Design Of Nonuniform Arrays And Underdetermined DOA Estimation

Array of direction of arrival(DOA)has a various application in many fields,such as wireless communication,astronomy,radar,and so on.For the array,the inter-element spacing may usually smaller than the half-wavelength of the incident sources to avoid the angle ambiguity.For a nonuniform array with specific structure,via exploiting the second-order statistics information,a virtual uniform array,called difference coarray(DCA)with inter-element spacing of half-wavelength can be constructed.Furthermore,the number of virtual sensors in the DCA is always larger than that in the physical array,which means more degrees of freedom(DOFs),and can solve the problem of underdetermined DOA estimation,i.e.resolving more sources than sensors in the term of directions simultaneously.In addition,more DOFs,better estimation performance.Thus,in this thesis,we focus on the designing of nonuniform arrays,and propose two nonunifom linear arrays and two nonuniforn planar arrays with increased DOF or reduced mutual coupling.Comparing with existing arrays,our proposed arrays can resolves more sources,and has better performance for DOA estimation.The specific works are summarized below:1.The design of nonuniform linear array.By improving the popular two-level nested array(NA),we propose the improved nested array(INA),which has all the good properties of the two-level NA,and enjoys larger aperture and more DOFs.The INA can provide a larger hole-free DCA.Simulations demonstrate that the INA provides more DOFs,resolves more sources,and has better DOA estimation performance than the two-level NA.For the MIMO(Multiple Input and Multiple Output)radar with multiple transmitting and receiving sensors,in order to obtain a hole-free DCA with as many DOFs as possible,we propose a nested MIMO array.The nested MIMO array can provide(46)(K~4)DOFs in the difference coarray of the sum coarray(DCSC)using only K physical sensors,and the resulting of DCSC is hole-free.Furthermore,the nested MIMO array has closed-form expressions for the sensor locations and applicable for the design of large-scale arrays.Simulations demonstrate that the nested MIMO array can provide more DOFs,resolve more sources,and have better DOA estimation performance than the existing coprime MIMO array.It should be noted that the array design approach is also applicable for all existing sparse arrays with hole-free DCA,such as the super nested array(SNA),the INA aforementioned,and so on.2.The design of nonuniform planar array.For the sake of increasing the DOF in the nonuniform planar array,we propose a minimum redundancy planar array(MRPA)based on the minimum redundancy linear array(MRA).The projections of our proposed nonuniform planar array on the X axis and Y axis are both MRAs,and the resulting DCA is hole-free.By comparing with the open box array(OBA)with good performance,the MRPA provides more DOFs,resolves more sources,and has better DOA estimation performance than the OBA.To reduce the mutual coupling among sensors in the planar array,we propose a new array called thermos array based on the hourglass array,which is more robust to the mutual coupling than the hourglass array by rearranging the positions of sensor pairs with smallest separation.For the geometry of our proposed array resembles a thermos,we name it as thermos array.Comparing with the thermos array,the thermos array has fewer number of sensor pairs with smallest separation,thus is more robust to the DOA estimation when the mutual coupling is existed.