Abstract
Energy consumption has become a critical issue for a portable and wireless embedded device because of its limited energy capacity of batteries. An application that runs on such a system often imposes real-time constraints that a task must complete its execution before its deadline. In this paper, we propose a low-power real-time scheduling algorithm called LPRM for fixed-priority periodic scheduling. LPRM, based on the technique of dynamic Voltage Scaling, is a low-power extension of the Rate-Monotonic algorithm. Due to its simplicity and low complexity, LPRM can be easily implemented in a power-aware real-time embedded system for energy reduction. To demonstrate its performance impact, we conduct a set of experiments to compare energy reduction of LPRM with other approaches. The experimental results show that LPRM, with a simple extension, achieves better energy reduction. Particularly, compared with the static speed approach, LPRM reduces energy consumption by 45% for a set of 25 tasks with variable execution times.