Research On Method Of Detecting Weak Signal In Sea Clutter Based On Improved Teaching Learning Based Optimization Algorithm

In complex marine environment,the detection of sea targets is greatly affected by sea clutter.sea clutter is affected by many factors,also,the mechanism of sea clutter is complex.It is difficult to describe features.The significance and application value to radar signal processing is great when the weak targets are detected from sea clutter.In this paper,weak signal detection model in sea clutter background is studied.In order to solve the problems of low accuracy and inefficiency of traditional prediction models,Bayesian Echo State Network and Combined Support Vector Machine based on the Improved Teaching Learning Based Optimization algorithm are proposed to detect weak signals in sea clutter background,the specific research contents are as following:In order to improve the generalization ability of echo state network model,Bayesian theory and echo state network are combined to form the main body of the model.Since it is difficult to select the parameters of Bayesian Echo State Network model,an Improved Teaching Learning Based Optimization algorithm with high global search ability is used to optimize the parameters of the model.A one-step prediction model in chaotic background is established by the optimal parameters.The magnitude of prediction error is observed to judge whether there is weak signal in chaotic background.The time series generated by Lorenz system and the measured sea clutter data are used as chaotic background noise.The results show that the prediction accuracy and efficiency of the optimized Bayesian Echo State Network model are higher than those Echo State Network model optimized by genetic algorithm and the Neural Network model.The optimized Bayesian Echo State Network model can effectively detect weak signals in the chaotic background and has a lower threshold.In order to improve the processing ability of Support Vector Machine for complex non-linear problems,a weak signal detection method based on Improved Teaching Learning Based Optimization algorithm and phase space reconstruction theory is proposed.The RBF kernel function and polynomial kernel function are linearly combined to form the combined Support Vector Machine.The Improved Teaching Learning Based Optimization algorithm with strong optimization ability is used to optimize the combined Support Vector Machine.According to the obtained optimal model parameters,the model is built,trained and used to predict the weak signals of chaotic background.The time series generated by Lorenz system and the measured sea clutter data are used as chaotic noise.The results show that the combined Support Vector Machine model can effectively detect the target signal in thechaotic background.Compared with other optimized support vector machines and Bayesian Echo State Network models,the prediction accuracy and efficiency of combined Support Vector Machine optimized by the Improved Teaching Learning Based Optimization algorithm are improved.