Research On Low Power Model For Embedded System

As the technology of system intergration keeps on moving forward, the functionality of embedded system becomes more and more powerful, but the power is becoming one of the first constraints of embedded system design. Power reduction in embedded system can be realized by static and dynamic techniques. However, static techniques can not satisfy the system requirements on power consumption with the improvement on system function. Recent research has concerntrated on dynamic power reduction techniques, which include DPM and DVS.The main contributions of this dissertation are as follows:1. We analyze the main principle of the dynamic power management (DPM), study three kind of DPM strategy: The Time-out strategy, the forecast strategy and stochastic decision model.2. We study the technology of Dynamic Voltage Scaling(DVS). We study the non-real-time DVS algorithm, with real-time DVS algorithm. We study FLAT, PAST, PEAK, the AVGn algorithman and compare with various algorithms good and bad points.3. We propose a novel approach to solve the problem of saving electricity energy in embedded real-time system. Observing that there are many types of running states in the adjusting microprocessors and network priority exists between the tasks from the task set, we construct an accurate mathematical model for this problem. Then we present a RBEDF algorithm to sort the task sets.4. In view of microprocessor with the many kinds of running status characteristic, we propose a novel approach to solve the problem of how to set the voltage for the task sets. We subdivide the task sets into small subsets. After the subdivision, we specify the work voltage by converting each small subset to the knapsack problem.