Research On Several Signal Processing Problems In Airborne Wideband Radar

Airborne wideband radar(AWR) has many advantages over traditional radar system in aspects like detecting range and spatial resolution. It has been widely used in many areas such as battlefield scout,precision-guidance,target identify,ground map,ground target imaging and etc. Combining with a research project, the work of this dissertation is focused on research on a few key techniques on AWR signal processing including signal receiving and detecting,imaging and real-time processing. The main contents of the dissertation are summarized as following:Since it is difficult to apply ideal matching filter to unstationary wideband echo signal. Chapter one tries to handle this problem via wavelet de-noising process after studying the principle of wavelet de-noising. An adaptive threshold de-nosing method based on stationary wavelet transform (SWT) is proposed, which can improve the de-noising and signal rebuilt performance when noise power become inconstant. The proposed method, on one hand, are able to overcome Gibbs effect which comes along with discrete wavelet transform (DWT); On the other hand, are able to adopt an adaptive algorithm to work out wavelet decomposing level and adjust the threshold value by estimating the current signal-to-noise ratio (SNR) of the raw signal. Furthermore, by simulating with real data gathered from real environment, we may draw to the conclusion that wavelet de-noising algorithm is effective and rationalitive when facing unstationary AWR signals.Aiming at parallel distributed detecting & fusing on wideband radar system. A theoretical analysis is presented in the sense of the Neyman-Pearson (N-P) test about the relationship between fusion rule and local decision rules in the parallel distributed detection fusion in chapter three. The determination of decision rules of two sub-optimal systems and globally optimal system are derived completely. The wavelet domain filtering method with unstationary wideband radar signals is combined with multi-sensor parallel distributed detecting & fusing system. The results obtained indicate that the detecting performance of wideband netted radar detecting system is remarkably improved by implementing wavelet domain filter algorithm and multi-sensor parallel distributed detecting and fusing system simultaneously.In chapter 4, the digital wideband quadrature receive techniques of wideband signal are discussed. A high speed wideband digital receiving system based on band-pass sampling and efficient polyphase filter banks technique is proposed. The problem that the calculating ability of current DSP devices can not satisfy the demand that wideband signal process requires is resolved. The math model of a high performance multi-channel digital receiver is deduced and established based on the definition of traditional channel division method and polyphase filter banks to achieve lossless receiving. A broadband/ultra-broadband signal receiving system based on frequency spectrum intersection and polyphase filter banks is proposed. Two key problems——high speed & precision sampling and high speed signal process in digital wideband receiver are resolved. According to the optimization to the low-pass filters and the simulations to the wideband signal pulse compression by sub-band process, the validity of this frequency spectrum intersection approach is well testified.Chapter 5 deals with the special problems of airborne wideband mechanical scan radar doppler beam sharpening (DBS) imaging. The influence of the parameters in DBS system are analyzed and gives the conclusion. The approach named "constant PRF using FFT method based on data reuse" is proposed. which keeps the sharpening rate constant using signal processing method basing on the condition that the antenna is running at constant speed and PRF. The proposed method can not only improve data utilizing rate, but also ensure real-time imaging speed of the entire system and reduce the aberrance probability for wideband imaging. Simulation and real data imaging results testified the validity of all the methods above. Some of the them have been implemented in an airborne fire-control radar.Chapter 6 discusses the despeckling problem of airborne radar DBS image. Radar image despeckling method is proposed based on Bayesian estimation and region division shrinkage in SWT domain. The despeckling problem of airborne radar DBS image is resolved. The statistic features of wavelet coefficients of radar DBS image in SWT domain is analyzed and described with mixed-gauss model. A shrinkage method is derived based on the minimum mean square error criteria with Bayesian estimation theory. The ideas of regions divided and shrinkage are adopted according to the intra-scale dependence of the wavelet coefficients. Wavelet coefficients are handled differently in different regions. Therefore, noise-free wavelet coefficients of each scale can be estimated more precisely and hence good despecling performance can be achieved.Chapter 7 investigates a few key techniques of AWR real-time parallel signal processing system such as system architecture and compatibility. A multi-DSP structure, based on distributed memory and local shared memory is proposed. This structure has good communicating performance among each node and good expansion capability. The scheme of an airborne radar parallel signal processing system based on standard bus and efficient processing module is designed. Part of the research results has been successfully applied in a few vast parallel signal process systems. Tests and outfield experiments show that compared to traditional designs, the scheme proposed in this paper is more universal,expandable and effective. It is very suitable for airborne wideband radar as it has high requirements like multi-modes,great task amount,high speed real time processing and reconfiguration function.Finally, a summary of the dissertation is given and suggestions of further work are also put forward.