Covert Communications in Cognitive Mobile Edge Computing Networks using Restless Multi-Armed Bandits

Covert communications in mobile edge computing networks are particularly challenging due to the mobility and multiple access opportunities. In this paper, we present a cognitive network where secondary nodes (SNs) Alice and Bob wish to communicate and offload computing tasks without detection by primary nodes (PNs) Willies. A distributed probing scheme is presented to find secure available connections that depend on the activity of PNs and belief in the covert outage probability due to uncertainty in the mobility of SNs. Since the available channels and computing resources are limited, a covert priority-based Restless Multi-Armed Bandit (RMAB) algorithm is developed to activate the connections that maximize the long-term reward in terms of secure data transmitted within a latency bound. Our numerical results show significant improvement in our schemes compared to existing solutions.