Abstract: For the non-stationary digitally modulated signal a novel feature; High Order Cross Cumulant (HOCC) was proposed. The non-linear dynamic modeling of an adaptive fuzzy modulation classifier, which based on the training mechanism within the neural network, was first presented. The model adopted the hierarchical decision-based structure,which made the features match the classifier and reduced the redundancy of the membership functions and fuzzy rules.According to the distribution of the feature samples, we established the Hierarchical Fuzzy Neural Network System (HFNNS) with initial experience to guarantee the controllability of the knowledge inference structure. By applying the training data,the algorithm adaptively adjusted and optimized the structure parameter and completed the approximation process. The simulation results verified the better robustness of the system in the presence of various environment parameters (SNR etc) , as well as the improvement of the average probability of correct classification and the algorithm efficiency, compared with the neural network classifier and fuzzy classifier.

Key words: Modulation recognition, HFNNS, HULL, Fuzzy inference, Adaptation