Reducing Energy Consumption in Android Devices with User Profile Analysis and AI-based Feedback

• Elian Souza UFAM
• Edwin Monteiro UFAM
• Raimundo Barreto UFAM
• Rosiane de Freitas UFAM

DOI: https://doi.org/10.5753/eniac.2024.245279

Resumo

In recent years, smartphones have greatly expanded in functionality and capability, incorporating advanced features and AI processing. However, battery evolution has lagged, creating challenges for usability and efficiency. Improving battery durability and health remains a critical concern for users. This study tackles mobile device energy consumption by using AI solutions to reduce it. The Tucandeira Data Collector (TDC) app was developed to collect daily data on factors like screen brightness, CPU usage, screen-on time, RAM usage, and unused features like Bluetooth. This data forms a database for analyzing consumption patterns and building a personalized user profile. Machine learning models, including decision trees, random forests, and neural networks, are trained to identify patterns that impact battery life. The Curica Smart Alert (CSA) app uses this profile to provide real-time data collection and personalized feedback, predicting potential battery gains in minutes or hours based on user actions. Accepting these suggestions can extend battery life. The study’s findings are promising, with the Random Forest (RF) model achieving high accuracy and the Deep Neural Network (DNN) model performing well. Integrated into CSA, these models offer effective recommendations for optimizing energy use without affecting smartphone performance or inconveniencing users.

Palavras-chave: Energy consumption, user profile, Android, energy efficiency

Referências

Barreto Neto, A. C. S., Farias, F., Mialaret, M. A. T., Cartaxo, B., Lima, P. A., e Maciel, P. R. M. (2020). Building energy consumption models based on smartphone user’s usage patterns. CoRR, abs/2012.10246. [link].

Duan, L.-T., Lawo, M., Rügge, I., e Yu, X. (2017). Power management of smartphones based on device usage patterns. In Dynamics in Logistics: Proceedings of the 5th International Conference LDIC, 2016 Bremen, Germany, pages 197–207. Springer.

Mehrotra, D., Srivastava, R., Nagpal, R., e Nagpal, D. (2021). Multiclass classification of mobile applications as per energy consumption. Journal of King Saud University - Computer and Information Sciences, 33(6):719–727.

Monteiro, E., Souza, E., José, R., Balico, L., Barreto, R., e de Freitas, R. (2024). A dataset from the daily use of features in android devices. Mendeley Data.

Pereira, R., Matalonga, H., Couto, M., Castor, F., Cabral, B., Carvalho, P., Sousa, S., e Fernandes, J. (2021). Greenhub: a large-scale collaborative dataset to battery consumption analysis of android devices. Empirical Software Engineering, 26.

SWPERFI (2024). Tucandeira data collector app. [link]. Accessed: 2024-06-30.