Araay Signals' DOA Tracking And Beamforming

The nuclear content in smart antennas' field is estimation of direction of arrival (DOA) and beamforming, and there are series of algorithms already. In this dissertation, comprehensive study on how to track those time-varing signals' DOA and how to do beamforming is presented. Much work has been done with subspace adaptive algorithms and neural networks.Firstly, we analyse and simulate some exist algorithms, and then concentrate on subspace adaptive algorithms and neural networks algorithms.Those beamforming algorithms based on eigenspace can work well when the steering vector is fixed or varies little with time. Yet in real mobile communication circumstance, the vector is always changing for user equipment's moving or other reasons. Based on ESPRIT (Estimating Signal Parameter via Rotational Invariance Techniques), an improved beamforming algorithm is presented via subspace tracking with constrained Perturbations application accordingly. For the character of updating with time, this algorithm can restrict the steering error. Simulation results show its effectiveness.Most algorithms about smart antennas are based on mathematical models. For often using matrix decomposing and other methods, much computation is needed. Besides, the realtime processing ability and adaptive ability are poor. Recently, neural network methods are considered in these fields, which have advantages including reduction in computation, fast converging speed, quality of being precise and adaptive ability. Based on radial basis function neural network (RBFNN), certain improved algorithm for DOA estimation is proposed. Simulation results show that after decorrelation, less samples are needed to track the direction-varying signals, which makes it easier for engineering application.Based on RBFNN, certain improved beamforming algorithms is also considered. Because of the character of Toeplitz matrix, only elements of the first row are needed to simplify the structure of neural network. Simulation results indicate that after simplification, its robust is strengthened and beamforming performance is improved.