MUSIC Algorithm And Direction Of Arrival Estimation

Direction of arrival (DOA) estimation is an important research direction in array signal processing field, which is a common problem in many fields, such as radar, sonar, active protection system, communication system, smart antenna and so on. The DOA estimation algorithm based on array signal processing can resolve the DOAs of multiple signal sources from various directions simultaneously with high resolution. The main research point of spatial spectrum estimation is the DOA processing of spatial signal. The MUSIC (Multiple Signal Classification) algorithm, presented by Schmidt in 1979, is the most classical super-resolution algorithm. Based on an exact model of antenna cirray, the MUSIC algorithm can estimate the DOA of uncorrelated signals accurately However, as far as adjacent signals with small SNR (Signal to Noise Ratio) is considered, MUSIC algorithm cannot estimate their DOA. The main approach to solve this problem is full utilization of signal sources' information. The MUSIC algorithms based on the cyclostationarity property have a better performance in noise suppressing and signal selectivity by using the cyclosationarity property of signal into array signal processing. However as far as signal with cyclic correlation is considered, the signal resolution of this algorithm is limited. To solve this problem, an improved cyclic MUSIC algorithm was presented in this paper. Both cyclic correlation matrices and conjugate cyclic correlationmatrices are considered in this improved algorithm. Emulation results have indicated the improved algorithm can estimate the DOAs of cyclic correlation signals and adjacent signals and has a better resolution and noise suppressing ability.In this paper, the principles of smart antenna and the application of DOA estimation in smart antenna are firstly introduced, and then the MUSIC algorithm is discussed in detail as a classical super-resolution algorithm in DOA estimation in smart antenna. An improved MUSIC algorithm based on the cyclostationary property of signal is presented to solve the shortcomings of the MUSIC algorithm.