本文選題：動(dòng)態(tài)電源管理 + 嵌入式測控系統(tǒng)　； 參考：《哈爾濱工業(yè)大學(xué)》2017年碩士論文

【摘要】：隨著電子系統(tǒng)功能的增強(qiáng)以及電路集成度不斷提高,電子系統(tǒng)的能量消耗也在迅速增加,而過高的功耗會(huì)影響到電子系統(tǒng)的使用壽命。因此為了提高電子系統(tǒng)可靠性,如何在保證系統(tǒng)性能的同時(shí)優(yōu)化電子系統(tǒng)的功耗成為了設(shè)計(jì)者面臨的關(guān)鍵問題。動(dòng)態(tài)電源管理可以從不同層次對(duì)嵌入式系統(tǒng)的功耗進(jìn)行優(yōu)化,成為了嵌入式系統(tǒng)低功耗設(shè)計(jì)的熱點(diǎn)。本文主要從系統(tǒng)級(jí)和體系結(jié)構(gòu)級(jí)對(duì)嵌入式系統(tǒng)進(jìn)行動(dòng)態(tài)電源管理的研究,根據(jù)嵌入式系統(tǒng)不同的運(yùn)行狀態(tài),分別采用結(jié)合動(dòng)態(tài)頻率調(diào)節(jié)技術(shù)的節(jié)能任務(wù)調(diào)度算法和動(dòng)態(tài)功耗管理策略對(duì)系統(tǒng)功耗進(jìn)行優(yōu)化。本文的硬件平臺(tái)為基于ZYNQ的嵌入式測控系統(tǒng),文中對(duì)該平臺(tái)劃分了不同的低功耗工作模式,并建立了功耗管理模型。本文通過分析嵌入式系統(tǒng)執(zhí)行任務(wù)過程中采用靜態(tài)優(yōu)先級(jí)調(diào)度算法和動(dòng)態(tài)優(yōu)先級(jí)算法的優(yōu)缺點(diǎn),在Linux系統(tǒng)中實(shí)現(xiàn)了算法復(fù)雜度低且性能較穩(wěn)定的RM調(diào)度算法,并結(jié)合動(dòng)態(tài)頻率調(diào)節(jié)技術(shù),對(duì)任務(wù)執(zhí)行過程中的功耗進(jìn)行優(yōu)化。針對(duì)RM算法調(diào)度過程中具有相同優(yōu)先級(jí)的任務(wù)可能無法及時(shí)公平地得到響應(yīng)的弱點(diǎn),本文設(shè)計(jì)了RM調(diào)度算法與時(shí)間片輪轉(zhuǎn)調(diào)度算法相結(jié)合的改進(jìn)算法,有效降低了任務(wù)的平均周轉(zhuǎn)時(shí)間。當(dāng)系統(tǒng)暫時(shí)沒有任務(wù)處于就緒隊(duì)列時(shí),Linux系統(tǒng)會(huì)執(zhí)行空閑進(jìn)程,進(jìn)入空閑模式。由于頻繁地模式切換會(huì)產(chǎn)生額外的能量消耗,針對(duì)這一情況,本文通過對(duì)Linux內(nèi)核中功耗管理模塊的研究與修改,設(shè)計(jì)并實(shí)現(xiàn)了采用雙超時(shí)閾值Timeout算法的動(dòng)態(tài)功耗管理策略,兩個(gè)超時(shí)閾值的設(shè)定是為了根據(jù)空閑時(shí)間的長度來控制系統(tǒng)的正常工作模式跟空閑模式和休眠模式之間切換的時(shí)機(jī),在硬件平臺(tái)測試中表明改進(jìn)的動(dòng)態(tài)功耗管理策略使得系統(tǒng)模式切換次數(shù)有所降低。在測試中表明本文設(shè)計(jì)的動(dòng)態(tài)電源管理方案能夠在保證系統(tǒng)性能的同時(shí),有效降低系統(tǒng)的功耗。
[Abstract]:With the enhancement of the function of electronic system and the improvement of circuit integration, the energy consumption of electronic system is increasing rapidly, and the high power consumption will affect the service life of electronic system. Therefore, in order to improve the reliability of electronic system, how to optimize the power consumption of electronic system while ensuring the system performance has become a key problem for designers. Dynamic power management can optimize the power consumption of embedded system from different levels and become the focus of low power design of embedded system. This paper mainly studies the dynamic power management of embedded system from the system level and architecture level, according to the different running state of the embedded system, The energy saving task scheduling algorithm and the dynamic power management strategy are used to optimize the power consumption of the system. The hardware platform of this paper is an embedded measurement and control system based on ZYNQ. This paper divides the platform into different low power working modes and establishes a power management model. This paper analyzes the advantages and disadvantages of static priority scheduling algorithm and dynamic priority scheduling algorithm in the execution of embedded system. The RM scheduling algorithm with low complexity and stable performance is implemented in Linux system. Combined with dynamic frequency regulation technology, the power consumption during task execution is optimized. Aiming at the weakness that tasks with the same priority in RM scheduling process may not be responded in a timely and fair manner, an improved algorithm combining RM scheduling algorithm with time slice rotation scheduling algorithm is designed in this paper. Effectively reduces the average turnaround time of the task. Linux systems execute idle processes and enter idle mode when there is no task in the ready queue for the time being. Because of the extra energy consumption caused by frequent mode switching, this paper designs and implements a dynamic power management strategy based on the dual-timeout algorithm through the research and modification of the power management module in the Linux kernel. The two timeout thresholds are set to control the timing of switching between the normal working mode of the system and the idle and dormant modes according to the length of the idle time. In the hardware platform test, it is shown that the improved dynamic power management strategy can reduce the number of system mode switching. The test results show that the dynamic power management scheme designed in this paper can effectively reduce the power consumption of the system while ensuring the performance of the system.
【學(xué)位授予單位】：哈爾濱工業(yè)大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：TN86

