A Dynamic Matching Scheme between Licensed User and Cognitive User Based on Deep Neural Network with Q-learning
Mengjie Tao, Fei Lin, Yichen Che · 2019
In the cognitive radio network where licensed user (LU) and cognitive user (CU) coexist, both LU and CU should benefit from their cooperation. So, we study the dynamic matching scheme of transmission modes between LU and CU in this paper. Firstly, we determine three transmission circumstances of CU according to LU's information transmission power. Secondly, When LU has more data to send, CU help LU as a relay to forward the information. We obtain power expression to get the goal of CU exploration in deep Q-learning network (DQN). Finally, we analyze the performance of the DQN algorithm and LU's channel capacity under dynamic matching scheme, aiming to improve the accuracy of information transmission. From the analyzation and simulation results, it can be seen that LU can get better channel capacity and CU can get more opportunities to access spectrum.