A Study On Adaptive Waveform Optimization Design Algorithm For Cognitive Radar

Cognitive radar is a new type of radar system for the closed-loop systems,and the optimal design of the waveform is the key technology of the cognitive radar.The introduction of the cognitive radar makes the adaptive technology extend to the waveform design of the transmitter.The prior information of the target environment can be used to optimize the next transmit waveform,thereby the performance of the radar detection and target recognition can be improved.There are two main optimization criteria to design waveform,which are the information theory related to radar target recognition and the signal-to-noise ration closely related to the detection probability of radar targets.Based on this,the constrained optimization problem is established with constraints such as energy and constant modulus,etc.Some effective algorithms to solve the constrained optimization problem are researched and proposed in this article.An IGA-NP algorithm is proposed to solve the optimal waveform problem of the maximum mutual information with energy constraint,under the condition that the spectral variance distribution of the target and clutter has been known in the target surroundings.To get a more accurately solution and accelerate the convergence speed,the algorithm built a crossover and mutation method which satisfies the constrained optimization problem and adds the search procedure and the nonlinear programming process that fit the problem.It is verified that the optimized power spectrum distributes the energy to the frequency band where the variance of clutter power spectrum is weak but the variance of target spectrum is strong.Under the condition that the spectral distribution of the target,clutter and noise is known,the adaptive waveform optimization problem of the maximum the signal-to-noise ratio under the constraints of the energy and constant modulus is studied.Firstly,based on the single-shot signal model,the IGA-SA algorithm is proposed.To improve the convergence and time complexity,the algorithm optimizes the phase of the signal directly and applies the simulated annealing strategy to the interfered chromosomes,making the algorithm win the ability of steadily rising and jumping out of the local optimal solution.The effectiveness of the algorithm has been verified by the simulations.Then,the waveform optimization problem that maximizes the signal-to-noise ratio in the multi-transmit system mode under the constant-mode constraint is studied.The simulation verifies that the SCNR under this model is indeed greatly improved,but the orthogonality among the multi-transmit waveform is lost.Therefore the constraint condition for guaranteeing the orthogonality between signals is added to the original problem model,and the method that selects the signal with the best orthogonality is added to the algorithm when a random signal satisfying the conditions is generated.Under the condition,a new optimization model is established and a better result has been obtained using the algorithm of semi-deterministic relaxation and improved Gaussian randomization.The simulation results show that the algorithm can improve SCNR of the system steadily on the basis of guaranteeing the orthogonality between multiple transmit waveform as much as possible.However,since the Gaussian randomization method is used in the algorithm,the stability of the entire algorithm needs to be further improved.