Spectrum handoff in unlicensed cognitive network: Competition and avoidance

In this paper, we investigate the spectrum handoff problem in the cognitive network consisting of unlicensed spectrum. The cognitive users are supposed to be independent and to be aware of the statistics of the other user´s characteristics only. An important issue is the user´s selection between switching over to another channel and keeping communicating in the current one when collision occurs. Switching can avoid current interference, but remaining active may get a better performance as the possibility that the other user who may stop transmitting should not be left out. Thereby, we propose a novel framework in the spectrum handoff process to evaluate the user´s utility with various service time. The utility of the user is its average discounted reward. The user makes a decision about when to switch out the current channel with the sensing results which are obtained from the suitable target channels. The service time is changing with various sensing results. Moreover, we provide and analyze the two types of equilibrium in our model. The simulation results show the strategies and utility in equilibrium.