Research On DOA Estimation Technology Based On FPGA

The Direction of Arrival(DOA)estimation of the signal refers to processing and calculating the spatial signal of interest and estimating the arrival direction of the signal received.However,the super-resolution direction-finding algorithm requires a large amount of computation and complicated hardware implementation,with a result that real-time performance is difficult to meet actual needs.Meanwhile,the accurate source number estimation is the premise of many spatial spectrum estimation algorithms,and the performance of the algorithm is seriously affected when the estimation of the number of radiation sources is in error.Therefore,it is of great significance to study the direction of arrival estimation under the condition of the number of unknown radiation sources,reduce the computational complexity of the algorithm,and implement the hardware in real time.The direction of arrival estimation of unknown source number,fast subspace directionfinding algorithm and FPGA implementation of MUSIC algorithm are the focus of this paper.The work of this article is as follows:(1)This paper studies the MUSIC algorithm and ROOT-MUSIC algorithm,analyzes the orientation vectors of uniform linear arrays,circular arrays and L-shaped array models,simulates and compares the performance of different source estimation algorithms.It also studies two solutions coherent algorithm,spatial smoothing algorithm and Toeplitz algorithm.(2)Two improved propagation operator estimation algorithms are proposed.One is to perform real-time processing on the propagator operator so that the complex operation of the algorithm is transformed into a real operation,and forward-backward spatial smoothing is introduced into the algorithm.Finally,the direction of the radiation source is estimated using the root-finding method.The improved algorithm further reduces the computational complexity of the algorithm and can be applied in the multipath propagation environment.The other one is to preprocess the received signal by wavelet denoising.The Toeplitz algorithm for low signal-to-noise ratio transforms the reconstructed matrix after wavelet denoising.Finally,it also uses the root method to estimate the direction of signal.This algorithm improves the output signal-to-noise ratio and the performance of propagation under low-SNR condition in multipath propagation environments.(3)This paper studies the direction of arrival estimation under the condition that thenumber of radiation sources is unknown.In this paper,a direction-finding algorithm suitable for coherent signals is proposed on the basis of MUSIC-like algorithm in order to solve the problem that MUSIC-like algorithm has large computational complexity and does not have the ability of decoherence and that pseudo peak is easy to appear.The idea of the algorithm is derived from the MUSIC algorithm and the beamforming algorithm.The constraint vector is used to obtain the optimal weight vector,so that the weight vector falls into the noise subspace and the spatial spectrum is constructed between the weight vector and the steering vector by the orthogonal relationship.The real feature decomposition and root instead of peak search can reduce the computational complexity of the algorithm.The improved algorithm does not need to know the number of radiation sources in advance,and under the condition of low SNR or few snapshots,it can still maintain a certain direction-finding performance.It also has decoherence capability,reduces the probability of a pseudo peak,and is able to accurately estimate the direction of the radiation source in a multipath propagation environment.(4)Finally,the hardware implementation of the MUSIC algorithm is proposed.In the hardware implementation based on the ZYNQ development platform,the PL side implements the configuration management of MUSIC algorithm and the bottom module.The PS side performs data statistical analysis and system-level communication processing functions.Meanwhile,the combination of the FPGA and the ARM expands the application range of the super-resolution direction-finding algorithm.Three modules of covariance matrix,feature decomposition and peak search of MUSIC algorithm are designed through FPGA.At last,the simulation output of Vivado software platform is given,which verifies the feasibility of the design scheme.