HomeChair: John R. Treichler, Applied Signal Technology, USA
Home
Wonzoo Chung, Cornell University (U.S.A.)
James P LeBlanc, New Mexico State University (U.S.A.)
We study the local minima relocation of the fractionally spaced Constant Modulus Algorithm(FSE-CMA) cost function in the presence of noise. Local minima move in a particular direction as the noise power increases and their number may be eventually reduced. In such cases the performance of FSE-CMA may fail to adequately reduce inter symbol interference (ISI), but achieve an approximated MMSE by reducing its equalizer noise gain under certain constraints. We analyze the mechanism of relocation of FSE-CMA cost function local minima in terms of the auto-correlation matrix of sub-channel convolution matrix and its eigenvectors.
ic982298.pdf (From Postscript)
TOP
Marie-Line Alberi, ETIS (France)
Inbar Fijalkow, ETIS (France)
James D Behm, U.S. Department of Defense (U.S.A.)
Thomas J Endres, Sarnoff Digital Comm (U.S.A.)
The improved convergence speed and tracking properties of fractionally-spaced equalizers are analyzed. We consider in particular the effect of a frequency offset between the transmitter baud rate and the receiver sampling clock that induces important time-variations. We show that a fractionally-spaced equalizer can handle the intersymbol interferences (ISI) induced when the propagetion channel doesn't introduce too much ISI.
ic982458.pdf (From Postscript) TOP
Philip Schniter, Cornell University (U.S.A.)
C. Richard Johnson Jr, Cornell University (U.S.A.)
Adaptive blind equalization has gained widespread use in communication systems that operate without training signals. In particular, the Constant Modulus Algorithm (CMA) has become a favorite of practitioners due to its LMS-like complexity and desirable robustness properties. The desire for further reduction in computational complexity has motivated signed-error versions of CMA, which have been found to lack the robustness properties of CMA. This paper presents a simple modification of signed error CMA, based on the judicious use of dither, that results in an algorithm with robustness properties closely resembling those of CMA. An approximation to the steady-state mean-squared error performance of the new algorithm is derived for comparison to that of CMA.
ic981509.pdf (From Postscript) TOP
Christian Luetkemeyer, University of Technology RWTH Aachen (Germany)
Hans-Martin Bluethgen, University of Technology RWTH Aachen (Germany)
Tobias G Noll, University of Technology RWTH Aachen (Germany)
A transversal equalizer with half Baud spaced taps in the center and extended with Baud spaced taps on both sides is presented. This hybrid equalizer combines the benefits of Baud spaced equalizers - like superior equalization of notches in the middle of the transmission band - and fractionally spaced equalizers, which have a superior performance when equalizing asymmetric notches in the slope of the transmission band, when the same number of coefficients are used. The hybrid equalizer offers the reduced sensitivity to sampling time changes and the ability to model the matched filter in the receiver as the fractionally spaced equalizer. The problem of tap-wandering, which is present in fractionally spaced equalizers, is reduced due to the reduced degree of freedom in the coefficient adjustment.
ic981966.pdf (From Postscript) TOP
Duncan J Brooks, Imperial College (U.K.)
Sangarapillai Lambotharan, Imperial College (U.K.)
Jonathon A Chambers, Imperial College (U.K.)
The performance of the Constant Modulus Algorithm can suffer because of the existence of local minima with large Mean Square Error(MSE). This paper presents a new way of obtaining the optimum MSE over all delays using a second equalizer under a mixed Constant Modulus and Cross Correlation Algorithm(CM-CCA). Proof of convergence is obtained for the noiseless case. Simulations demonstrate the potential of the method.
ic981774.pdf (From Postscript) TOPMonisha Ghosh, Philips Research (U.S.A.)
The two criteria most commonly used in blind equalization are Sato's cost function and Godard's cost function. In this paper we analyze a sign-error cost function for real signals which gives an error term that can be viewed as the sign of either the Sato or the Godard error. We show that the conventional definition of equalizer convergence is not suitable for analyzing this cost function. A more realistic definition of convergence for low to medium SNR situations is presented and used to analyze this sign-error cost function. The performance of this cost function is evaluated via simulations and shown to have excellent performance as compared to the Godard cost function, with substantially less complexity.
ic981500.pdf (From Postscript) TOP
Chong-Meng See, DSO National Laboratories (Singapore)
Colin F.N. Cowan, The Queens University of Belfast (Northern Ireland)
In this paper, we present an adaptive algorithm for space-time channel identification and equalization. The proposed algorithm performs joint channel estimation and sequence detection by optimizing a least squares cost function iteratively in a forward and backward manner. Simulation results demonstrate the proposed algorithm to be data efficient and fast converging. In addition, good BER performance is achieved in time-varying channels at relatively low SNR and with an extremely short start-up sequence. These attributes render it suitable for wireless mobile communications using short burst data format.
ic981547.pdf (From Postscript) TOP
Palaniswamy Chandra Kumar, Nanyang Technological University (Singapore)
Paramasivan Saratchandran, Nanyang Technological University (Singapore)
Narasimhan Sundararajan, Nanyang Technological University (Singapore)
This paper presents the study results of non-linear channel equalisation problems in data communications using a recently developed minimal radial basis function neural network structure, referred to as MRAN(Minimal Resource Allocation Network). MRAN algorithm uses on-line learning and has the capability to grow and prune the RBF network's hidden neurons ensuring a parsimonious network structure. Compared to earlier methods, the proposed scheme does not have to estimate the channel order first, and fix the model parameters. Results showing the superior performance of the MRAN algorithm for two different non-linear channel equalisation problems, along with a linear non-minimum phase problem, are presented.
ic981151.pdf (From Postscript) TOP
Steven Bouchired, ENSEEIHT-SIC (France)
Mohamed Ibnkahla, ENSEEIHT-SIC (France)
Daniel Roviras, ENSEEIHT-SIC (France)
Francis Castanié, ENSEEIHT-SIC (France)
The paper proposes a neural network approach to equalize time varying nonlinear channels. The approach is applied to a satellite UMTS channel composed of time invariant linear filters, a non-linear memoryless amplifier and a time varying multipath propagation channel. The neural network equalizer has a Radial Basis Function structure. The usual k-mean clustering algorithm is replaced by a Kohonen learning rule. This results in an RBF-SOM equalizer which outperforms the LMS equalizer, and which has better recovering abilities (after passing through a high fading area) than the former RBF equalizer.
ic982013.pdf (From Postscript) TOPArnaud Bouttier, Thomson-CSF (France)
This paper describes a new adaptive blind equalization algorithm based on a truly IIR structure that enables the correction of ISI over severely distorted channels. The recursive feedback filter is in lattice form to allow an easy monitoring of the filter stability. During blind training, the adaptation of the equalizer is carried out via the usual stochastic gradient algorithm by minimizing the Shtrom-Fan cost function, a CMA like functional robust to ill-convergence. Once in steady state, the algorithm switches automatically into a classical DFE structure adapted via the DD-MMSE criterion. Simulation results show that this new equalizer outperforms most of the traditional blind FIR equalizers.
ic982557.pdf (From Postscript) TOP