Federated Learning-based Secure Mobile Edge Computing for Intelligent Android Applications

Abstract

The increasing adoption of intelligent Android applications in areas such as healthcare, smart transportation, mobile commerce, and personalized services has created significant demand for secure, efficient, and low-latency computing solutions. Traditional cloud-based architectures often face challenges related to communication delays, bandwidth consumption, and user privacy exposure. To address these limitations, this paper proposes a Federated Learning-Based Secure Mobile Edge Computing (FLS-MEC) framework for intelligent Android applications. The proposed framework integrates federated learning with mobile edge computing to enable collaborative model training without transferring sensitive user data to centralized servers. Security mechanisms including secure aggregation, differential privacy, and trust-based validation are incorporated to protect model updates and enhance system reliability. The framework supports efficient edge intelligence by reducing computational latency and communication overhead while maintaining high prediction accuracy. Experimental evaluation demonstrates that the proposed approach significantly improves privacy preservation, resource utilization, and overall system performance compared to conventional cloud-centric solutions. The results indicate that the FLS-MEC framework provides a scalable, secure, and intelligent environment for deploying next-generation Android applications in dynamic mobile computing ecosystems.