Adaptive Waveform Optimization For Cognitive Radar Parameter Estimation

In existing radar systems, transmitted waveforms and working modes are typically fixed, and receiving systems passively process radar signals and data. Therefore even the well-designed radar systems are difficult to obtain the desired processing performance in a complex spatial and temporal changing environment. With a closed-loop feedback mechanism from the receiver to the transmitter, cognitive radar is able to dynamically tailor the radar transmitter and receiver according to a priori and real-time knowledge of environments and targets. It alters the open loop information processing mode of traditional radar system fundamentally and has exhibited significant performance improvements and adaptability to changing environments. Cognitive radar has important theoretical and military value and attracted widespread attentions from many scholars and military experts at home and abroad. Adaptive transmit waveform optimization technology is one of the key technologies of cognitive radar system and the paper mainly studies the adaptive radar transmission system optimization for target parameter estimation. Specifically, the content of the paper can be divided into two aspects: target tracking(moving target state estimation) and radar imaging(stationary target electromagnetic scattering coefficient estimation).In the aspect of target tracking, the paper investigates the adaptive waveform selection technique to improve the maneuvering target tracking accuracy in clutter. Firstly, the receiving system and the process of radar measurement acquisition are modeled in detail,followed by an analysis and quantitative calculation of the effects of different transmitted waveforms on the statistical characteristics of radar measurements under different SNR. A novel CF-HFM(Constant Frequency-Hyperbolic Frequency Modulation) radar burst waveform is also designed to mimic the bats emitted acoustic signals, and then a task-based dynamic waveform selection algorithm and a joint waveform and detection threshold optimization algorithm for maneuvering target tracking in clutter are proposed.Secondly, In order to more accurately estimate the posterior hybrid state probability density function of maneuvering target, a novel interacting multiple model probabilistic data association particle filter(IMMPDAPF) are proposed, in which base state estimation and modal state estimation are completely separated. In this new filter, only base state particles are needed and the number of particles in each maneuvering mode can be controlled.The more robust Amari α-divergence is introduced to measure the information gain of radar measurements for different transmitted waveforms, and then an information based dynamic waveform selection algorithm for maneuvering target tracking in clutter is proposed, which is close related to IMMPDAPF implementation procedure.In the aspect of radar imaging, the paper investigates the adaptive transmit diversity to reduce the complexities of radar imaging and transmit diversity based curvilinear trajectory synthetic aperture radar(SAR) imaging algorithms. Firstly, as for the curvilinear trajectory SAR imaging problem for image matching guidance, an adaptive adjustment approach of the radar beam pointing for imaging in an equivalent broadside geometry is proposed, which can significantly reduce the difficulty of range migration correction and the coupling between range and azimuth. And then the quick look Range DopplerDechirp(RD-Dechirp)imaging algorithm is designed based on partial aperture data and efficient spectral analysis method. Secondly, as for the high squint curvilinear trajectory SAR imaging problem for homing guidance, the modified nonlinear chirp scaling(NCS) algorithm for curvilinear trajectory SAR imaging is proposed after introducing an adjustable nonlinear frequency modulated component to the transmitted linear frequency modulated signal. The new algorithm can simultaneously handle the dependence of range cell migration correction and secondary range compression on azimuth frequency and range, while its computational complexity is at the same level with the standard CS and RD algorithms. Therefore it can provide a perfect compromise between the imaging performance and the computational complexity. Finally, as for the spotlight SAR imaging problem with arbitrary trajectory and squint angle, a pulse repetition interval(PRI)diversity based polar format algorithm(PFA) is proposed, which can avoid interpolation operation and take advantage of the full wavenumber domain data. It only requires FFT and complex multiplication operations to format high-quality SAR image and take into account the trajectory flexibility, imaging precision and efficiency simultaneously.By making full use of the feedback mechanism from the receiver to the transmitter and a priori knowledge, the paper investigates the transmission system optimization technology for target parameter estimation problem. Its research results is helpful to change the working mode of existing radar systems fundamentally and to improve the radar system performance and adaptability. Our research provides theoretical and technical supports for the development of intelligent radar and cognitive radar.