| Traditional radar waveforms usually consist of one or more fixed waveforms,which are received by the receiver and then resolved by the corresponding signal-processing algorithms for target information.Since traditional radar systems cannot utilize the environment as a priori information,it greatly limits the performance of radar systems.The emergence of cognitive radar overcomes this challenge by sensing a priori information such as the environment and then feeding this a priori information back to the transmitter,where it is used to optimize the transmit waveform for the next moment,improving the performance of the radar target detection and estimation,among others.Therefore,the study of waveform optimization design for cognitive radar has become a hot issue at present.Due to the complexity and variability of radar application scenarios,which lead to uncertainties in a priori information such as environment,Lagrange multiplier waveform design methods based on different criteria have been widely studied by scholars.In this thesis,two radar transmit waveform design algorithms are proposed for the target detection and estimation tasks,and for the problem that Lagrange multiplier methods are not easy to implement.Divide And Conquer(DAC)algorithm,which is a single-target radar transmit waveform design algorithm.Considering two cases of noise presence only and noise and clutter simultaneously,the traditional water-filling algorithm and DAC algorithm are used for the design of optimal single-target radar transmit waveform based on the mutual information and signal-to-interference-plus-noise ratio criteria,respectively.Based on the energy constraints,the performance and efficiency of the two algorithms are compared and analyzed respectively.The simulation analysis shows that the waveform designed by the DAC algorithm has a better energy spectrum and a larger mutual information value(signal-to-interference-plus-noise ratio value),and thus can improve the radar target detection and estimation,while achieving a substantial improvement in efficiency,making it more conducive to engineering implementation.Divide And Conquer Linear Weighted Sum(DACLWS)algorithm,which is an extension of single-target radar waveform design to multi-target radar transmit waveform design.Based on the traditional single-target water-filling algorithm,a multi-target model is established by using the weighted summation method,and finally,the objective function of the transmit waveform is derived.Based on the energy constraint,the simulation comparison analysis of DACLWS and LWS algorithms under the mutual information,signal-to-interference-plus-noise ratio and detection probability criteria,respectively,shows that the performance(optimal waveform spectrum,mutual information value,signal-to-interference-plus-noise ratio value and detection probability)of DACLWS algorithm is better,i.e.,the target detection and estimation capability is better,while the efficiency improvement is more than that in single-target.The location of the optimal solution of the transmit waveform changes at different transmit energies,and the traditional optimization methods cannot design the optimal transmit waveform at different transmit energies simultaneously due to the influence of the search accuracy,which affects the performance in radar target detection and estimation.To address this problem,a radar transmit waveform design algorithm with stable energy utilization is proposed in this thesis.The Golden Section Genetic Algorithm(GSGA),which is a multi-target-based radar transmit waveform design algorithm,is improved based on the golden section strategy.Based on the energy constraint,the energy spectrum,energy loss rate and detection probability of the waveforms designed by GSGA and LWS algorithms are compared and analyzed under the criteria of mutual information,signal-to-interference-plus-noise ratio and detection probability,respectively.Finally,the energy loss rates of GSGA and DACLWS algorithms are compared and analyzed under different criteria with different accuracy constraints.The simulation analysis shows that the GSGA algorithm has a lower and more stable energy loss rate and can be applied to target detection and estimation tasks at different energies. |
