A note on the convergence analysis of LMS adaptive filters with Gaussian data
نویسندگان
چکیده
1087 puts of each module in a special way to cause the output of the total algorithm to be in proper order 111, 191. This results in the fxtest algorithm, but the program must be written for a specific length. A fourth method is similar, but achieves the unscrambling by properly choosing the mulitplier constants in the modules [IO]. A program based on this method could transform many different lengths by using a stored table of appropriate multipliers. The fifth method uses a separate indexing method for the input and output of each module [ I ] , [ I l l. This results in a slightly slower program than methods three or four, but it is compact and can transform all lengths possible from the combination of modules. ALGORITHMS FOR N = 2 " ' A Cooley-Tukey radix-2 FFT that is calculated in place produces the output in bit-reversed order. The unscrambler usually uses a bit-reverse counter [I]. 1121 or nested loops 131. A particularly efficient form has been programmed by Morris [131. A fast unscrambler has been developed by Evans [ 14). but it requires a precalculated " seed " table. The normal in-place Cooley-Tukey FFT does not allow the un-scrambling to be done in the modules (butterflies) as is possible with the PFA. That is an intrinsic property of the type of index map used and the fact that the module lengths are not relatively prime. It is possible to design an in-place, in-order radix-2 FFT if the butterflies are calculated two at a time and a special indexing used. This is described in [15], [I61 and a structure is shown in The important observation contained in this correspondence is that a radix-4 or any radix-2 " ' FFT can be modified so the output is in bit-reversed order. If a normal radix-4 FFT with the output of each butterfly in normal order is used, the output occurs in base-4 reversed order, and similarly for radix-8 and others. However, if for the radix-4 FFT, the short length-4 butterflies are modified to have their outputs in bit-reversed order, the output of the total ra-dix-4 FFT will be in bit-reversed order, not base-4 reversed order. Likewise, if the output of the length-8 butterflies in a radix-8 FFT are placed in bit-reversed order, the output of the total radix-8 FFT will be in bit-reversed order. This allows converting the output …
منابع مشابه
Utilizing Kernel Adaptive Filters for Speech Enhancement within the ALE Framework
Performance of the linear models, widely used within the framework of adaptive line enhancement (ALE), deteriorates dramatically in the presence of non-Gaussian noises. On the other hand, adaptive implementation of nonlinear models, e.g. the Volterra filters, suffers from the severe problems of large number of parameters and slow convergence. Nonetheless, kernel methods are emerging solutions t...
متن کاملSpeech Enhancement by Modified Convex Combination of Fractional Adaptive Filtering
This paper presents new adaptive filtering techniques used in speech enhancement system. Adaptive filtering schemes are subjected to different trade-offs regarding their steady-state misadjustment, speed of convergence, and tracking performance. Fractional Least-Mean-Square (FLMS) is a new adaptive algorithm which has better performance than the conventional LMS algorithm. Normalization of LMS ...
متن کاملOptimization Capabilities of LMS and SMI Algorithm for Smart Antenna Systems (RESEARCH NOTE)
In the present paper convergence characteristics of Sample matrix Inversion (SMI) and Least Mean Square (LMS) adaptive beam-forming algorithms (ABFA) are compared for a Smart Antenna System (SAS) in a multipath environment. SAS are employed at base stations for radiating narrow beams at the desired mobile users. The ABFA are incorporated in the digital signal processors for adjusting the weight...
متن کاملConvergence analysis of LMS filters with uncorrelated Gaussian data
Statistical analysis of the least mean-squares (LMS) adaptive algorithm with uncorrelated Gaussian datais presented. Exact analytical expressions for the steady-state mean-square error (mse) and the performance degradation due to weight vector misadjustment are derived. Necessary and sufficient conditions for the convergence of the algorithm to the optimal (Wiener) solution within a finite vari...
متن کاملThe Wavelet Transform-Domain LMS Adaptive Filter Algorithm with Variable Step-Size
The wavelet transform-domain least-mean square (WTDLMS) algorithm uses the self-orthogonalizing technique to improve the convergence performance of LMS. In WTDLMS algorithm, the trade-off between the steady-state error and the convergence rate is obtained by the fixed step-size. In this paper, the WTDLMS adaptive algorithm with variable step-size (VSS) is established. The step-size in each subf...
متن کاملA Family of Selective Partial Update Affine Projection Adaptive Filtering Algorithms
In this paper we present a general formalism for the establishment of the family of selective partial update affine projection algorithms (SPU-APA). The SPU-APA, the SPU regularized APA (SPU-R-APA), the SPU partial rank algorithm (SPU-PRA), the SPU binormalized data reusing least mean squares (SPU-BNDR-LMS), and the SPU normalized LMS with orthogonal correction factors (SPU-NLMS-OCF) algorithms...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید
ثبت ناماگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید
ورودعنوان ژورنال:
- IEEE Trans. Acoustics, Speech, and Signal Processing
دوره 36 شماره
صفحات -
تاریخ انتشار 1988