Research On Method Of DOA/TDOA Estimation Based On Spatio-temporal Cyclostationarity

In recent years,direction of arrival(DOA)and time difference of arrival(TDOA)estimation as an important means of radiation source location has been a hot topic in the field of radio reconnaissance.In this dissertation,the problems of DOA/TDOA estimation of cyclostationary signal under multipath conditions,high-speed motion scenes,spatio-temporal reception conditions or existing amplitude and phase errors are studied.DOA/TDOA estimation methods based on spatio-temporal cyclostationary are proposed.The main contents are summarized as follows:1.Based on the cyclostationary DOA estimation model,Cyclic-MUSIC algorithm and Toeplitz-Cyclic-MUSIC algorithm are proposed to improve the DOA estimation performance of first-order statistically uncorrelated but cyclic second-order statistical correlation signals.Based on the almost cyclostationarity(ACS)DOA estimation model,method of constructing virtual nested array and DOA estimation algorithm based on compressed sensing are proposed.The algorithm effectively extends the array degreeof-freedoms(DOFs)and improves the DOA estimation performance when the number of signals is larger than the number of array elements.Based on generalized almost cyclostationarity(GACS)DOA estimation model,a GACS DOA estimation algorithm based on delay correlation is proposed to improve DOA estimation performance for ACS signals in the high-speed motion scene.2.Based on the single-channel spatio-temporal cyclostationarity DOA estimation model,a single-channel Cyclic-MUSIC DOA estimation algorithm is proposed.Compared with the conventional spatio-temporal DOA estimation algorithm,the performance of the algorithm is not affected by the signal bandwidth and switching time,and the algorithm can accurately estimate DOA of the target signal when the switching time error is known.Based on the dual channel spatio-temporal cyclostationarity DOA estimation model,a dual channel Cyclic-MUSIC DOA estimation algorithm is proposed,and the algorithm can accurately estimate DOA of the target signal when the switching time exists unknown error.Toeplitz-Cyclic-MUSIC is extended to the spatio-temporal model,and the spatio-temporal Toeplitz-Cyclic-MUSIC algorithm is proposed.By decoherent processing of the cyclic autocorrelation data,the DOA estimation performance for first-order statistically uncorrelated but cyclic second-order statistical correlation signals is improved.By applying the dual-channel Cyclic-MUSIC algorithm,the volume and weight of the direction finding receiver is reduced because only two receiving channels are needed and the size of the antenna is reduced by the expansion of the antenna aperture using cyclostationarity characteristic.The size and volume of the positioning system satisfies the requirements of the UAV platform.3.Based on the cyclostationary characteristics of the external auxiliary source signal,a algorithm based on unknown azimuth auxiliary source signal is proposed.Using the cyclostationary characteristics of the external auxiliary source signal,the amplitude and phase errors of the array are corrected.Then based on the spatial smoothing ESPRIT algorithm,the direction of arrival of the auxiliary source signal and the target signal is roughly estimated.Finally,vased on the rough estimation results,the iterative algorithm are used to accurately estimate DOAs of signals,amplitude and phase error of the array,and multipath decline coefficients,which reduces the loss of estimation accuracy due to the reduction of the effective aperture because of spatial smoothing.4.Based on the generalized almost cyclostationarity of chirp signals,a phase difference estimation algorithm based on multi-cycle frequency and a spectral correlation TDOA estimation algorithm are proposed and a method is given to resolve the phase fuzzy problem of multi-cycle frequency phase difference estimation method.Based on generalized almost cyclostationarity,a TDOA estimation algorithm based on fractional Fourier transform is proposed for high-speed moving target signals.By applying the multi-cycle frequency phase difference estimation algorithm to the distributed positioning system which estimates the time difference according to the phase of the cyclic autocorrelation data,the transmission bandwidth requirement of the actual system is reduced because only the cyclic auto-correlation data calculated according to the time domain data are needed to be transmitted to the center station from the receivers,which reduces the data transmission amount of the distributed positioning system.