Research On Modulation Signal Recognition Method Based On Fractal Theory

With the continuous development of modern communication technology,the shortage of radio spectrum resources and the complexity of communication methods have gradually become the main bottleneck restricting the development of communication technology.How to effectively manage communication signals has become a problem that needs to be solved urgently.The first thing to manage signals is to be able to quickly and accurately identify the modulation methods of different signals.Since communication signals are susceptible to noise interference during the propagation process,the extracted time-frequency domain features are more sensitive to noise,and it is difficult to improve the classification and recognition ability when the signal-to-noise ratio is unknown.Therefore,this article introduces the fractal theory commonly used in nonlinear systems into modulation signal recognition.As a function of time,the fractal features of the communication signal extracted from the change of its waveform distribution have good antinoise performance,and the classifier design is simple,the recognition rate is high,and it has good engineering application value.In this paper,the commonly used fractal dimension is simulated and analyzed,and it is verified that the box dimension,Katz dimension and Sevcik dimension have good classification characteristics and robustness to noise changes.At the same time,this paper also analyzes and studies the commonly used classification machine learning algorithms for pattern recognition.Through simulation and comparative analysis,it is known that the decision tree algorithm and the neural network algorithm have better classification results.In order to solve the problem of aliasing in the signal dimension when the signal-to-noise ratio is low,a new joint feature vector is constructed in this paper.With the help of the signal's statistical variance,it also has a good degree of discrimination in the low signal-to-noise ratio to make the classification The accuracy rate is improved.When the signal-to-noise ratio is 5dB,the overall recognition rate can reach99%,which is nearly 19% higher than the recognition rate that only uses the fractal dimension as the feature parameter under the same conditions.