Research On The Doa Estimation Method Of Mixed Source Based On Uniform Circular Array

Array signal processing has always been a major concern, it has been used in many fields, and the DOA estimation is an important component. Due to the complexity of the physical environment, actually, signals are often not exist alone. This paper studies on two common kinds of mixed source estimation methods: the first is one and two dimensional estimation methods of mixed coherent and incoherent signals; the second is two-dimensional estimation methods of mixed circular and noncircular signals. Due to the advantages of uniform circular array, this paper methods are based on this array model. Furthermore, the estimation methods of mixed source under the two circumstances has all combined with low computational methods, making it more practical.The specific work of this paper is as follows:Firstly, introduced two widely used array model: ULA and UCA. Then describled definitions and properties of coherent signal, incoherent signal, circular signal and noncircular signal. And expounds the two kinds of noise environment used in this paper: Gauss and impulsive noise.Secondly, studied the one and two dimensional estimation methods of mixed coherent and incoherent signals under UCA. For one-dimensional estimation, added impulsive noise environment, first use fractional lower order moments to deal with the data that affected by impulsive noise, then combine the principle of mode conversion, construct difference matrix to obtain the DOAs of signal. This method not only eliminates the influence of impulse noise, but also can separate independent and coherent signals, but it can only estimate one-dimensional angle. For the case of two-dimensional estimation, studied on the decorrelation method of double circular array smoothing. The method use the way of calculating instead of spectral peak searching to obtain the two dimensional angle information of signals, making the amount of calculation reduced to a large extent.Finally, for the estimation method research of mixed circular and noncircular signals, there has exists one-dimensional DOA estimation methods under ULA, but few studied two-dimensional estimation methods based on this kind of mixed signal. This paper first combined non-circular feature and UCA to obtain two-dimensional angle information of mixed sources. This method has accurate estimation results, but large amount of calculation. Based on this, this paper used the principle of UEPM(unitary extended propagator method) algorithm under this condition, and then proposed an improved method which is of low complexity. The method uses real arithmetic instead of complex operation. And no longer use eigen decomposition of covariance matrix but the methods of propagator to obtain the noise subspace, Then use it to get the DOAs of signal. Which is of low complexity and high practicability. Simulation results show that the algorithm can separate the two signals, estimate more signals than the number of array elements, and the estimation performance is also a great advantage.