A COMPREHENSIVE ANALYSIS ON TRUST-AWARE FEDERATED LEARNING IN IOT EDGE COMPUTING

Abstract

Federated Learning has been among the prevalent methods that enable cooperation in machine learning as well as maintaining privacy in decentralized networks like Internet of Things (IoT), healthcare, edge computing and smart systems. Nevertheless, federated learning has experienced some challenges including privacy issues, increased communication costs, managing node trust issues, client malicious updates, data poisoning and scalability issues. This survey will review 26 research papers about federated learning which have been published between 2024 and 2026. The fields related to the above-mentioned research papers include various fields such as differential privacy, trust aware learning, block chain based federated learning, federated learning, Explainable AI, edge intelligence, intrusion detection and health care analytics. Various techniques utilized in the above-mentioned research papers include Differential Privacy (DP), Generative AI, Convolutional Neural Network (CNN), LSTM, GNN, Blockchain, MARL and Trust ManagementHowever, based on the findings from the report, the efficiency is quite clear and accurate, while at the same time the increase in level of trust , the reduction of communication overhead and the ability to resist any attacks. The most important thing that one should keep in mind based on the findings from the survey is that despite all the successes, the development of the next generation FL framework that consists of four key factors is very necessary.