Improved surrogate data tests for sea clutter

The first part of the paper is a comparison between the two versions of Tough and Ward's (T/W) model in the generation of surrogate data to a compound stochastic k-distribution prescription. An overview of both versions of the model is presented together with predictor results which allow for direct comparison of the models. The second part of the paper describes the concept of surrogate data testing and introduces a new surrogate data test. The aim is to provide a new statistical hypothesis test which employs the method of surrogate data specifically for sea clutter. The test provides one with a significance measure of how appropriate the k-distribution model of Tough and Ward is for a particular sea clutter data set that is under test. This test incorporates the most recent T/W version 2 model for the generation of its surrogate data. In addition, a new surrogate statistic is introduced which is used to reject/accept the null hypothesis. This statistic is the normalised mean square error (NMSE) from a predictor and is a statistic which can be applied to any type of time-series. An overview of the method is presented together with results for two sea clutter data sets.

[1]  Bernie Mulgrew,et al.  False detection of chaotic behaviour in the stochastic compound k-distribution model of radar sea clutter , 2000, Proceedings of the Tenth IEEE Workshop on Statistical Signal and Array Processing (Cat. No.00TH8496).

[2]  Bede Liu,et al.  Generation of a random sequence having a jointly specified marginal distribution and autocovariance , 1982 .

[3]  Robert N. McDonough,et al.  Detection of signals in noise , 1971 .

[4]  C. Baker,et al.  Maritime surveillance radar Part 1 : Radar scattering from the ocean surface , 1990 .

[5]  Takens,et al.  Estimation of the dimension of a noisy attractor. , 1994, Physical review. E, Statistical physics, plasmas, fluids, and related interdisciplinary topics.

[6]  Cornelis J. Stam,et al.  Reliable detection of nonlinearity in experimental time series with strong periodic components , 1998 .

[7]  Nan He,et al.  Chaotic modelling of sea clutter , 1992 .

[8]  K. Ward.  Compound representation of high resolution sea clutter , 1980 .

[9]  V. J. Mathews.  Adaptive polynomial filters , 1991, IEEE Signal Processing Magazine.

[10]  C. Baker,et al.  Maritime surveillance radar. II. Detection performance prediction in sea clutter , 1990 .

[11]  Alfonso M Albano,et al.  Phase-randomized surrogates can produce spurious identifications of non-random structure , 1994 .

[12]  B. Anderson,et al.  Optimal Filtering , 1979, IEEE Transactions on Systems, Man, and Cybernetics.

[13]  Simon Haykin,et al.  Detection of signals in chaos , 1995, Proc. IEEE.

[14]  S. Haykin,et al.  Is there a radar clutter attractor , 1990 .

[15]  H. J. Arnold.  Introduction to the Practice of Statistics , 1990 .

[16]  K. Ward,et al.  The correlation properties of gamma and other non-Gaussian processes generated by memoryless nonlinear transformation , 1999 .

[17]  James Theiler,et al.  Testing for nonlinearity in time series: the method of surrogate data , 1992 .

[18]  Mark Cowper,et al.  Nonlinear processing of non-Gaussian stochastic and chaotic deterministic time series , 2000 .

[19]  S. Haykin,et al.  Chaotic dynamics of sea clutter. , 1999, Chaos.

[20]  William H. Press,et al.  Numerical recipes in C , 2002 .