Analysis of Energy Consumption on Android Devices for Developers: A Systematic Mapping Study

Reducing energy consumption is a major challenge that mobile computing must deal with. Smartphones are constantly evolving to match traditional computers in many aspects, especially in processing and memory. However, user experience is severely impacted by the rapid discharge of batteries in smartphones. This article aims to identify which metrics and assessment techniques are applied to evaluate energy consumption on Android smartphones, by summarizing factors that cause high energy consumption through a Systematic Mapping Study (SMS). The methodology of this SMS consisted of performing searches on the digital libraries ACM, IEEE, and Scopus. Sixty articles were obtained, of which 17 were identified as relevant for this study. Among the main methods identified, energy consumption is collected at time intervals based on information on battery voltage/current or on information from features such as Wi-Fi, cellular networks, screen brightness, screen duration on, Bluetooth usage, and others. Regarding the tools and applications that collect such information, there are Android batterystats, applications developed specially for each research like BatteryHub. The data is mainly analyzed by techniques such as clustering (17.65%), covariance (17.65%), Bayesian classification, and decision trees (11.76%). From these techniques, it was identified that user profile is the main factor affecting battery performance, being present in 23.53% of the articles, followed by mobile networks and Wi-Fi (17.65%), in addition to applications and services in background present at 11.76%. Finally, two articles in this SMS provide recommendations to reduce consumption based on users’ usage profiles. In summary, the analysis has shown a significant correlation between user habits and energy consumption. As a result, it is recommended that developers prioritize the exploration of artificial intelligence techniques to automatically adjust smartphone usage settings based on usage context. This approach can lead to significant battery power savings.