Cancellation Of Sea Clutter Based On Chaos Theory

Sea clutter is the backscatter from fluctuant sea surface. Sea clutter is the major noise interference of seaborne radar. Complex in signal structure, Abundance in Spike, it is the major factor degrading detection performance of radar. It is traditionally processed by statistical theory through probability distribution such as log-normal, weibull, compound K distribution. Discription in only macro and statistic rather than in exact and physical, statistical theory has limited effect.A new way turns out with establishment and development of chaos theory. Haykin point out that sea clutter is derived from certain unkown chaotic system rather than from random noise after analysis of correlation dimension and Lyapunov exponent. He presents a method of sea clutter modelization and cancellation utilizing artificial neuro network and get good result. But there are still some key problems:Is sea clutter chaos or random? The anser is the basis of Haykin's method. While Haykin and some researcher provide many proofs, there are still some questioners.With chaos in mind, how to develop an effective method of sea clutter cancellation and detection? So much choice, it is hard to choose and develop an effective method.How to adapt to time-variance of sea clutter? And how much is the interference?This paper deals with these key problems. Real sea clutter data recorded by radar is used in analysis. Statistical analysis, phase space reconstruction is processed. And then the correlation dimension, Lyapunov exponent is calculated. The result is compared to known chaos system and random noise, and lead to the conclusion that sea clutter is similar to chaos system rather than random noise.Predictable is the key charater of chaos and can be used to distinguish chaos and random noise. This paper present a improved one order predict method which performs well in prediction of sea clutter. The precise prediction conforms that sea clutter is chaotic rather than random.Based on these result, a predict model is presented, whose parameters are got by utilizing artificial neural network. The model is then used in sea clutter cancellation and target detection. Real sea clutter data collect by radar is processed and shows good result.All the data used in analysis and process is collected in different time. The result shows that sea clutter is slow time variant and can be corrected by adaptive processing.