最小均方差適應性演算法研究及其在IS-54等化器之應用

Abstract

在時變或未知系統中，適應性訊號處理是十分重要的課題。在本文中， 針對減低運算量，我們對於一已提出之新式高效適應性演算法 (the Chen's LMS algorithm) 做進一步的理論分析並尋求更進一步的應用。其 次，針對加快收斂速度上，我們提出一新式外插適應性演算法，在尋求系 統穩定並減小均方誤差的前題下，能大幅提高系統收斂速度。最後，我們 研讀北美數位蜂巢行動電話系統。為對抗由於多重路徑衰減所造成之交越 訊號雜訊，使用一低複雜度之等化器 (MT-MT LMS) 以符合標準之規範。 並且以計算機輔助設計工具(DSP Station) 從事硬體之實踐。 In this thesis, to reduce the computation complexity,we discuss a new effi- cient LMS adaptive filtering algorithm ( the Chen's algorithm ) in more detail about the properties of stability and its variants. An extrapolated LMS adaptive filtering algorithm is also proposed here to speed up the rate of convergence. For stability and small excess mean square error considerations, the ELMS algorithm converges much faster than LMS algorithm, at the cost of minor extra computation. For practice , we study the North American Digital Cellular (NADC, IS-54 standard) system. A low complexity Multiple Training and Multiple Tracking LMS ( MT - MT LMS) is presented here in order to match the recommended minimum performance specification. An implementation is then given with the help of CAD tools (Dsp Station).