Research On Radar Clutter Simulation Technology

Radar signal simulation is extremely important to the development,testing and debugging of radar,and the simulation of radar clutter is an indispensable part of it.Studying the distribution characteristics of radar clutter and mastering its distribution law will not only help improve the detection and tracking performance of the radar to the measured target,but also help the advancement and development of radar clutter suppression technology.In summary,the simulation of radar clutter has very important practical significance and engineering value.The research content of this article is mainly from the following aspects:The statistical model of clutter is elaborated from two aspects: clutter amplitude probability density distribution and power spectrum model;based on this model,simulations of Rayleigh clutter,log-normal clutter and Weibull clutter are given.Methods and steps,and then put forward the overall process of generating one-dimensional radar clutter with various amplitude distribution based on the zero memory nonlinear transformation method(ZMNL)to transplant the algorithm to the hardware platform;Research and design the sub-modules in the clutter simulation process.The uniformly distributed random number generation module is the starting module of hardware clutter simulation.To solve the problem of poor whiteness of the uniformly distributed random number sequence generated by the linear feedback shift register(LFSR)commonly used in hardware design,this article focuses on clutter The first step of the simulation introduces the "one-step multi-hop" technology.It has been verified that the "whiteness" of the resulting uniformly distributed random number sequence is effectively improved and has excellent "uniformity";a comparative study based on transformation and based on Rejected-accepted two types of Gaussian white noise sequence generation algorithms,and obtained experimental conclusions: the transformation-based Box-Muller & CLT algorithm has a simple implementation structure,less hardware resource overhead,and the obtained Gaussian random number sequence accuracy meets the needs of the project and is more suitable for use in hybrid Wave hardware simulation.The Box-Muller & CLT algorithm is used to realize the generation of high-precision Gaussian distributed random number sequences.In order tomeet the requirement of the time-correlation of the radar clutter generated in the final simulation,this paper studies and designs the FIR Gaussian spectral filter.First,the theoretical expression of the filter coefficients is derived and processed,and then the length of the filter is analyzed and discussed in conjunction with simulation experiments,and the length is determined to be 24.The Gaussian white noise sequence is used to obtain the power spectrum as a Gaussian model.Related Gaussian noise.Finally,combined with the overall flow of the proposed algorithm transplantation and the design of each sub-module,the hardware simulation generation of Rayleigh clutter,log-normal clutter,and Weibull clutter is realized.The actual probability density curve,power spectrum curve and their theoretical curve of the generated clutter data are fitted and compared,and the maximum likelihood estimation method is used to estimate the data parameters,and the Chi-square test and K-S test are used to simulate the data.Test of goodness.The qualitative and quantitative test results verify the validity and reliability of each module design and algorithm transplantation in this paper.The relevant research and design of this paper have certain guiding significance for the engineering implementation of radar clutter simulator or clutter simulation module in radar simulator based on FPGA and other hardware platforms.