2012 · 23 citations · 3 references
ComputingComputer ScienceElectrical EngineeringDigital CommunicationFeature EngineeringMachine LearningDigital Signal ProcessingModulationPattern RecognitionRepresentation LearningSignal ProcessingSupport Vector MachinesAutomatic RecognitionCommunication SignalsDecision Threshold SelectionDigital ModulationModulation ClassificationModulation RecognitionPattern Recognition In SignalsSignal Classification
In order to solve the problem of low modulation recognition rate of digital communication signals and the difficulty of selecting the appropriate decision threshold, the paper constructs characteristic parameters (CP) for recognizing signals in the cumulant domain, and uses support vector machines based on binary tree as a classifier to identify the characteristics vector mapping to high-dimensional space, which achieves automatic recognition of digital modulation. The algorithm has a high recognition rate not only, but also it is simply and efficiently. And it solves the problem of sample inseparable in low-dimensional space. It's good at generalization ability.When signal to noise ratio (SNR) is higher than −1dB, the recognition rate achieves 94%. Compared with existing algorithms, simulation results show the superioity of the algotithm.
3