| Frequency diverse array(FDA)radar forms range and angle dependent beam pointing by transmitting different signal carrier frequencies.This makes FDA radar have a great potential for applications in the field of secure communications and target location.Multiple Input Multiple Output(MIMO)radar combines FDA technology with MIMO technology,which not merely has the characteristics of range and angle-dependent of FDA radar beam,but also has the advantage of the high freedom of MIMO radar.In recent years,it has become a hot topic in the radar field.At present,most of the multidimensional parameter estimation algorithms for FDA-MIMO radar based on matrix analysis suffer from low accuracy and high complexity.The thesis investigates the joint multidimensional parameter estimation method for FDA-MIMO radar using tensor analysis as the mathematical theory.First,the mathematical theoretical basis of the thesis is introduced: the basics of tensor algebra.Second,the antenna characteristics and structure of the FDA are introduced,and the explanation of the causes of the FDA array with range and angle dependent beam pointing is given.Then,the derived signal model for the FDA-MIMO radar.Lastly,traditional subspace algorithms and traditional tensor decomposition algorithms for the joint parameter estimation of the FDA-MIMO radar are introduced.Simulation experiments are used to analyze the performance of each algorithm.For the problem of degraded performance of parameter estimation of FDA-MIMO radar with small samples,a joint parameter estimation algorithm of compressed unitary PARAFAC decomposition is proposed.First,the algorithm transforms the complexvalued tensor into a real-valued tensor with twice samples using the unitary transformation and forward-backward technique,solving the problem of performance degradation in small sample cases while reducing complexity.Next,a small real-valued tensor is obtained using the compressed matrix obtained from the HOSVD decomposition,which further reduces the complexity while retaining most of the tensor information.Then,the estimations of the direction matrices are obtained using PARAFAC decomposition.Finally,the angle and range estimations are obtained using least squares fitting.The superiority of the proposed algorithm is verified through complexity analysis and simulation experimental results.The joint parameter estimation algorithm based on fast PARAFAC decomposition of data expansion is proposed for the FDA-MIMO radar with high complexity and low accuracy.First,the proposed algorithm structures a signal model with the larger receiving array aperture by using unitary transformation and data expansion.Next,transmit direction and receive direction matrices are initialized by the idea of PM,solving the situation where the complexity of the PARAFAC decomposition process may be too high due to poorly chosen initial values.Then,the PARAFAC decomposition is performed to obtain the estimations of the direction matrices.Finally,the angle and range are estimated using the least squares fitting method.The experimental results demonstrate the superiority of the proposed algorithm. |
