一种基于用户行为的嵌入式功耗优化方法  

Embedded Power Optimization Method Based on User Behavior

作　　者：王海[1] 高岭[1] 陈东棋[1] 任杰[1] 

机构地区：[1]西北大学信息科学与技术学院,西安710127

出　　处：《系统仿真学报》2015年第2期320-326,共7页Journal of System Simulation

基　　金：国家自然科学基金(61373176);教育部博士点基金(20116101110016);陕西省自然科学基金(2012JQ8047);陕西省工业攻关项目(2014K05-42)

摘　　要：随着以手机、平板电脑等为代表的嵌入式设备的飞速发展,嵌入式低功耗问题已经成为了一个研究热点。经常有用户在户外面临手机没电所带来的问题,目前主要给嵌入式设备供电的电池却受到其体积、重量等因素的制约,只能提供有限的电量。从用户行为识别出发,用机器学习的方法识别设备用户当前的行为状态,根据既定的策略来获取该行为状态下用户对设备的使用习惯,进而通过主动关闭设备不使用的部件或者提醒用户应用动态交互优化策略来降低设备功耗。In recent years, with the rapid development of embedded device represented by mobile phone and tablet computer, low power technology has been one of the hotspots in the embedded research field. Because the battery capacity of embedded device is limited due to its restricted volume and weight, there are often users suffering the problem that their phone battery being dead. There are many research directions in embedded low power field at present. The relationship between low power and user behavior recognition was aimed, which started with recognizing user behavior using machine learning and then obtains the user's daily usage habits in specific behavior. Part of device components could be turned off or Dynamic Interactive Optimize Strategy was applied to reduce the power consumption.

关 键 词：嵌入式系统 低功耗 行为识别 机器学习 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TP368.1[自动化与计算机技术—控制科学与工程]