Research On DOA Estimate Technologies And Related Beamforming Technologies In Smart Antenna

Smart antenna is an important research field of modern signal processing. It has found wide applications in many fields, such as radar, communications, sonar, seismology, radio astronomy and biomedicine engineering, etc. Firstly, the model and theory of Smart antenna are introduced in this dissertation. Secondly, the property and weaknesses of the former various DOA (Direction Of Arrival) estimate algorithms, such as Capon, MUSIC (Multiple Signals Classification) and ESPRIT (Estimation of Signal Parameters via Rotational Invariance Techniques) and so on are introduced. Introduction for those of the related beamforming technologies is followed. And then new DOA algorithms and related beamforming technologies are discussed in detail.It's found that the traditional algorithms, such as Capon, MUSIC and ESPRIT and so on, are not prefer to estimate DOA quickly and timely because of relative larger computation load and they perform poorly if SNR (Signal to Noise Ratio) is relatively low and signals are rich through studying on the algorithms for DOA estimate and related beamforming technologies based on the second moment. A novel algorithm, NCFPMA (the Non-Circular Fast Projection Matrix Algorithm) is proposed to smooth away the problem. NCFPMA can estimate DOA of more signals better and more fast than the former, because it expands the array aperture virtually under the assumption that the signals impinging on the array are non-circular, and uses the data of virtual covariance matrix to approximate the signal subspace. The theoretical analysis and computer simulations show the effectiveness of NCFPMA.It's found that the popular algorithms, including MUSIC-LIKE, VMUSIC (Virtual MUSIC) and so on, can only work well under the condition that the number of the signals is known as a priori and when the information about the number of signals is wrong, those algorithms will make poor DOA estimate through studying on the algorithms for DOA estimate and related beamforming technologies based on the fourth order cumulant (FOC). A FOC DOA algorithm based on the noise subspace of fixed dimensions is proposed under the assumption that the signals impinging on the real array are resoluable for the virtual array to smooth away the problem. The new FOC algorithm based on the noise subspace of fixed dimensions can obtain the estimate of a substitute DOA set containing DOA of signals. Then it can find the estimate of true DOA in the substitute DOA set and estimate the number of the signals by certain principle at the same time. The theoretical analysis and computer simulations testify the effectiveness of the new method.True noise environments are not often the Gauss noise and some behave rather impulsive. For this reason, DOA algorithms are surveyed in impulse noise condition and several algorithms, such as ROC-MUSIC( RObust Covariation MUSIC), FLOM-MUSIC(Fraction Lower Order Moment MUSIC) and so on are turned out. But the heavy calculation burdens of these methods prevent them from estimating DOA quickly and efficiently. A fast real-valued FLOM algorithm under impulse noise environment is proposed to smooth away the problem. The fast real-valued algorithm can obtain a real FLOM matrix by recombining the data out of a former FLOM matrix, use the data of real FLOM to approximate the noise subspace and estimate the DOA of signals. The new one can obtain the good estimate for DOA of signals more fast under impulse noise environment, because it does that in real number field without exhaustive eigen-decomposition. The theoretical analysis and computer simulations testify the effectiveness of the fast real-valued algorithm under impulse noise environment.