Research On Underwater DOA Estimation Algorithm Based On Wideband Signals

Direction of Arrival(DOA)estimation has a wide range of applications in radar detection and underwater target positioning.With the increasing frequency of human development and exploration activities in the ocean,underwater DOA estimation has become increasingly important role.Underwater signals are often in a harsh signal-to-noise ratio environment.The method of increasing the number of sensor elements to ensure the accuracy of the estimation will increase the hardware cost.How to improve the underwater DOA estimation accuracy becomes an important issue.Underwater DOA estimation generally selects sound waves as the signal carrier.The propagation speed of sound waves is constantly affected by various factors.The existence of sound velocity errors will reduce the estimation performance of conventional DOA estimation algorithms and become an important issue for underwater DOA estimation.Existing narrow-band signal velocity-independent algorithm eliminate the sound velocity error by adding arrays and relying on the angular relationship between the two arrays,but the sound velocity search process performed to eliminate the speed of sound increases the computational complexity and the estimation accuracy is affected by the search step.Compared with narrowband signals,wideband signals usual y contain more information about the angle characteristics,which is beneficial to obtain better DOA estimation results.Therefore,DOA estimation of wideband signals has received more attention.In this paper,combining the wideband signal model,the underwater DOA estimation accuracy problem and underwater sound velocity error problem are studied.The main work is as follows:1.Introduced the theoretical basis of underwater DOA estimation,including narrowband and wideband signal models,coherent source model and decoherence technology,acoustic velocity theory and DOA estimation performance indicators;For existing wideband signal underwater DOA estimation algorithms,Including ISSM-MUSIC,TOPS,RSS,TCT and consistency focus algorithm are introduced and simulated experiments,and the characteristics of existing algorithms are analyzed.Aiming at the problem of sound speed error in underwater DOA estimation,the influence of sound speed error on underwater DOA estimation is analyzed,and a current speed-independent algorithm namely the VI-MUSIC algorithm is introduced.2.The Cramér–Rao Bound derivation of the wideband signal model is given,and the theoretical lower limit of the performance of the wideband signal underwater DOA estimation is given.The final expression shows that the CRB value estimated by the wideband signal model DOA is related to the number of snapshots,array element spacing,signal-to-noise ratio,source angle,number of array elements,and selected frequency points.3.Aiming at the problem that conventional wideband signal underwater DOA estimation algorithm does not improve the estimation performance when the bandwidth and frequency points increase,an underwater DOA estimation method based on multiple frequency wideband signals is proposed.The received signals at each frequency point are subjected to joint matrix reconstruction,so that the elements in the obtained joint covariance matrix include phase difference information corresponding to a larger array aperture.Simulation results show that the method has smaller root mean square error of DOA estimation under the same number of array elements and the same frequency.4.Aiming at the problem of increased computational complexity caused by the sound speed search process of the VI-MUSIC algorithm,a sound speed-independent DOA estimation algorithm(ROOT-ISSM-VI-MUSIC algorithm)based on wideband signals without the need for sound speed search is proposed.Obtain multiple sets of roots that contain the source angle and sound speed information corresponding to each frequency point signal,and average them to obtain the average roots corresponding to the sub-arrays.According to the angle relationship between the sub-arrays,get the expression of the actual sound velocity about the two sets of average roots,Obtaining the DOA estimation value independent of the sound speed,avoiding the sound speed search process.At the same time,the simulation results show that the ROOTISSM-VI-MUSIC algorithm has lower RMSE of the estimated sound velocity and angle.