AI-Driven Energy-Efficient mHealth Applications for Chronic Disease Management: A Review of Optimization Techniques (Preprint)

Abstract

BACKGROUND Mobile health (mHealth) applications offer real-time monitoring and feedback, which can help in chronic disease management. However, their continuous operation drains device battery excessively, making long-term use challenging. To mitigate this, AI-driven techniques like adaptive sampling, task scheduling, and federated learning offer promising solutions for optimizing energy consumption while maintaining functionality. OBJECTIVE This study provides a systematic review focusing on AI-powered approaches for energy-efficient mHealth applications. It investigates methods such as adaptive sampling, task scheduling, and predictive behavior modeling for improved energy efficiency, enabling long-term battery-powered services without affecting monitoring. METHODS A systematic literature review of AI-based optimization techniques in mHealth was conducted, focusing on energy-saving characteristics of adaptive sampling and task scheduling. The analysis included 30 studies from 2016 to 2024, quantified by percentage improvements in energy savings and battery life. RESULTS Task scheduling achieved up to 40% energy savings, extending battery life by several hours. Adaptive sampling reduced energy consumption by 25-30%. Federated learning showed energy savings of up to 25% by minimizing data transmission. Predictive behavior modeling further optimized energy by adjusting resource use based on user interactions. CONCLUSIONS AI-driven techniques significantly reduce energy consumption in mHealth applications, particularly in chronic disease management. Task scheduling and adaptive sampling show the most promise for long-term monitoring without frequent recharging. Future research should explore advanced machine learning models and energy-harvesting technologies for more sustainable applications. CLINICALTRIAL None