| With the continuous development of the communication equipment, network and system,modern military science is in constant perfection and iterative update, and the communication interference is faced with increasingly complex environment and severe challenge. With a variety of advanced technology appearing and rapidly developed, the technology of communications interference must be improved followed, in order to ensure the efficiency of the communication interference. At the same time, it is being more and more important and urgent for evaluating the jamming system to research and develop the newest technology and method timely.Therefore, this article studies the interference measure evaluation system based on speech signal. Processing speech signal interfered by communication interference equipment, we can get an objective evaluation on the degree of interference, and establish an speech communication interference evaluation system that can satisfy the demands of the project.The main works of this paper are as follows:First, the character representation of speech signal is researched and analyzed. The principle and calculation method of the Mel frequency cepstrum coefficient (MFCC), wavelet transform and cognitive features are mainly introduced.Second, this paper introduces the commonly used data fitting algorithms, including the least squares method and BP neural network model. We also introduce the multi-measure fusion based on RF random forest. Explain the basic principles of the three algorithms and the application in the interference evaluation systemFinally,based on subjective and objective data fitting,the interference evaluation results of speech signals under different characteristics are evaluated. Application of the system is summarized in the experiment. Through the experiments, the multi-measure fusion based on RF random forest has a better correlation coefficient and the projects achieve the research requirements and expected results. |
PDF Full Text Request