IMPROVING THE RESILIENCY OF COMMUNICATION NETWORKS BY APPLYING ARTIFICIAL INTELLIGENCE

One of the major concerns in wireless spectrum management is the underutilization of the active and licensed wireless spectrum. To enable the spectrum access to more users, it is of utmost importance to utilize these unused parts of the spectrum. Spectrum sensing and cognitive radio-based communication systems have emerged as a means to employ these unoccupied segments. These systems work by monitoring the spectrum for any licensed user activity and then transmitting when any licensed user activity is not present or transmitting while causing minimum interference when a licensed user is present. In a situation where an entity is attempting to transmit various traffic streams to multiple end-users by leveraging multiple cognitive radios, it would be appropriate to consider the entire set of cognitive radios as a single logical unit. The reason is that each cognitive radio transmission affects the transmission quality of other radios.

Various cognitive communication systems that are capable of transmitting QoS and QoE-aware traffic to end-users have been proposed in the literature. However, there are significant technical challenges in effectively leveraging cooperation among cognitive radios to ensure reliable task-oriented communication, particularly in constrained communication environments. This comprehensive research tackles these technical challenges by exploring emerging AI and blockchain technologies. The research mainly consists of four main components: 1) development of blockchain-powered QoE-aware cooperative cognitive radio infrastructure, 2) exploration of trustworthy AI-powered spectrum sensing, 3) designing a QoE-aware communication infrastructure for surveilling wildfire, and 4) development of a semantic communication system for remote robot operations.