Two Dimensional DOA Estimation Base On MST And Its Implemention

Direction-of-arrival(DOA)estimation is animportant branch of array processing.Since decades of development,the theoretical basis is very rich and varied,either the earlier stage of minimum variance method made by Capon,or the Multiple signal classification method by Schmidt,even the maximum likelihood estimation,the computational complexity isn't quite acceptable.Especially some of these methods refer to the matrix operation,including the matrix mathematical decomposition,addition,subtraction,etc.This always makes limitation to the application of the algorithm.With the development of the semiconductor industry,monolithic integrated circuit gets more higher density,that allows the implementation of complex algorithm coming true.At an earlier stage,most of the developers chosen digital signal processors for most,but later on,things are benefit from the development of Field-Programmable Gate Array,more and more developers choose FPGA chips instead.Based on the two-dimensional DOA estimation algorithm based on MST,this thesis is mainly focused on the implementation and its detail improvement.This thesis is summarized as follows:First of all,we briefly summarize theresearch progress of adaptivetwo-dimensional DOA estimation,and analyze the structure model of antenna array.We use uniform linear array and uniform circular array for examples.Secondly,based on the algorithm of DOA,we introduce the Multiple signal classification method.Furthermore,we discussed the advantage of MUSIC,such as higher resolution and asymptotically infinite border.Subsequently,the disadvantage is also notable,the complexity,we mean it's hard to calculate and achieve.Finally,we discussed MST and two-dimensional DOA estimation algorithm based on MST.We achieved it on an FPGA chip.the most difficult and complex part is matrix decomposition operation,we overcame it by using bilateral Jcobialgorithm.during the procedure,this thesis proposed some improvement method making use of FPGA's characteristics to process the non-linear computation.thecordic method is adopted to simplify the calculation of matrix's decomposition,it significantly reduced the usage of FPGA's resource,and increased the accuracy.Through the simulation analysis and comparison,it's proved that the adaptive algorithm is effective.