SDN-based Federated Learning approach for Satellite-IoT Framework to Enhance Data Security and Privacy in Space Communication

Abstract

The proliferation of IoT devices and integration of machine learning technologies paved the path towards automation in various sectors guided by Artificial intelligence (AI). It enables multitudes of use cases ranging from mass scale cloud-edge computing based robust communication between smart IoT devices, weather variation detecting low powered remote sensor nodes residing on a harsh terrain, AI-assisted driverless vehicles immaculately cruising through traffic to industrial robots performing sophisticated tasks with precision and finesse. As space colonization is a becoming a myth of the past and venturing towards reality, this AI-based IoT ubiquity will also be a major mart of those space colonies where autonomous infrastructures with be the norm. These IoT integrated networks will also boast a wide area of coverage reaching the furthest of the horizons with low orbit satellite integration. However, the mass deployment of these modern technologies is heavily contingent to the fact that data is safeguarded from malicious intrusions. Therefore, in this paper we have proposed an approach to thwart data breach that can plague satellite-IoT framework with respect to space communication. The framework is based on software defined networking that uses federated learning techniques for distributed systems and employs deferential privacy while sharing data among devices to ensure secured critical data transmission between IoT devices.