Wyniki wyszukiwania - BazTech

1

Blind Estimation of Linear and Nonlinear Sparse Channels

Georgoulakis K.

Journal of Telecommunications and Information Technology

|

2013

|

nr 1

65-71

EN

This paper presents a Clustering Based Blind Channel Estimator for a special case of sparse channels - the zero pad channels. The proposed algorithm uses an unsupervised clustering technique for the estimation of data clusters. Clusters labelling is performed by a Hidden Markov Model of the observation sequence appropriately modified to exploit channel sparsity. The algorithm achieves a substantial complexity reduction compared to the fully evaluated technique. The proposed algorithm is used in conjunction with a Parallel Trellis Viterbi Algorithm for data detection and simulation results show that the overall scheme exhibits the reduced complexity benefits without performance reduction.

2

Pipelined architectures for the frequency domain linear equalizer

Glentis G. O., Georgoulakis K.

International Journal of Applied Mathematics and Computer Science

|

2006

|

Vol. 16, no 4

525-535

EN

In this paper, novel pipelined architectures for the implementation of the frequency domain linear equalizer are presented. The Frequency Domain (FD) LMS algorithm is utilized for the adaptation of equalizer coefficients. The pipelining of the FD LMS linear equalizer is achieved by introducing an amount of time delay into the original adaptive scheme, and following proper delay retiming. Simulation results are presented that illustrate the performance of the effect of the time delay introduced into the adaptation algorithm. The proposed architectures for efficient pipelining of the FD LMS linear equalization algorithm are suitable for implementation on special purpose hardware by means of the ASIC, ASIP or FPGA VLSI processors.

3

Pipelined architectures for the LMS adaptive Volterra filter

Glentis G.-O., Georgoulakis K.

Foundations of Computing and Decision Sciences

|

2006

|

Vol. 31, No. 2

135-155

EN

In this paper, efficient pipelined architectures for Least Mean Square (LMS) adaptive filtering and system identification of discrete-time Volterra models is presented. First, the multichannel embedding is adopted for the transformation of the discrete-time Volterra model to an equivalent multi-input single output format. Then, the LMS algorithm with delayed coefficients adaptation is applied, for the identification of the model parameters. The adaptation delay introduced in the computational flow of the adaptive scheme, allows for a pipelined implementation, however, the convergence and tracking properties of the algorithm are affected. Proper correction terms are subsequently introduced that compensate the adaptation delay and give results identical to the original LMS algorithm, subject to a latency delay.

4

Adaptive least squares algorithms for two dimensional system identification and filtering: a unified approach

Glentis G.-O., Georgoulakis K.

Foundations of Computing and Decision Sciences

|

2003

|

Vol. 28, No. 2

95-112

EN

In this paper, a unified approach is presented for adaptive Least Squares Two-Dimensional system identification and linear filtering. First, a unified deterministic Least Squares criterion is introduced, and subsequently utilized for the derivation of a general algorithmic framework for the adaptive Two-Dimensional system identification and filtering. Overdetermined, as well as underdetermined Least Squares Two-Dimensional adaptive algorithms are derived, from the proposed general adaptive scheme. In this way, known Two-Dimensional adaptive algorithms are interpreted as special cases of a general algorithmic form. Moreover, new adaptive algorithms are derived, following the proposed methodology.

5

Performance analysis of the frequency domain LMS adaptive filter using the sliding DFT

Glentis G.-O., Georgoulakis K.

Foundations of Computing and Decision Sciences

|

2002

|

Vol. 27, No. 1

27-41

EN

In this paper the performance of the Frequency Domain LMS adaptive filter is investigated, for the case when the Sliding Discrete Fourier Transform is utilized for the frequency domain data transformation. A statistical performance analysis in terms of the mean, and the mean squared error of the filter parameters is presented. The convergence speed of the algorithm is analyzed in terms of the eigenvalue spread of the input signal autocorrelation matrix. The theoretical analysis results are verified by computer simulations.