Research On High Precision And Super-Revolution Direction Finding Techniques In Passive Radar Seeker

Spatial spectrum estimation technique can accomplish super-resolution of direction of arrival (DOA) of signals. It can resolve multiple signals which arrive simultaneously, and it is used by passive radar seeker (PRS) to accomplish super-resolution direction finding. The forth-order cumulants (FOC) has performance of virtual array extending and inhibition from Gaussian colored noise, so FOC is widely applied to pretreatment of spatial spectrum estimation algorithms. Spatial spectrum estimation algorithms based on FOC and solving ambiguity methods in direction finding of spatial spectrum estimation were studied primarily in this dissertation. The design and realization of the soft-hardware scheme in direction finding processor of spatial spectrum estimation was finished and correlative indexes were tested.PRS can't resolve multiple signals with limited antenna disk volume. On this problem, MUSIC algorithm based on FOC was studied, and a modified algorithm which has low computational complexity was proposed. Lossless dimensionality reduction of FOC matrix removed data redundancy and improved real-time performance. Performance of virtual array extending was reserved at the same time. That can resolve multiple signals using fewer antennas with limited antenna disk volume. Both computer simulation and practical measurement data simulation prove that the algorithm had good estimation performance with Gaussian colored noise, and could estimate multiple signals.Replacing MUSIC algorithm by estimation of signal parameters via rotational invariance techniques (ESPRIT) algorithm, searching spectrum peaks was avoided. That reduced computational complexity of FOC spatial spectrum estimation algorithms effectively. A modified FOC ESPRIT algorithm was proposed. Based on the new array created by virtual array extending, ESPRIT algorithm was used to estimate DOA of signals. FOC matrix which has no redundancy was regarded as data covariance matrix of new array. And a method of dividing sub-array which adapted all linear arrays was designed. Modified FOC matrix and ESPRIT algorithm increased computation velocity of FOC spatial spectrum estimation algorithm.FOC MUSIC algorithm and ESPRIT algorithm couldn't resolve coherent signals. So introducing spatial smoothing technique to FOC algorithms, FOC spatial smoothing algorithm was studied. Spatial smoothing technique was used to estimate coherent signals with new array created by virtual array extending. Simulation indicates that the algorithm can estimate DOA of coherent signals, and couldn't lose antenna aperture, when the antenna number of sub-array isn't less than that of real array.Direction finding processor of spatial spectrum estimation should produce angle ambiguity. For this, the methods of solving ambiguity by twice searching based on MUSIC algorithm and virtual array extending were proposed. So covering frequency of PRS would be widened. The method of twice searching can estimate DOA of single signal quickly and effectively with the limited searching section which just had a spectrum peak of real angle. The method of virtual array extending resolves ambiguity effectively with non-uniform linear array with Gaussian colored noise. That offered a new thought for solving ambiguity.Direction finding processor of spatial spectrum estimation needs good precision and high real-time performance in practice, so four chips of float ADSP-TS101S were used to accomplish searching spectrum peaks of MUSIC algorithm. Spatial spectrum estimation algorithm represents good performance in the index test of PRS. It doesn't only estimate DOA of single signal with high precision, but also estimates DOA of multiple signals which arrived simultaneously. The design and implementation of the hardware system settled a basis for the new type of PRS, and promoted engineering application steps of spatial spectrum estimation technique.