Study On The DOA Estimation Algorithm For Coherent Signals

The DOA(Direction-Of-Arrival) estimation of coherent signals is a research hot spot of array signal processing, which is widely used in radar, communication and sonar etc. In order to solve the problem of the DOA estimation of coherent signals, a lot of algorithm were proposed by the researcher of all over the world. All the algorithms can be departed into two types: decreasing dimension algorithms and nondecreasing dimension algorithms.The spatial smoothing algorithm is an effectively DOA estimation algorithm of coherent signals, Its performance is ideal with little complexity and easier to be accomplished. However. The effective aperture of array will be reduced because the ability of decorrelation was gain based on sacrificing the effective number of arrays, and, in the circumstance of lower SNR(signal-noise-ratio) , the estimation performance is deteriorated. In the thesis, The author proposed some modified algorithm after analyzing the algorithm of spatial smoothing and the questions of existing in them. The main work of the thesis is in several aspects as follows1. Firstly, The conventional algorithm of DOA estimation, such as MUSIC algorithm and ESPRIT algorithm, are analyzed. Then, the problem that the performance of DOA estimation will be deteriorated with conventional algorithm under the circumstance of coherent signals, is demonstrated. Some simulation results are given.2. The decorrelation principle of the kind of spatial smoothing algorithm is demonstrated. Then, some effective algorithm of spatial smoothing, such as forward spatial smoothing(FSS), forward backward spatial smoothing(FBSS) and improved spatial smoothing algorithm , were introduced. In the FSS algorithm and FBSS algorithm, only the autocorrelation of outputs of the individual subarrays are utilized to obtain the final estimate of the covariance matrix, without considering the cross correlations of the subarray outputs. The algorithm of Improved spatial smoothing which is capable of increasing the effectiveaperture of the array over that of conventional spatial smoothing methods can fully utilize the correlation of array outputs and produce a more stable estimate of the covariance matrix. Up to now, the algorithm we have mentioned are based on the reducing of the effective aperture of the array to decorrelation. the numbers of signals which can be distinguished are reduced consequently.3. Contraposition the problem of the spatial smoothing algorithms > A modified MUSIC algorithm(MMUSIC) is proposed which doesn't reduce the effective aperture of array. It has been demonstrated that the average correlation coefficient will theoretically fall to 63 percent. The fault of MMUSIC algorithm is that when the DOA of coherent signals is vary closer , it may fall to yield the desired performance of DOA estimation. Then, a modified MMUSIC algorithm is proposed in the thesis , which can discern the DOA of one degree interval of two coherent signals, offers better estimation performance than the MMUSIC algorithm.4. In the case of ideal circumstance, the covariance matrix of received data is Toeplitz matrix. However, it is not so in the case of real circumstance. Then, the performance of DOA estimation is deteriorated. In the thesis, a new approach is proposed through establishment of a new Toeplitz matrix based on the correlation of the incept data of array. The new approach, without reducing the aperture of array, increases the ability of discernment under the same SNR, the simulation results show that the proposed method is effective.