Research And Realization On Adaptive Parameter Estimation And Signal Synthesis Algorithms In Multi-Antenna Arraying

Utilizing multi-antenna technique to receive signals is one of the most important approach for wireless communication at low Signal-to-Noise Ratio. This paper makes a research on the key technologies of randomly-distributed multi-antenna array in the Gaussian channel, which contains the estimate of parameter differences and combining weights. It tries to use adaptive algorithms to solve the problem of parameter estimate made by time varying. At the end, based on FPGA and DSP, it completes the work of algorithms designing and realizing for four-channel multi-antenna signal synthesizer. The details are as follows:1. At first, this paper illuminates the signal processing structure of randomly-distributed multi-antenna array. Based on the signal model in Gaussian channel, it presents an approach of"pre-processing + precise estimation"to deal with the problem of parameter differences'estimation and compensation. (1) For pre-processing, it uses the method of joint estimation which is based on ambiguous function to deal with the problem when time delay difference and frequency offset exist simultaneously, and presents a joint estimation algorithm based on FFT. It has a good performance in project realization. (2) This paper also makes a research on precise estimation of frequency offset. It analyzes the performance of the frequency estimation algorithm which is based on single frequency signal at low SNR, and presents an algorithm based on the phase of correlation function, and modify it by adaptive filtering. It has some tracking ability for time varying frequency offset even at low SNR. (3) For precise estimation of time delay difference, this paper uses ETDE algorithm to track the residual delay difference, and presents an adaptive estimation structure based on LS optimization. It has a good performance on the tracking of time-varying delay at low SNR.2. This paper also makes a research on signal synthesis algorithms. Base on the research of some common synthesis algorithms, it presents a novel weight estimation algorithm which utilizes three signals. The algorithm is based on SNR rule, get the SNR and the raio of amplitude attenuating factor from the correlation of every two signals. It has a good performance for both uniform and non-uniform array. This paper also gives an algorithm which is based on iteration of correlation matrix. It assumes that signals have the same power of noise after normaliztion at low SNR. It makes the estimation of phase differences and the raio of amplitude attenuating factor from the iteration of correlation matrix, and then, get the optimal weights. It has a good performance at low SNR, could make up the ddeficiency of the former one whose performance is not quite well at low SNR.3. Based on the platform of FPGA+DSP, it designs the signal processing flow for four-channel synthesizer, and completes the procedure development of the DSP. In the end, it tests the synthesizer by self-made signals and satellite signals. The result shows that the synthesizer could modify the quality of the signal and cut down the BER effectively.