Research On Algorithm Of Modulation Recognition On Graphs For Alpha Stable Distribution Noise

Automatic modulation classification(AMC),which identifies the modulation scheme of an unknown received signal with little or no a priori knowledge about the signal,is considered as an indispensable but sophisticated mechanism between signal detection and demodulation.AMC plays a pivotal role in many military and civilian communication applications,such as cognitive radio,adaptive modulation and coding,spectrum surveillance,and electronic warfare systems.In practical communication scenario,the additive noise,due to a variety of natural and man-made sources,usually presents impulse characteristic in the statistical sense,which could only be modeled as Alpha stable distribution.This leads to the conventional AMC algorithms designed for optimal performance in Gaussian noise are significantly deteriorated when they are directly employed in Alpha stable distribution noise environment.This thesis introduces a new AMC approach using graph-based generalized second-order cyclic spectrum(GSCS)analysis under Alpha stable distribution noise background.The new AMC approach can be generally divided into several categories.1.The generalized second-order cyclic spectrum analysis is first invoked to make the polyspectra of the contaminated signal obtainable.Then,the graph-based AMC mechanism is systematically established on the graph representations of the GSCS to identify the modulation scheme.Those graph features can be easily expressed by the corresponding adjacency matrices.2.A new feature-extraction paradigm for graph-based automatic modulation classification is proposed in this thesis.The modulation features are optimally constructed using the Kullback-Leibler divergence of the dominant entries in the adjacency matrices associated with the graph presentation of the GSCS.Then,the normalized Hamming distance(NHD)is invoked to measure the discrepancies between the features derived from the training and test data to determine the modulation type.3.Simulation results demonstrate the effectiveness and superiority of the proposed AMC algorithm.The effect of sample size,timing and frequency offset on the proposed method is evaluated via simulation under different conditions.The computational complexity of the proposed method is investigated in detail.Simulation results demonstrate that our proposed method can achieve much better classification accuracy than the existing technique when the cyclostationary theory is used.The timing error and frequency offset of the input signal have a potential deleterious impact on the performance of the proposed algorithm.But a systematic feature-selection approach is proposed in our method,which can select the most distinguishable features,leads to the promising solution to automatic modulation classification.