Multi-channel Based Dynamic Spectrum Allocation Strategy With Imperfect Channel Sensing And Performance Study

With the high-speed development of radio communication, the problem of spectrum resources shortage is prominent increasingly. Cognitive radio network of licenced spectrum "secondary utilization" can effectively improve the spectrum utilization. Based on the centralized cognitive radio network with multi channel, considering the imperfect sensing results, we give the study of dynamic spectrum allocation strategy、performance analysis and system optimization.Firstly, according to different business requirements of licenced users and cognitive users, we propose a CB/PT(Channel Bonding or Parallel Transmission) dynamic spectrum allocation strategy, in which the transmissions of licenced users are using channel bonding while the transmissions of cognitive users are paralleled. Considering the possible sensing errors, we establish a modified preemptive priority queuing model with multi-server synchronous interruption. Using the matrix geometric solution method for the steady-state analysis, we derive the performance measures in terms of interference ratio of licenced users packets, channel utilization and average delay of cognitive users packets.Secondly, on the basis of CB/PT strategy, to further improve the transmission quality of cognitive users, setting reserved channels for cognitive users packets, we put forward a CB-CR(Channel Bonding and Channel Reservation) dynamic spectrum allocation strategy which combine the channel bonding mechanism with the channel reservation mechanism together. Combining with the imperfect sensing results, we establish a three dimensional Markov model. Using the matrix geometric solution method for the steady-state analysis, we derive the performance measures in terms of interference rate of licences users packets, average delay of cognitive users packets and throughput of cognitive users packets.Then, in view of the established system models, we give the mathematical analytical algorithm for the rate matrix and the distribution of the steady-state; in view of the CB/PT strategy and the CB-CR strategy, we give the simulation statistics algorithm for the performance measures. Using the results of numerical analysis and the results of monte carlo simulation, we reveal the influence of spectrum capacity and reserved channel number on system performance targets, quantitatively evaluate system performance under the CB/PT strategy and the CB-CR strategy based on imperfect channel sensing.Finally, in view of the CB/PT strategy, based on the trade-off relationship between different performance targets, we give the optimization scheme of the capacity spectrum through setting up the system cost function. In view of the CB-CR strategy, we analyze the personal optimal behavior of cognitive users and the social optimal behavior of the strategy, then we achieve the social optimization of the strategy through giving a pricing schemes for cognitive users.