Research On Waveform Temporal Diversity And Optimization Design For Cognitive Radar

Increasingly complicated working scenarios,as well as demanding performance requirements and diversified functional requirements from users are calling for new breakthroughs in radar technology.The development of the ability to control new degrees of freedom is the source of power to promote the progress of radar technology.The hardware of modern radar systems allows the transmission of waveforms that vary across space,time,and frequency and that can be changed in rapid succession,thus liberating a huge degree of freedom at the transmitter and giving birth to the vibrant research field of radar waveform diversity and design.Under such a technical background,the concept of cognitive radar came into being.Unlike a conventional radar which transmits a fixed waveform and works in a feedforward mode,a cognitive radar can learn about the target and the background through the interaction with the environment,adaptively adjust the transmitted waveform according to the feedback information from the receiver to better meet the mission requirements,and forms a dynamic closed feedback loop.This thesis is devoted to study temporal waveform diversity and design under the framework of cognitive radar.Focusing on improving the performance of radar in weak target detection in dense-passive-interference environments as well as in extended target parameter estimation and recognition,this thesis utilizes the idea of temporal waveform diversity and investigates the construction and application framework of two classic waveforms: ideal waveform with white amplitude spectrum,and matched waveform with coloured amplitude spectrum.The former is mainly used to prohibit the range sidelobes of strong scatterers from masking weak targets and improve the estimation accuracy of extended target parameters;the latter is aimed to improve the parameter estimation and recognition performance of extended targets.A waveform with an impulse-like aperiodic autocorrelation function or a white amplitude spectrum is called a ideal waveform because of its excellent performance in weak target detection,multi-target resolution,extended target parameter estimation,etc.Complementary codes with temporal diversity can be regarded as an implementation scheme of ideal waveforms,but it is seldom applied in practice because of the problem of Doppler sensitivity.To overcome this shortcoming,this thesis proposes a joint design method of transmitting pulse sequence and receiving pulse weights,which is used to construct a Doppler resilient complementary waveform from classical Golay complementary codes.The key characteristic of the waveform is that the waveform can always maintain excellent range sidelobe cancellation ability when the Doppler shift is within a given interval,and has a certain Doppler sidelobe suppression capability.The above-mentioned ideal waveform based on complementary codes is simple and easy to generate,especally suitable for the scenario with the requirement for Doppler resilience being not very high.For the scenario where high Doppler resilience is required and Doppler resolution and signal-to-noise ratio should be also taken into account simultaneously,this thesis relaxes the constraints of constant modulus,finite phase,and complementarity,and proposes a design method for Doppler resilient quasi complementary waveforms based on ambiguity function shaping.Because of the relaxed constraints and the full use of the degree of freedom provided by temporal diversity,this kind of waveforms has higher Doppler resilience,plasticity,flexibility and diversity.Based on the above research on wavefrom design,this thesis proposes a cognitive transmision scheme of Doppler resilience(quasi)complementary waveform to detect weak targets in dense-passive-interference environment.The scheme alternately transmits a traditional coherent pulse train and a Doppler resilient(quasi)complementary waveform.The former will perceive the scene and acquire the knowledge of the interference/target Doppler shift to guide the waveform design.The latter is used to suppress the sidelobe in order to improve the visibility of the weak target which is close to the strong interference both in distance and velocity.When the prior knowledge of the target is insufficient,the waveform with white amplitude spectrum can be regarded as an excellent parameter estimation waveform.However,once a certain level of target prior knowledge is available,the waveform with colored amplitude spectrum matching the scattering signature of the target tends to perform better.Based on this consideration,under the assumption that the target signature follows a Gaussian mixture distribution,this thesis proposes a matched waveform design method to improve the performance of target parameter estimation and recognition.The waveform thus obtained can maintain a good parameter estimation ability for individual target hypotheses and maximize the divergence between hypotheses.As an application platform of the above matched waveform,a cognitive radar architecture for extended target parameter estimation and recognition is proposed.Considering that the traditional way is easily affected by the target aspect sensitivity and fails to make full use of the correlation of the target signatures between observations,this thesis introduces the multiple model approach and exponentially-correlated model to describe the dynamic of the target signatures,and then develop a method to update the estimate of target signatures and the probability distribution of target hypotheses recursively in response to the latest echo.Using the recently updated knowledge of the target,the radar dynamically optimizes the next transmitted waveform to improve the estimation and recognition performance of extended targets.