Research On Dynamic Power Management Of Data Acquisition System Based On Embedded Platform

In the increasingly developing age of embedded system's System-On-Chip technology,raising the system integration density also brings about the increasing of system power consumption and temperature.This problem directly leads to the decline of integrated system's working reliability and service life,thus it limits the development of the integrated technology.Therefore,how to optimize the system power consumption as well as temperature and achieve the lowest power consumption without affecting its performance,is a key issue in the embedded integrated systems at present.The object of this paper is to study the dynamic power management of long-term charged integrated embedded systems through a real-time operating system.The system platform is built based on ZYNQ 7000.Then,the real-time data acquisition system for the voltage and temperature of the ZYNQ So C chip is established.Through the analyzing of the working process of ZYNQ,the working modes are designed.In the design of low-power strategy technology,through the comparison of voltage regulation and frequency adjustment,dynamic frequency adjustment technology was selected to optimize power consumption.Also,a frequency adjustment model was established for analysis.According to the comparison as well as analysis of timeout algorithm,exponential average prediction algorithm and random algorithm based on Markov chain,Timeout algorithm is selected as the basis of low-power strategy.This paper transplants Free RTOS real-time operating system in ZYNQ platform to realize the scheduling of system tasks.Through the analysis and comparison of the task scheduling algorithm of Free RTOS,?C/OS-II and RT-Thread operating system,the respective advantages and disadvantages of these algorithms are found.Then,it also founds that there is a problem in the task scheduling of Free RTOS.Under the condition of tasks in the ready state having large differences in the task priority,searching the necessary task priority takes a long time.Afterward,according to the ready-table lookup algorithm,the algorithm of searching the necessary task priority is improved.The improved algorithm is able to reduce the time of Free RTOS task scheduling effectively when tasks in the ready state having large differences in the task priority.So it reduces the power consumption of embedded integrated system consumed in the case of long-term charging.Through the test on the hardware platform,the low-power strategy as well as arithmetic improvement results are verified,and 28 percent power consumption can be reduced in this plan.