Modeling Of Compressed Perception Observation Sequences Based On GEP Algorithm

The listening, speaking, reading and writing are mainly ways to exchange the information among peoples, the first three are communicated with language, so language has become the most important communication way in people’s daily lives. Language owns the maximum information capacity and the highest level of intelligence. With the development of science and technology, the production, transmission and receiving process of speech are constantly researched, and a speech signal processing technology is gradually formed and has been applied in all aspects of the society.Compressed sensing theory, greferred CS, which is a collection of information processing theory. CS is different information processing from classical Nyquist theory, compressed sensing is a new signal processing theory, which can achieve compression and sampling simultaneously. Once compressed sensing theory was proposed, aroused extensive attention of domestic and foreign experts and scholars. In this paper, based on compressed sensing theory, focus on the research of speech signals measurements sequence. Establishing a speech signals measurements sequence analysis model is the most basic research. This paper modeling with measurements sequence.The simulation experiment discovery, the model makes up the lack of speech signals measurements sequence research, which has the important means with theoretical and practical.The core issue of this study, summarized as follows:(1) According to the characteristics of the speech signals and the compressed sensi-ng constraints, the framework of compressed sensing used in measurements sequence modeling was proposed. After analysis of the relationship between the speech signals measurements sequence and the time series, getting the conclusions of measurements sequence still belongs to the scope of the time series, and then introduced the algorithm of GEP, this paper included its concepts, principles, advantages and disadvantages, providing theoretical support for next research.(2) The paper based on the compressed sensing theory and the speech signal processing theory, introducing classical linear predictive analysis to better research speech signals measurements sequence and modeling. The simulation experiment discovery that, speech signals measurements sequence are different with the original speech signal. The data of speech signals measurements sequence is manifested more of a non-linear relationship. Obtained the conclude that using nonlinear modeling methods for analysis of measurements sequence.(3) The gene expression programming (GEP) algorithm is used on nonlinear time series modeling study, sending end is not need to send a long measurements sequence to the receiving end, just send a small part measurements sequence and several model parameters to the receiving end, then the receiving end receive them, use reconstruction algorithm to restore speech signals. The simulation experiment discovery, after GEP algorithm was used to build nonlinear modeling and guarantee the performance of reconstructed speech signal, which can achieve the purpose of reducing the length of speech signals measurements sequence and secondary compression speech signals.