| Direction of arrival (DOA) estimation technology, which is an important aspect in arraysignal processing field, can be used to estimate the direction angle of received signals byextracting the information in signals. In this dissertation, two problems of the DOA estimationin the array antenna are researched deeply, which are mutual coupling calibration and methodof DOA for coherent signals.1. A self-calibration algorithm for an L-shaped uniform array (SAL) in the presence ofmutual coupling is presented in the paper. Cascade estimation instead of joint estimation isused to reduce the computation. The precision of DOA performs better in low SNRenvironment.Based on the structure of L-shaped uniform array, the algorithm regards theorigin of coordinates as a virtual reference, and this point does not exist a physical sensor. Toreconstruct and decompose the mutual coupling matrix, a new form of decoupling matrix isderivate. With the new decoupling matrix and MUSIC algorithm, the two-dimensional DOAof the signals and mutual coupling coefficient of the L-shaped uniform array will be estimated.The simulation results show that when the snapshot was100and the Signal to Noise Ratio(SNR) was less than10dB after two iterations, the precision of DOA estimation increased by18.13%on average, and the correction precision of mutual coupling coefficients increased by24.15%. When the SNR was0dB and the snapshot was less than400after two iterations, theprecision of DOA estimation increased by0.32%,the correction precision of mutual couplingcoefficients increased by21.35%.2. An improved spatial differencing method for DOA estimation is proposed byconstructing a spatial differencing matrix with neighboring subarrays under the coexistence ofboth uncorrelated and coherent signals, as well as the condition of uniform linear arrays(ULAs). The color noise in the new matrix is repressed with the property that the color noiseis correlative, so the proposed method performs better in an spatial color noise environment.Then the expression of the maximum detectable signals versus sensors and subarrays isderived. The relationship between DOA estimation precision and subarrays is discussed.Simulations show that the estimation accuracy is improved under low SNR environment. Thenumber of maximum detectable signals can be more than the number of sensors, varying withthe number of subarrays. The precision of DOA can be improved by increasing the number ofsubarrays, however, the maximum detectable signals will decrease when the number ofsubarrrays is out of a threshold, so there is a balance between precision and maximumdetectable signals. |
