Research On Anti-interference Waveform Optimization For Radar

A traditional radar system cannot alter its probe waveform at run time, hence the waveform design algorithm should take all of the performance indices into consideration to generate a perfect waveform. However, the degree of freedom for waveform synthesis is limited, consequently, it is impossible to design a perfect waveform for all scenarios.Nowadays, with the development of radar hardware, on-line waveform synthesis becomes possible, consequently more and more researchers started to study algorithms for on-line waveform design. Since the radar can change its probe waveform on the fly, there is no need to design a perfect waveform for all scenarios. The designed waveform can be custom-tailored for current scene, thus the number of the constraints is reduced, the designed waveform will perform better than its traditional counterpart in a specific scene.Generally, the on-line waveform synthesis technique is comprised of two components: 1) an estimator to estimate the parameters of the current scene, and 2) a waveform design algorithm. Up to now, researchers have already proposed numerous algorithms, unfortunately, all of these may have some certain flaws, which make them unsuitable for some specific radar scenes. This project is dedicated to provide new algorithms for solving the problems encountered.Firstly, An algorithm framework named Iterative Spectral Approximation Algorithm(ISAA) is proposed to solve the range sidelobe masking problem. After the concept of pulse compression had been introduced in radar systems, the range sidelobe masking becomes a problem that every radar designer must takes it into consideration. The range profile output from matched filter can be seen as the convolution of the radar scene and the autocorrelation of the probe waveform. For this model, the autocorrelation function acts as the point spread function in image blur model, which spread the echo energy from one range cell to other range cells, and consequently, the range profile is blurred. Unlike optical system, an active radar can control its probe waveform, which means once the blurred range profile is obtained, the radar can design a waveform with specific autocorrelation shape to minimize the range sidelobe of strong scatterers located in specific range intervals. Based on this idea, ISAA can synthesize waveforms with low autocorrelation magnitude in specified intervals. This algorithm framework is based on power spectral density fitting, phase retrieval, and alternating projections, and a unique method named Dynamic Ideal Autocorrelation Construction(DIAC) is proposed. An effective waveform design algorithm can be obtained by combining the ISAA framework and DIAC method, which is faster than similar algorithms.Secondly, an algorithm framework named Phase-Only Nonlinear Programming(PONLP) is proposed. Similar to ISAA, PONLP can be used to synthesize waveforms with specified autocorrelation shape. However, PONLP has its unique advantage: it can be used to synthesize waveforms with many narrow autocorrelation nulls. This advantage makes PONLP an ideal method for designing clutter suppression waveforms. Unlike ISAA, PONLP is a derivative-based optimization algorithm. Compared with traditional derivative-based algorithms, PONLP replaces the gradient and line search with phase-only gradient and phase-only line search; as a consequence, the envelope constraint is included implicitly, and the convergence rate is improved.Thirdly, considering that many active interferences may exist in the scene for today’s radar systems, the concept Rank-Deficient Fourier(RDF) transform is introduced and being applied to the ISAA framework. The result is an extended algorithm framework which is capable of designing waveforms with specified autocorrelation and spectral shape. Traditional radar system can utilize a band rejection filter to suppress active interferences located in certain bands. However, if the probe waveform and interference are overlapped, not only the interference but also the partial energy of the echo will be suppressed. On the other hand, by designing the spectral shape of the waveform, the overlapped band width of the waveform and interference can be minimized. Similar to the extended ISAA framework, an extended PONLP framework based on RDF is proposed. Both of these frameworks can accomplish their assigned tasks.Fourthly, several extended versions of ISAA are proposed for multiple waveforms design. These versions include an ISAA based algorithm named ISAA-Cross Correlation Optimization Inner Product Constraint(CSOIPC) for transmit-receive joint optimization, and algorithm named Multi-Dimensional ISAA(MDISAA) for MIMO orthogonal waveforms design, and an algorithm framework named Alternating Projections for designing Locally Orthogonal Waveform Pairs(APLOWP) based on Generalized ISAA(GISAA) which can be used to synthesize waveform pairs for Instantaneous Polarization Radar(IPR).Finally, based on the characteristics of the practical radar scenes, several simulated radar environments are provided for testing the proposed algorithms. Numerical simulations have demonstrated the effectiveness of the proposed algorithms.