Methods For Direction-of-arrival Estimation Of The Wideband OFDM Signals

Orthogonal Frequency Division Multiplexing(OFDM) technology which has high spectrum efficiency and can resistant the frequency selective fading has been concerned by researchers who study on array signal processing. Therefore the ability to effectively estimate the direction-of-arrival(DOA) of wideband OFDM signals has become a new research in the field of array signal processing. Since OFDM signals are sensitive to carrier frequency offset, the article mainly focuses on the DOA estimation of wideband OFDM signals with or without carrier frequency offset.1. This paper firstly studies in detail the principles and defects of the current traditional algorithms which are used to estimate the DOA of wideband OFDM signals. Since incoherent Signal-Subspace Method(ISM) only uses the part of data information of the wideband signals, the estimated performance is not high and at the same time, it can’t distinguish coherent sources because of using the idea of MUSIC algorithm. Although the performance of modified ISM algorithm is better than ISM algorithm and can estimate the DOA of coherent signals under certain condition by de-correlation operation of each sub-narrowband, the performance of the modified ISM algorithm is to be improved due to using part of data information of the wideband signals too. Though coherent signal subspace method(CSM) can distinguish the DOA of coherent signals and its performance is superior to the ISM algorithm and the modified ISM algorithm, there exists errors between steering vector matrix at reference frequency and the one after focusing operation, therefore the performance is still to be improved.2. When the noise is colored noise rather than white noise, CSM algorithm should be whitened in advance, otherwise it will lead to larger estimation error for signals. To solve this problem, a new method called the fourth-order cumulants method is derived, this proposed algorithm can resistant the impact of colored noise by using the features of fourth-order cumulants and the estimated performance is improved, at the same time, sparse reconstruction is used to solve the problem of DOA estimation of coherent signals.3. This paper studies in detail how carrier frequency offset effects OFDM symbols and DOA estimation. Theoretical analysis and simulation results show that carrier frequency offset leads to amplitude attenuation of OFDM received signals while the amplitude of noise is not affected, therefore the SNR decreases, at eventual the performance of the estimation of DOA become worse. To solve this problem, a method whose idea is to compensate carrier frequency offset whose value is estimated by a carrier frequency offset estimation technique is introduced. The simulation results show that carrier frequency offset leads root mean square error of DOA estimation to become higher, however, the root mean square error after compensating carrier frequency offset become lower than the case of existing carrier frequency offset. So the effectiveness of the proposed method is proved.