Research On Array DOA Estimation Methods Based On Mixed-bit Quantization

Direction-of-arrival(DOA)estimation is to extract the direction of incident electromagnetic wave by using the signal received by antenna arrays,which possesses wide emerging applications,such as location,detection.In most current studies,DOA is directly estimated based on signals collected by arrays,that is,DOA estimation is based on unquantized data.However,in an actual system,the signal needs to be quantized by the analog-to-digital converters(ADCs)before subsequent signal processing operations.Although 1-bit quantization is currently used in DOA estimation,1-bit quantization will cause the loss of signal amplitude information,thus greatly affecting the accuracy of DOA estimation.Therefore,in order to meet the requirements of the actual system and reduce the DOA estimation performance loss caused by quantization,this thesis proposes several DOA estimation methods based on mixed-bit quantization under different scenarios,such as nested array,coprime array and coprime multiple input multiple output(MIMO)radar,aiming to reconstruct the unquantized covariance matrices by constructing and solving the corresponding optimization problems.The research content and novelty of this thesis are as follows:(1)A mixed-bit quantization method is proposed for DOA estimation on nested array.Firstly,the observed signal model of nested array under the mixed-bit quantization is derived,and the quantized covariance matrix is constructed by using the data after the mixed-bit quantization.Then,based on the low-rank matrix reconstruction theory,the optimization problem of restoring the unquantized covariance matrix in the difference coarray domain is constructed,and the convex optimization tool is used to solve it.Finally,the subspace-based method is used to estimate DOA based on the recovered unquantized covariance matrix.(2)A method of DOA estimation on coprime array by using mixed-bit quantization is proposed.Firstly,the received signal model under mixed-bit quantization is derived on the coprime array,and the virtual difference co-array signal is constructed by using the quantized data.Then,in order to make full use of the quantized information of the hole position in the virtual difference co-array,an interpolated quantized virtual difference coarray signal is constructed based on interpolation theory.Finally,the atomic norm theory is adopted to reconstruct an unquantized,noise-free matrix.This method can not only recover the unknown information of the hole position in the interpolated virtual difference co-array,but also reduce the information loss caused by mixed-bit quantization.(3)A method of DOA estimation using mixed-bit quantized data in the case of MIMO radar is proposed.Firstly,the signal model of the MIMO radar under mixed-bit quantization is derived,and the sum-difference co-array of the transmitting array and receiving array is interpolated by additional virtual sensors.Then an atomic norm minimization problem is established to further recover an augmented noise-free covariance matrix to estimate the DOA of the target.This method can not only recover the unknown information of interpolation position,but also suppress the influence of noise and quantization error.