Research And Implementation On Weight Estimation Algorithm In Multi-Antenna Signal Combining Techniques

Due to the explosive growth of the wireless communication and increasing demand for the wireless telecommunications services, signals become more intensive, and the received signals become weaker in some application environments. As a complement for the traditional receiving means, multi-antenna signal combining techniques (MASCT) has captured the attention of many researchers. MASCT can effectively improve the received signal, consequently, the weight estimation algorithm for MASCT has become a hotspot.This dissertation studies the key technologies to the MASCT thoroughly, and focuses on the weight estimation algorithm in different composition schemes, and designs a multi-antenna signal combining software. The contributions obtained in this thesis can be summarized as follows.1. Measures to solve the weight estimation algorithm in symbol weight oriented, a multi-antenna joint weight amplitude and phase estimation algorithm based on the EM algorithm is proposed. Based on the EM algorithm framework, the new algorithm introduces the unknown transmitted data as the "missing data" to construct "complete data" model, and the maximum likelihood estimation of the amplitude and phase is derived from the conditional a posteriori expectation of the complete data likelihood function, which is calculated iteratively. The simulation results show that the proposed algorithm is highly precise and is applicable to different modulation modes. And both the amplitude and phase estimation can achieve the Cramer-Rao low bound(CRLB) in the case of high signal-to noise ratio(SNR). The SNR loss of the proposed algorithm can be ignored.2. Based on further study of SUMPLE algorithm, an improved algorithm is proposed. Firstly, choose a signal and fix the weight phase of the signal. Then the signal of each antenna, except the fixed phase one, is correlated with the weighted sum of all the antenna signals left to obtain its weight phase. Comprare to SUMPLE algorithm, the new algorithm has a lower computation and a slightly covergenence rate. Simulation results show that the phase wandering problem of the SUMPLE algorithm is greatly eliminated without deteriorating the performance.3. To solve the problem of wave oriented weight estimation in multi-antenna system, a joint multi-antenna weight amplitude algorithm based on cross correlation matrix is proposed, which groups the cross correlation matrix with three antennas and obtains multiple estimations of one antenna, after then the weight amplitude can be estimated by averaging those estimation values. The last weight amplitude estimation is the best estimation based on least-square algorithm. The simulation results show that the proposed algorithm operates well and is suitable for multi-antenna combining system in low SNR environments.4. The multi-antenna signal combining software is designed. The structure of the software is divided into four modules, which are the interface control and display module, the data reading and writing module, the data processing module and the signal combining module. The principle and realization of eath module has been introduced respectively. Visual C++ 2008 is used as the development tool. In order to accelerate the software, three major mixed programming mode of the VC++ are compared, namely, MATLAB mixed programming mode, IPP programming mode and CUDA programming mode. And the mixed Programming of IPP and VC++ is chosen in this dissertation. The results of simulation and testing indicate that the performance of the software is good.