安卓智能手机功耗管理机制及评估模型研究

发布时间：2018-03-08 04:31

本文选题：安卓　切入点：功耗管理　出处：《西安电子科技大学》2014年硕士论文　论文类型：学位论文

【摘要】：移动终端的多元化及社交广泛性为互联网发展带来一场深刻的变革,智能手机作为移动终端产品的代表,在市场上占据主导地位。根据市场分析,80%以上的智能手机运行安卓系统。随着用户体验的要求不断提高,理解和优化智能手机的功耗已成为一个重要的研究领域。现有的工作已经提出多种功耗模型来评估功耗,但这些模型缺乏细化粒度和精确度。本文对安卓系统自身的电源管理机制和功耗模型构建进行了系统的研究,所取得的主要研究成果如下:1.对安卓系统的电源管理机制进行研究,分析安卓电源管理模块的源码实现,特别是与其密切相关的唤醒锁机制、功耗管理以及评测机制。结果表明安卓系统不能统计应用程序或硬件模块在特定运行时间内的具体功耗值。2.针对安卓系统自身功耗管理机制的不足,提出基于硬件使用率和延时功耗的动态功耗评估模型,模型的构建基于非线性回归结构,首先通过模块化智能手机各硬件组件的相关参数来确定模型,然后设计功耗测试用例用于模型辨识,确定各个功耗相关系数,给出模型具体的关系表达式。模型具有以下优势:重点提出了无线通讯设备具有延时功耗的问题,建立了相应的延时功耗模型;将Flash、Camera、Bluetooth硬件组件加入到手机功耗模型中,完善了已有的基于硬件使用率的功耗模型。因此,该模型可达到改善整体功耗评估精度的目标。3.采用当前主流的智能手机设备进行模型验证,结果表明常用场景下该模型实测功耗和评估功耗之间的平均绝对误差都小于4.6%。依赖于上述模型,我们开发了一款安卓功耗监测工具,可以实时显示系统功耗变化趋势,实时统计安卓各个硬件组件和整体的功耗。
[Abstract]:The diversity and social universality of mobile terminals have brought a profound change to the development of the Internet. Smartphone is the representative of mobile terminal products. More than 80 percent of smartphones run Android, according to market analysis. Understanding and optimizing the power consumption of smart phones has become an important research field. However, these models lack of granularity and precision. In this paper, the power management mechanism and power model construction of Android system are systematically studied. The main research results are as follows: 1.The power management mechanism of Android system is studied, and the source code implementation of Android power management module is analyzed, especially the wake-up lock mechanism, which is closely related to Android power management module. The results show that the Android system can not count the specific power consumption value of the application or hardware module in a given running time. A dynamic power evaluation model based on hardware usage and delay power consumption is proposed. The model is constructed based on nonlinear regression structure. Firstly, the model is determined by the parameters of each hardware component of a modular smart phone. Then the power test cases are designed for model identification, each power consumption correlation coefficient is determined, and the specific relational expression of the model is given. The model has the following advantages: the problem that wireless communication equipment has delay power consumption is put forward emphatically. The corresponding delay power model is established, and the Flash Camera-Bluetooth hardware component is added to the mobile phone power model, which improves the existing power consumption model based on hardware utilization. The model can achieve the goal of improving the accuracy of overall power evaluation. The results show that the average absolute error between measured and evaluated power consumption is less than 4.6 in common scenarios. Depending on the above model, we have developed an Android power monitoring tool, which can display the power trends of the system in real time. Real-time statistics of Android hardware components and the overall power consumption.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN929.53;TP316

【相似文献】