Research On The Energy Efficiency And Power Optimization Strategy For Mobile Devices

In recent years,a large number of mobile devices emerge such as smartphones,tablet computers,traditional users experience the transition from computer terminal to mobile terminal.Android operating system is becoming more and more favored by the people,which attracts a large number of users in the field of intelligent terminal operating system for its open source,free and designed especially for mobile devices.However,many people often faced with the phenomenon that the battery life of device is becoming shorter and shorter while enjoying the high technology.To solve this problem,this issue analyses the shortage and the inadequate of the Android current power management and makes some improvement based on the further research on Android system,designing the corresponding optimization and improvement for energy saving at the same time.Firstly,this issue researches the wake lock mechanism in Android power management system.Through tracking the principle and the procedure of using the wake lock mechanism from user space to kernel layer,and detecting the silence-consume of energy while the screen of the device is powered off.On the basis of this,developing a mechanism called WLDealer to handle the wake lock bugs.WLDealer runs in the background of the system,it tracks the wake locks in active state when the screen is powered off,and judges the abnormal use situation of wake locks according to the process state information.WLDealer can handle the abnormal state of wake lock automatically without the participation of user.Secondly,the inadequate of the current power management mode on Android platform is analyzed,then a power management optimization system is designed and implemented.By collecting real context information and power consumption optimization strategies,and training the data with C4.5 algorithm in the Weka intelligent platform,then generates the power consume optimization model for users under different scenarios,which has a better classification performance than the Na?ve Bayes algorithm.And the main energy consuming modules of the equipment are optimized based on this model.In the end,the traditional DVFS strategy in Android system is researched,and the problem which caused extra consumption of energy in the process of adjusting the frequency of CPU is also analysed.To solve this problem,a DVFS optimization scheme is designed,which by acquiring and analysising the real time load information,then adjusting the CPU to the appropriate operatingfrequency promptly which taking into account both of the performance and power consumption which avoid the problem of power consumption on CPU in the traditional DVFS process.By testing on the real terminal equipment,the results show that WLDealer can effectively deal with the abnormal use of the wake locks in the system when the screen is powered off,reducing the phenomenon of silence-consume of energy in the system at the same time.And the intelligent power management system has achieved a good performance in the optimization of energy saving,which greatly improve the energy efficiency of the system.The proposed DVFS optimization scheme achieves better energy saving effect than the traditional CPU frequency conversion adjustment strategy in the process of continuously changing of the system load.