A Nash learning based trust aware opportunistic routing in cognitive IoT

The explosive growth of communication technologies has led to massive opportunistic connections between smart devices and the Internet of Things (IoT), resulting in severe spectrum scarcity accompanied by false detection. As a key enabler for future IoT, cognitive radio (CR) gives rise to cognitive IoT (CIoT), which significantly improves quality of experience. However, the dynamic and heterogeneous spectrum environment makes it extremely challenging to design robust data transmission strategies for unreliable IoT applications. Opportunistic routing, which takes advantage of the broadcast nature of wireless communications, provides an effective way to improve opportunistic transmission performance. To address these issues, we propose NTLODT, a Nash Learning and Transfer Learning based trust aware Opportunistic Data Transmission scheme, which jointly optimizes efficiency and reliability. NTLODT provides a new learning-based opportunistic routing for differentiated services in CIoT. In particular, we use Nash learning to refine the candidate relay set and transfer learning under a trust-aware mechanism to support intelligent transmission decisions. We also prove the convergence and analyze the complexity of the proposed algorithm. Simulation results show that NTLODT outperforms existing methods in average energy cost per bit, average delay, expected routing cost, and throughput.