Research And Application Of Typical Signals Pattern Recognition Algorithms

The Applications of typical signal pattern recognition, in the field of communications, mainly are non-cooperative communication, such as signal confirmation of civilian communications, interference identification, electronic warfare, and electronic surveillance of military communications. In the biomedical field, the applications are mainly to provide more clinical diagnosis and provide more effective treatment for patients. So the pattern recognition of typical signal is broadly applied.In recent years, many scholars have made a lot of exploration in the signal pattern recognition whether in the field of communications, or in the biomedical field. But many of them are just stayed in the simulation phase. The simulation algorithm is not applied to the actual signal. A feature extraction method of digital signals in typical communication signals is proposed, which is relied on laboratory projects and based on spectral correlation characteristics and spectrum analysis. And the extracted features are applied to threshold method and support vector machine methods which are used to classification. The simulation results show the effectiveness of the algorithm. Then, a modulation type identification plan for R&S ESMD is designed, which is based on the combination features of time domain and frequency domain. The system is validated by the effective and stable experiments results. Finally, the EEG features extraction method based on echo state network is explored. Then the marginal fisher analysis method is used to reduce the dimension of extraction feature vector and the support vector machine is applied to final EEG pattern recognition. The simulation results prove the validity of the methods.