Research And Application Of Low-power Technologies In Embedded System

Early research of the embedded systems usually focused on the performance, real-time and reliability. But in recent years with the rapid development of embedded technologies, the frequency and integration of embedded systems have become higher and higher. The use of embedded devices especially the mobile embedded devices (mobile phones, MP3, PDA, etc.) has put forward a new question for the designers of the embedded systems-how to effectively reduce the power consumption of embedded systems in order to extend the batter lifetime and increase the lifetime of the device. At present, the embedded low-power optimization methods have become the research focus attracting more and more scholars and designers of embedded systems. On certain occasions, the issue of power is even more important than other conditions such as performance and reliability and become the overriding consideration in the design of embedded systems.Embedded system consists of two parts that are embedded hardware and embedded software. So the power optimization technologies can be expanded in the hardware level and software level. Because the power consumption of embedded system is directly generated by the underlying hardware units, early studies mainly focused on the hardware level, including the update of the production materials and manufacturing process and optimization of circuit-level and micro-structural level. But in reality, the upper software drives the underlying hardware, for example different instruction execution and data access affect the underlying hardware circuit directly and result in different power generation. Modern research has proven that the power consumption of the embedded software is an important part of the whole embedded system. And collaborative optimization of the hardware and software can minimize the power consumption of embedded systems.In this paper Ⅰ first analyze the causes of power consumption in embedded systems, and then introduce the existing embedded power optimization technologies at home and abroad. I analyze the power optimization technologies of the software level in detail and build the instruction-level power model of C language based on EMSIM. Then use the source-level and algorithm-level software power optimization techniques to optimize the commonly used embedded real-time system μC/OS-II. The original and improved source codes of μC/OS-II are put in the EMSIM which has been installed on the Red hat Linux9.0. Through comparing the change of the power consumption and the execution time of the codes to analyze how the source code level optimization affect the performance and power consumption of the embedded systems. The final experimental data shows that optimization techniques in the software could reduce the power consumption of the whole system without significantly affecting the performance of the embedded system as the optimization techniques in the hardware level.