Research On Spectrum Sensing Algorithm Based On Random Linear Network Coding

With the rapid development of the Internet and intelligent mobile terminals, the contradiction of spectrum supply and spectrum demand becomes increasingly prominent. It becomes the bottleneck to restrict the development of wireless communications technology. However, according to the report of the FCC, the current average spectrum utilization is only between 15% and 85%, resulting in a tremendous waste of spectrum resource. Cognitive radio can improve spectrum utilization by using vacant spectrum for communication, so it has become one of the key technologies to solve the problem of the shortage of spectrum resource. Spectrum sensing is the precondition for cognitive radio. Therefore, how to sense the spectrum state rapidly is an important problem for cognitive radio research. Firstly, with the goal of reducing the detection delay and increasing the system throughput, the random linear network coding is introduced into cognitive radio, and a fast spectrum detection algorithm based on random linear network coding is studied. Secondly, in order to reduce the impact of detection delay and false alarm probability, a Hidden Markov Model is applied to model and analysis primary user channels, and a spectrum prediction algorithm based on random linear network coding is studied deeply. The main contents and innovations of this paper are summarized as follows:1. A cumulative summation spectrum detection algorithm based on random linear network coding (RLNC-CUSUM) is proposed to address the problem that spectrum states transfer frequently and the detection delay is large in the traditional cognitive radio (CR) network. In the proposed algorithm, random linear network coding (RLNC) is introduced to primary user (PU) subnetwork in the CR network and shapes traffic. Consequently, transitions between spectrum states are expected to be less frequent and the spectrum structure is of more regularity, which, in turn, results in great improvement in detection delay and throughput rate. In addition, since the anti-fading performance of the traditional cumulative summation (CUSUM) algorithm is poor, five kinds of fading channels are modeled and analyzed, and the detection performance of RLNC-SUSUM algorithm over various fading channels is compared, and then it is proved that the RLNC-SUSUM algorithm has good anti-fading performance. Simulation results show that the proposed algorithm provides great improvement in detection delay, throughput rate and anti-fading performance at a certain false alarm probability, and it can be better adapted to the complex fading channel environment.2. Based on the predictability induced by random linear network coding in cognitive radio and the spectrum predictability of Hidden Markov Model, a back-off spectrum prediction algorithm based on random linear network coding (Back-off-SP) is proposed. In order to improve the sensing performance, a Hidden Markov Model is applied to model and analysis primary user channels, a prediction technology based on back-off is introduced into the proposed algorithm, and improve the traditional spectrum prediction algorithm based on Hidden Markov model. Simulation results show that the proposed algorithm provides great improvement in spectrum sensing performance, including increasing the prediction probability and reducing the false alarm probability, and then reduces the interference of secondary users to the primary users and improves the throughput rate of secondary users.