Spectrum prediction method in cognitive radio system based on convolutional neural network

Abstract

The invention discloses a spectrum prediction method in a cognitive radio system based on a convolutional neural network. The method can obtain the higher prediction accuracy through the shorter-timetraining, and belongs to the field of information and communication. The method comprises the steps: 1, performing the uninterrupted spectrum sensing of a frequency band F through a period T, recording S<t0>, S<t1>, S<t2>,... of the spectrum sensing from the moment t0 and I<t0>, I<t1>, I<t2>,... corresponding to S<t0>, S<t1>, S<t2>,..., marking t0+kT as tk, wherein k=0, 1, ..., wherein S<tk> represents the occupation conditions of all channels of the frequency band F at the moment tk, I<tk> represents the number of idle sensing periods of each channel of S<tk> from the moment tk+T; 2, inputting S<t0>, S<t1>, S<t2>,... and I<t0>, I<t1>, I<t2>,... into the convolutional neural network, and performing the training of the convolutional neural network, wherein I<t0>, I<t1>, I<t2>,... are used for making labels in the training; 3, continuously predicting the channel C with the maximum future idle probability according to S<t0>, S<t1>, S<t2>,... through the trained convolutional neural network.

Description

technical field [0001] The present invention relates to the field of information and communication, in particular to an intelligent idle channel allocation algorithm after spectrum sensing in a cognitive radio system. Background technique [0002] In recent years, with the rapid development of wireless communication technology, user business types and business requirements are also constantly explosively increasing, which leads to an increasing user demand for radio spectrum resources. However, the average occupancy rate of authorized spectrum in the time domain, space domain, frequency domain, and energy domain is very low, and the utilization rate of spectrum resources is generally not high. The contradiction between the shortage of spectrum resources and the waste of spectrum resources is becoming more and more prominent. Cognitive Radio technology (Cognitive Radio, CR) can intelligently search for idle spectrum access in the current frequency band, thereby effectively i...

AI Technical Summary

Problems solved by technology

[0005] However, this mechanism is accompanied by a problem: if the number of PUs in the current frequency band frequently accesses and exits channels or there is high-power dynamic malicious interference, the SU may need to avoid PUs frequently, re-spectrum sensing, and re-access. retransmit packets

[0008] However, there are many problems in the existing spectrum prediction methods. For example, the spectrum prediction based on Markov chain and the spectrum prediction based on regression analysis have low prediction accuracy. In some cases, the collision rate is even higher than random selection of idle channel access. Method collision rate

Although the spectrum prediction method based on BP neural network has higher prediction accuracy than the previous two, it needs to train and use a fully connected neural network, and the error backpropagation in the training stage takes a long time and the data forward propagation in the prediction stage takes a long time. , and the training is easy to fall into local optimum, and more training samples are needed to obtain better generalization

Moreover, once the parameters of the above method are determined, it is not easy to change. When the spectrum environment changes or there is dynamic malicious interference in the frequency band, the prediction accuracy will drop, and the method may fail in severe cases.

Images