Research And Improvement On DOA Estimation Algorithm In Smart Antenna

In these years, with the fast increase of the mobile users and the appearance of new type of wireless services, the wireless services become lager and lager. This results in too much serious Co-Channel Interruption and the Multiple Address Interruption. The band of the transported signals and the signal rate become restricted. At the same time, the transport environment of the wireless signals is complicated and uncertain; all this has strong bad effect on the capacity and performance of wireless communication system. The smart antenna technology is an effective solution to take good use of the frequency sources and to promote the capacity and the service quality of the communication system. Smart Antenna has two critical tasks, one is to filter the uplink signals, and the other is to form the downlink beam, and we need adaptive algorithm and digital signal processing(DSP) technology to fulfill these work .Adaptive algorithm is one of the most important technologies of Smart Antenna, and it determines Smart antenna's speed to the wanted to signal and the complexity of the circuit of the communication system. Consequently, the system should select reasonable algorithms in various communication environments and algorithm diversity technology could optimize the system to work well.Smart antenna is an antenna with the capability of spatial and time handling, which combines the array antenna and the advanced signal processing technology and which is very important Wideband Adaptive Array in a signal processing domain. The smart antenna system has three kind of main structures: Switched Beam Systems,Adaptive Array Structure,Wideband Adaptive Array. Switched Beam Systems is low in the price of the project, but its capability of dealing with environment change is bad. The self-adaptive system can effectively make up this kind of insufficiency, and its application is more widespread. The principle of work of the smart antenna technology chooses the appropriate directional diagram through the change power value, namely the antenna pattern, does its best enlarging the useful signal, simultaneously suppresses the unwanted signal the directional diagram, aims at making the main petal in the directional diagram and the direction of the useful signal consistent, but lowest gain (called falls zero) in directional diagram aims at the unwanted signal direction in other words. But it happens only in the ideal situation. In fact because the wireless correspondence environment is complex, for example the interfering signal is in abundance, multiple access interference exists, and the number of the antenna array is finite, makes the ideal plan not be possible to realize. Therefore, whether the directional diagram is appropriate may use the principle of Maximum Signal-to-Noise Ratio to weight.A very important content of smart antenna studies the estimate of the spatial spectrum the basic structure of the estimate of the spatial spectrum is divided into two parts: one is pretreatment algorithm; the other is the DOA estimate algorithm. The main goal of the pretreatment algorithm is data processing, including the hypothesized transformation, removing correlation and enhancing signal and so on, but the DOA estimate algorithm estimate the direction of concrete signal. This article mainly does the key research on the smart antenna in DOA estimate algorithm.DOA estimate algorithm is divided into four kinds: conventional technique, subspace based technique, maximum likelihood technique and integrated technique. The conventional technique is based on the classic method of beam form, needs a great deal of array elements to obtain the high resolution. The subspace based technique use the characteristic structure data-in array, which is the most superior method of obtaining the high resolution. Maximum likelihood technique is the most superior plan, even if the signal-to- noise ratio is very low in the environment, it also can obtain the very good performance, but the computation quantity is usually very big. The integrated technique is a very hopeful method to CDMA, uses the plan of restoring the characteristic to distinguish many signals, and estimate the spatial characteristic, and then uses subspace based technique ensure the direction of signal's arrival.The MUSIC algorithm is a kind of algorithm used extremely widespread, which uses the orthogonal characteristic of signal subspace and noise subspace, structures space spectrum function, examines the signal DOA through the spectrum peak searching,. Although in stationary random process the average value of the signal is zero, the signal is mutually independent with the noise, the noise of a processor is additive white Gaussian noise, and the number of signals is less than the number of the antenna arrays, the MUSIC algorithm in theory may have willful resolution. But the actual situation is extremely complex, and cannot have such ideal condition. Through the simulation there are some conclusions: resolving power of the incident signal DOA and the estimate precision under using the MUSIC algorithm become better and better, with the number of the antenna array,the signal-to-noise ratio and the system sampling points increasing, while when he number of the antenna array,the signal-to-noise ratio and the system sampling points decrease, the characteristic of the MUSIC algorithm can drop, even expires. In addition, when signal is coherent or correlation signal, because the covariance matrix R_ss of source turns strange, the MUSIC algorithm will expire.In order to solve these problems, we proposed the Root-MUSIC algorithm; its basic thought is that the method of multinomial root replaces the MUSIC algorithm the spectrum peak search. Through the analysis of theory and the simulation there are some conclusions obtained: Root-MUSIC algorithm can reduce its computation quantity, moreover there is the enhancement in the resolution, especially in the low signal-to-noise ratio situation. But it is emphasized that Root-MUSIC algorithm only adapts in uniform line, and this point and the MUSIC method is different.In Root-MUSIC algorithm, when the signal is coherent, Root-MUSIC algorithm expires. But through the characteristic of Toeplitz, a plan of amending Root-MUSIC algorithm may be proposed to make the covariance matrix become Toeplitz matrix and make it hold the capability of distinguishing coherent signal. Through the theoretical analysis and the experimental simulation there are some conclusions obtained: the revision Root-MUSIC algorithm may effectively resolve coherent signal, its whole resolution have the enhancement compared with the Root-MUSIC algorithm, two signals whose angle of arrival is very close still are able to be distinguished.Computer simulation is carried out to validate the above feasibilities by the research tool of the software MATLAB.