Federated Edge Computing for Privacy-Preserving Analytics in Healthcare and IoT Systems
Ravi Kumar Vankayalapati, Sudha Rani P. R., Shashikala Valiki, Venkatesh Ganti · International Journal of Computer Trends and Technology · 2025
Federated Learning and Edge Computing, also known as Federated Edge Computing, have emerged as breakthrough technologies that enhance the privacy of patients and overall data privacy compliance. Federated Learning has been developed as a distributed learning approach to train machine learning models that exchange and learn only the model parameters from different databases. At the same time, the data remains local to the clients, hence avoiding the possibility of privacy erosion. The combination of Federated Learning and Edge Computing essentially extends the aggregation of learning parameters happening in Federated Learning from traditional edge nodes to increasingly underutilized resources at mobile phones, IoT, or other devices acting as smaller-scale edge nodes. A lot of work has been published to improve machine learning in a federated learning environment. Privacy, especially in the healthcare domain and in IoT, is a vital issue. More and more data is collected by IoT sensors, wearables, tablets, networks, compute resources, and even facilities. The collected data can range from energy management, visitor monitoring, equipment performance analytics, home/building security, surveillance, and even patient vitals, among others. There are many challenges in storing and processing this data. Given its sensitive nature, this data must be securely stored. They should possibly also be computed upon, creating insights or making decisions based on the data before importing it to less secure environments. In addition, compliance with privacy rules and regulations transmitted in healthcare communication acts must be met when storing and processing medical data/protected health information. Federated Learning provides a model set to instantiate and execute. In contrast, federated edge computing provides an instantiation on edge hardware to run the workloads and functionalities in a federated learning environment. With federated learning, all the processing of data, specifically medical data, will be kept within pockets of edge nodes themselves. Thus, this approach addresses safe and secure processing and the privacy of the user, patient, or community being observed.