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 張豪;王益涵;吳飛;;計(jì)算機(jī)節(jié)能技術(shù)綜述[J];計(jì)算機(jī)技術(shù)與發(fā)展;2016年01期

2 童義山;蘇開榮;劉解華;;VxWorks操作系統(tǒng)電源管理研究[J];無線通信技術(shù);2015年01期

3 李宏升;;嵌入式系統(tǒng)電源管理架構(gòu)分析與設(shè)計(jì)[J];網(wǎng)絡(luò)安全技術(shù)與應(yīng)用;2015年01期

4 陳運(yùn)文;吳飛;劉博;;智能終端的節(jié)能控制策略研究[J];上海工程技術(shù)大學(xué)學(xué)報(bào);2014年04期

5 劉彩霞;;提高軍用手持機(jī)續(xù)航能力設(shè)計(jì)[J];計(jì)算機(jī)與網(wǎng)絡(luò);2014年13期

6 張小來;;便攜衛(wèi)星通信終端的動(dòng)態(tài)電源管理設(shè)計(jì)[J];電子世界;2013年14期

7 曹哲;尤政;;超時(shí)策略動(dòng)態(tài)閾值的閾值選擇影響因素[J];哈爾濱工業(yè)大學(xué)學(xué)報(bào);2013年06期

8 孔繁鑫;林宇晗;徐惠婷;任碧巖;鄧慶緒;;單處理器多外設(shè)實(shí)時(shí)系統(tǒng)全局能耗優(yōu)化算法研究[J];小型微型計(jì)算機(jī)系統(tǒng);2013年04期

9 李偉生;王冬;;改進(jìn)自適應(yīng)學(xué)習(xí)樹電源管理預(yù)測策略[J];計(jì)算機(jī)工程與設(shè)計(jì);2013年01期

10 卜愛國;;基于Markov模型的動(dòng)態(tài)電壓調(diào)節(jié)策略[J];計(jì)算機(jī)應(yīng)用研究;2011年10期

相關(guān)博士學(xué)位論文 前2條

1 江琦;半Markov切換空間控制過程及其應(yīng)用[D];中國科學(xué)技術(shù)大學(xué);2008年

2 王保進(jìn);嵌入式實(shí)時(shí)系統(tǒng)的任務(wù)調(diào)度與資源共享模型及算法研究[D];中國人民解放軍信息工程大學(xué);2005年

相關(guān)碩士學(xué)位論文 前1條

1 李曄;基于Cortex-A8架構(gòu)嵌入式開發(fā)板的設(shè)計(jì)與實(shí)現(xiàn)[D];華北電力大學(xué);2014年

，

本文編號(hào)：2037937