Research On Dynamic Spectrum Access And Tcp Optimization Strategies In Cognitive Radio Networks

As an intelligent spectrum sharing technology, cognitive radio is concerned widespread by researchers. In order to utilize the inactive frequency bands, cognitive users must perform the spectrum sensing for the CR-used frequency bands frequently or periodically. Wireless devices with cognitive ability can intelligently sense its surrounding spectral environment and adaptively adjust its operating parameters in the physical or data-link layers. This approach improves the spectrum utilization effectively and resolves the scare spectrum problem. Dynamic spectrum access and cross-layer optimization technology, is the two key technologies in cognitive radio networks.The former primarily studys the problem which channel should be accessed by CR user, and the latter focuses on how to improve the TCP performance.This article firstly introduces the spectrum sensing technology, dynamic spectrum access and cross-layer optimization technology in cognitive radio networks, which provides a theoretical foundation for the following research. Then we propose the HMM based spectrum access strategy for cognitive radio networks, which is used to describe the unpredictable nature of wireless channel and establish the channel model. We implement HMM predict next channel state based on past ones. Set the true state of the channel, including idle and occupied, and use SNR as observation status, then input it to HMM model as observation value sequence. We train the HMM model repeated until it converges, when we get the channel state transition probability matrix. After that, we formulate the dynamic spectrum access problem as a markov decision model to obtain the optimal solution that makes better spectral efficiency. The simulation results prove the schema we propose in this paper is much better than the existing random access strategy.Combination of the above and the existing research, we presents a cross-layer optimization algorithm based on the Markov decision process in order to improve the performance of end-to-end transmission in cognitive radio network, which through jointly optimizing of channel access decision, the physical layer modulation encoding and link layer frame size and other parameters to improve the throughput of TCP. The simulation results show that the proposed algorithm is able to improve the performance of TCP significantly, compared with the existing non-cross-layer algorithm.