A Study On Low Power Consumption Optimizations In Embedded System Based On DPM And DVFS

With the rapid development of computer and mobile internet technologies, embedded mobile terminals (such as mobile phones and pads) have become an indispensable part in our daily life. Embedded system devices, especially portable mobile devices, are powered from batteries which are limited in size and weight. Therefore power becomes an important challenge to embedded system devices’design.There are a variety of methods to reduce the power consumption of embedded system. Dynamic Power Management (DPM) and Dynamic Voltage and Frequency Scaling (DVFS) are two mainly key technologies reducing the power consumption of embedded system. In this paper, two strategies DPM and DVFS are used to optimize and reduce the power consumption of mobile embedded system.DPM changes device’s working status primarily based on the workload on it in order to achieve the minimum power consumption. For example, shut down the peripheral device or make it go to sleep mode to reduce power consumption when the idle time is long enough. This paper first analyzes the power management mechanism of Google Android from Application level, Application Framework level and Linux Kernel level. Then this paper makes an experiment about Android network module in both the small flow mode (10Kib/min) and big flow mode (2Mbit/min) and obtains the conclusion that Android network module plays an important role in the system power consumption. In the end this paper presents a solution to optimize the power management of Android network module dynamically. This solution implements dynamic power management of network module on Android system considering four factors:network traffic, battery power, time and whether is rush hours. Form the experiments we can see that the power consumption of network module (GPRS/WIFI) plays an important role in the total system power consumption. Therefore this DPM solution has great significance in improving the energy utilization of mobile terminals and reducing the power consumption of mobile terminals.DVFS is another important way to reduce the power consumption of embedded system. DVFS scales the CPU clock frequency and CPU power supply voltage dynamically based on the workload of running tasks to achieve the tradeoff of tasks’ deadline and system power consumption. The processors of mobile terminals commonly have the ability of Dynamic Voltage and Frequency Scaling. How to effectively using the DVFS ability which is provided by the CPU to reduce the power consumption of mobile terminals in the operating system has become a deserving research point. This paper presents an optimized and fine grained DVFS algorithm and realizes it on Android operating system based on linear prediction aiming at mobile operating systems based on Linux kernel. This algorithm uses the linear prediction mechanism based on historical data to adjust the range and orientation of voltage and frequency. The parameters are obtained through regression method. Experimental results show that the optimized DVFS strategy could save13.55%energy compared to the default coarse grained frequency scaling strategy. So the optimized DVFS strategy could reach a better power saving effect and it enhances the cruising ability of mobile terminals.