An Artificial Neural Network Spectrum Sensing Method Based on Wolf Pack Optimization

Abstract

The invention relates to an artificial neural network spectrum sensing method based on wolf pack optimization. The decision-making center turns on the main user signal simulation transmitters in each sub-area in turn, and the decision-making center receives the local detection results from the cognitive users, and converts the received cognitive The local detection information sent by the user and the statistical and calculated detection probability information of the cognitive user generate a training sample set, and a neural network test function is generated according to the neural network structure and training samples; based on the neural network test function and the neuron weights generated after training Value matrix, use the wolf pack optimization method to optimize the weight matrix, and input the optimized weight matrix into the neural network working module; the cognitive user detects the main user signal, and the decision center will receive the cognitive user's signal The local detection results are fused with the accumulated cognitive user detection probability, and the fusion value is input into the optimized neural network to determine whether the main user signal appears.

Description

technical field [0001] The invention relates to an artificial neural network spectrum sensing method based on wolf group optimization. Background technique [0002] "Cognitive radio" was first proposed by Joseph Mitola in 1999. It uses software radio as an extended platform and is a new intelligent wireless communication technology. It can perceive the surrounding wireless environment, and adjust transmission parameters in real time through understanding the environment and active learning to adapt to changes in the external wireless environment. Spectrum sensing technology means that cognitive users obtain spectrum usage information in wireless networks through various signal detection and processing methods. At present, most researches on spectrum sensing technology focus on three aspects: local sensing, cooperative sensing and sensing mechanism optimization. [0003] There are currently three main methods of local detection. The first is matched filter detection, which ...

AI Technical Summary

Problems solved by technology

[0007] The prior art most relevant to the present invention is that in January 2012, Liu Qing of Beijing University of Posts and Telecommunications proposed a self-organizing neural network-based The spectrum sensing method of the network, the false alarm probability detected by this method is 0, and in the low signal-to-noise ratio environment, it still has a certain probability of correct perception, but the correct probability is not high

Images