Research On Intelligent Cooperative Spectrum Sensing Algorithm Based On Probability Vector

Spectrum sensing is one of the key technologies for efficient use of spectrum resources and has important research value.The subject is derived from a military cognitive communication anti-interference pre-research project,and the cluster cognitive communication algorithm is deeply studied.In the research process,it is found that the more secondary users(the larger the scale of the cooperative cluster in the project),the accuracy of cooperative spectrum sensing.The higher the value,the higher the perceived data dimension and the increased data processing overhead.In addition,the noise uncertainty caused by thermal noise and quantization noise generated by the digital signal processing process will lead to a decrease in the perceived accuracy of the cooperative spectrum.Aiming at the problems that the number of secondary users increases,the dimension of perceived data increases and the data processing overhead increases,this research designs a mapping algorithm that converts high-dimensional energy vectors into constant two-dimensional vectors.The data in the two-dimensional vector can represent the probability that the original data belongs to a certain channel state,so it is called a "probability vector".This paper compares the performance of probability vectors and energy vectors in different algorithms.After the high-dimensional energy vectors are converted into two-dimensional probability vectors by this mapping algorithm,K-Mediods and fuzzy support vector machines are used to train and classify the data..Focus on the comparison of the time consumption of the training phase and the classification phase when the probability and energy vectors are applied to the two algorithms.Simulation results show that for the same algorithm,when the number of secondary users is 16,using the probability vector takes less time than using the energy vector: the training duration is reduced by more than 31%,and the classification delay is reduced by more than 12%.Aiming at the problem that the contribution of each secondary user to the data fusion in the uncertain noise scenario leads to the deviation of the sensing accuracy,a weighted cooperative spectrum sensing algorithm based on particle swarm optimization neural network(PSO-BP)is proposed.The algorithm uses PSO-BP to optimize the weight coefficients of each secondary user.The weighted cooperative spectrum sensing algorithm and the signal-to-noise ratio weighted cooperative spectrum sensing algorithm are compared and studied through simulation experiments.The results show that the weighted cooperative spectrum sensing algorithm Cooperative spectrum sensing algorithms have better detection performance.At the same time,for this weighted cooperative spectrum sensing algorithm,the number of iterations using the probability vector is less than that using the energy vector.Taking16 secondary users as an example,the number of iterations is reduced by more than 40%.