Research On MIMO Radar Waveform Design Based On Manifold Optimization Algorithms

MIMO radar is a novel radar system that breaks through the limitations of conventional phased array radar and can be flexibly optimized via waveform design algorithm.Waveform design is the key technology to achieve the advantages of MIMO radar systems,possessed more degrees of freedom at transmitting and processing.However,waveform design problem with constant modulus constraint is usually nonconvex high-dimensional NP-hard,generally cannot be solved in polynomial time.Fortunately,a kind of novel optimization algorithms developed on manifold optimization framework can efficiently solve such problems.In this thesis,according to the importance of waveform design and the advantages of manifold optimization algorithm,some method innovations are developed to suppress signal-related interference with definite directions.The main workload of this thesis is as follows:1.Focus on the two-step transmit waveform design problem based on maximizing SINR.With the idea of setting radiation nulls on the interference directions,the problem is modeled similar to waveform synthesis method,then decomposed and reconstructed the problem to obtain a closed-form solution.The beampattern performance and SINR with proposed closed-form method are better than contrast methods,the computational complexity of proposed method is also lower.2.Focus on the one-step radar waveform design problem based on maximizing SINR.To address with the difficulties of this problem,this thesis introduces a novel optimization algorithm framework: the Riemannian manifold optimization framework.The Riemannian optimization algorithm transforms the original NP-hard non-convex high-dimensional problem into an unconstrained optimization problem in a constrained space,then efficiently solved it via quasi-Newton algorithm.Numerical experiments show that the proposed method has good convergence properties,and achieving higher SINR compared to contrast methods.3.Focus on the MIMO radar transmit-receive joint design problem based on maximizing SINR.The original problem can be modeled as a joint optimization problem of the transmit waveform and the receive filter weight vector.Then,an alternating optimization method was proposed for this problem.The proposed method solved receive filter weight vector by the minimum variance distortionless response method,solved transmit waveform by quasi-Newton method.Numerical experiments show that the proposed method achieve better runtime and SINR compared with the contrast methods.