无线传感器网络异步任务集双效节能延迟调度算法研究

深亚微米技术的发展,使得漏电功耗在CMOS电路总功耗中所占比重日益增大,传统的传感器节点CPU节能研究主要针对动态功耗,其能耗估计和优化方法已凸显局限.针对此问题,提出动态电压调节(DVS)和动态功耗管理(DPM)相结合的双效节能延迟调度算法.从相对截止期小于等于周期的异步实时任务调度出发,结合DVS技术,综合考虑动态功耗和漏电功耗的影响,在满足任务实时性的前提下,选取每个任务的CPU执行速度,以降低总能耗,并通过任务的延迟调度对CPU空闲时段加以合并,采用DPM方法使CPU在空闲时段有选择性的进入低功耗状态,从而进一步降低漏电能耗.仿真实验验证了该算法的有效性.

Dual Power-Aware procrastination scheduling Algorithm for Asynchronous Task Sets in Wireless Sensor Network

The development of deep-submicron CMOS technology results in the fact that the leakage power dissipation becomes comparable to dynamic power consumption. Most traditional low power technologies to reduce the dynamic energy consumption of sensor processor without considering the leakage power are less effective to reduce the overall energy consumption. In this paper, an algorithm that combines the dynamic voltage scaling(DVS) and dynamic power management(DPM) is proposed for asynchronous periodic real-time tasks with specified relative deadlines to decrease the overall power dissipation. Without causing any deadline miss, processor speed for each task is selected by using DVS technology. Then, the speed will be modulated to balance the dynamic and leakage energy consumption by comparing with the critical speed which minimizes the dynamic and leakage energy consumption per unit work. The higher-than-necessary speeds can result in some idle intervals. We present a procrastination scheduling scheme to merge these scattered intervals into larger ones to switch CPU to low power state by using DPM technology. The simulation experiment shows that this algorithm can decrease the CPU power effectively.