illustrate the ideas and methods.