Dynamic Power Management On Smartphone’s Screen

Smartphones win the favor of users for its operating system, free installation of various software and secure touch-screen operating characteristics, and there is a huge amount of users to use them. Smartphone is battery powered. However, the development of battery technology slows down and has entered the bottleneck stage, because of the limitation of volume, limited battery capacity can’t meet the large energy consumption on surfing the internet, navigation, full-screen touch and so on. To solve the energy consumption from the software and hardware, power management policy is of great significance.Android accounts for 86% of smartphones’ operating system, which is based on the Linux kernel. The kernel has its own power management framework. Combined with its screen management mode, the energy optimization of screen can be obtained. First, use the power monitor to monitor the energy consumption of the screen, analyze its energy consumption characteristics and establish the energy consumption model. Secondly, in order to minimize energy consumption, deterministic algorithm based on the model is designed to obtain the timeout value. To take into account the users’ different interaction habits, probabilistic algorithm which adds the distribution of the interaction idle time is proposed. Finally, parameters of the model are measured by using the benchmarking methodology. Android background service program has been written to collect users’ interaction data. The specific distribution can be obtained by analyzing its distribution characteristics and doing hypothesis test. Use two algorithms to calculate optimal screen timeout values and develop evaluation criteria for energy consumption, then to analyze energy consumption of the default timeout value and obtained optimal timeout values.The collected data presents the idle interaction time of the smartphone lines with right truncated exponential distribution. The optimal timeout value isn’t the most energy-efficient compared with different timeout values’ energy consumption. Deterministic algorithm has fixed timeout values, but it doesn’t have to save energy which determined by users’ habits. Probabilistic algorithm considers the user’s habits, can saves from 6% to 16% energy. Therefore, the users can obtain the best timeout values through the probabilistic algorithm. By using the optimal timeout values, power management module can switch the screen effectively to save energy.