Sparse Array Geometry Optimization And Parameter Estimation Based On Polarization Sensitive Array

In recent years,a growing number of electromagnetic devices make us convenience but also cause the electromagnetic environment complicated gradually.The new greater challenges have occurred in the various fields,such as electronic reconnaissance,navigation radar,etc.So how to accurately estimate the target parameters in complex electromagnetic environment is worth researching.The traditional scalar antenna can only obtain amplitude,phase and frequency of the target signal,and cannot pick up the polarization information of the signal and utilize it.Thus,people began to pay attention to the polarization sensitive array?PSA?,which can achieve and effectively use the polarization information of the target signal,so this array does a great deal to improve performance of the signal processing.Conventional array signal processing is mostly based on the homogeneous array with half of wavelength as element spacing.In order to achieve the required radiation performance with as few elements as possible,we have used the simulated annealing algorithm to optimize the homogeneous array with fixed aperture and regarded spatial gain of the antenna as the cost function,which can minimize the number of elements in the array while reducing the sidelobe level.In practical project,due to the high dynamics of the array or the fast movement of the target,the number of samples is not enough,the subspace class algorithm can not accurately estimate the covariance matrix,which leads to the decline of algorithm's performance.Moreover,the multipath delay wave caused by environment reflection is coherent with the target signal direct wave.Thus,Subspace class algorithm is no longer valid.Aiming at such problems,people have been studying the sparse recovery method.Often the sparse recovery method of existing technology cannot accurately estimate the parameters because the grid is too dense and the correlation among the columns of dictionaries is too high to lead to the sparse recovery performance loss,in addition,the sparse recovery method is of high computation cost.In order to solve the problems mentioned above,this thesis proposed a method to estimate the target parameters with alternating grid optimization.This method constructs the data observation vector by performing element denoising on the array receiving data.In the iterative process,use the dynamic increment to estimate the source num.Using null space tuning with hard threshold and feedback to obtain the 1^st target DOA vector and the recovery vector.And based on the recovery vector,we need to find the peak index,define the ratio of multi-target peak power and the sum of other grid power as the cost function and compute the current value of cost function,and compare the current value with the values after grid left or right shifting.Finally,we can update the current iterative parameters,and improve the estimation accuracy of multi-target parameters.The results of simulations show that the method we proposed is suitable to both uniform and sparse arrays,and can also be used to estimate the incoherent or coherent sources parameters in the case of small samples.In this thesis,first,we utilize the proposed method to estimate the direction of the incoming wave by the monopole array,and then improve the estimated precision of spatial angles using weighted fusion,finally,we select single dipole and single magnetic ring of the common position to estimate the polarization parameters of signal.In conclusion,compared with the traditional parameter estimation method,the proposed method can obtain better parameter estimation performance in the case of small sample and coherent source by combining with polarization sensitive array.