Research And Application Of Adaptive Filter Algorithm And Its Fpga Implementation

Adaptive filtering is the key technology of adaptive beamformer,which is a central part of smart antenna signal processing system, thestability and high precision of algorithm and running speed of hardwareare of the main standards to evaluate the performances of the system.Firstly, the most common transversal least mean square (LMS)algorithm in the fields of engineering is chose as research object. Ascheme that the optimized parameters of adaptive filter can be found outis advanced with the equivalent relationships between output signals andthe noise signals in the main channel. A series of performance charactersof the noise cancel filter such as the convergence speed, the stability ofweight coefficient, the ability for tracking input signals and theimprovement of signal-noise-ratio (SNR) and so on are analyzed instationary and non-stationary noise circumstance.Based on the analysis of gradient adaptive lattice (GAL) algorithm, aimproved algorithm that the fixed step parameters of GAL algorithm isadvanced according to the astringency and stability of reflectance, thegradient adaptive lattice joint processing algorithm combining new GALand transversal LMS algorithm is studied, all the performance index aresuperior to the transversal LMS algorithm in the same circumstance, itneeds less hardware resource further more.The circuit models of transversal LMS filter and GALJP filter areestablished, the pipelining optimization with the technology of delayleading transfer is also utilized for the two filters. Then the filters aredesigned out on EP2C5T144C6 of CycloneⅡseries of Altera usingmulti-EDA tools.An adaptive beamformer with the 3 section GALJP implemented byFPGA is designed for TD-SCDMA system. The analysis of algorithmshows that it can work better for the downlink beamforming with thereference signal offering by the system.