Cognitive Networks with Dynamic User Classification for Tactical Communications

A novel cognitive radio paradigm for tactical networks is presented. The proposed framework employs a dynamic classification of users into primary and secondary users as a function of time-varying packet priority to enable reliable and timely delivery of critical information generated from users. The stochastic model of the network captures the coupling between the random processes tracking the packet service state of the individual users due to mutual interference and networking protocols. The optimal channel access strategy maximizes the quality of service of packets conveying non-critical information under constraints on the delivery probability of critical information. Numerical results show that the optimal strategy maps the channel access probability onto the packet service and priority state of packets being transmitted by the users. The proposed framework, thus, defines primary and secondary packets, rather than primary and secondary users.