Efficient Use Of Sparse Adaptive Filters

Size: px
Start display at page:

Download "Efficient Use Of Sparse Adaptive Filters"

Transcription

1 Efficient Use Of Sparse Adaptive Filters Andy W.H. Khong and Patrick A. Naylor Department of Electrical and Electronic Engineering, Imperial College ondon {andy.khong, Abstract We present a novel adaptive algorithm exploiting the sparseness of an impulse response for network echo cancellation. This sparseness-controlled improved proportionate normalized least mean square (SC-) algorithm is based on which allocates a step-size gain proportional to each filter coefficient. The proposed SC- algorithm achieves improved convergence over by estimating the sparseness of the impulse response and allocating gains for each stepsize such that a higher weighting is given to the proportionate term of the for sparse impulse responses. For a less sparse impulse response, a higher weighting will be allocated to the term. Simulation results presented show improved performance over the algorithm during convergence before and after an echo path change has been introduced. We also discuss the computational complexity of the proposed algorithm. xn ( ) h ˆ( n) en ( ) yn ( ) Fig.. yn ˆ( ) h( n) Schematic diagram of an echo canceller. wn ( ) I. INTRODUCTION Adaptive algorithms have been employed extensively in signal processing applications for system identification, channel equalization and prediction []. Classical adaptive algorithms with a uniform step-size across all filter coefficients, such as the normalized least-mean-square () algorithm, have slow convergence when estimating sparse impulse responses such as those that occur in some telecommunication networks. In such systems where traditional telephony equipment is connected to the packet switched network [], the echo path impulse response, which is typically of length 64-8 ms, exhibits an active region in the range of 8- ms duration and consequently, the impulse response is dominated by regions where magnitudes are close to zero, making the impulse response sparse. The inactive region is due to the presence of bulk delay caused by network propagation, encoding and jitter buffer delays [3]. Other applications for sparse system identification include wavelet identification using marine seismic signals [4] and geophysical seismic applications [5][6]. One of the first algorithms for network echo cancellation (NEC) is the proportionate (P) algorithm [7] where each filter coefficient is updated with an independent step-size that is proportional to the magnitude of that estimated filter coefficient. The P algorithm achieves fast initial convergence but subsequently slows down due to the slow convergence of coefficients having significantly smaller magnitudes. Subsequent improved versions such as the [8] and I [9] algorithms were proposed which achieve improved convergence by introducing a controlled mixture of proportionate (P) and non-proportionate () adaptation. Consequently, these algorithms perform better than P for sparse network impulse responses and in some cases for non-sparse acoustic impulse responses []. To reduce the computational complexity of P, the sparse partial update (SP) algorithm was proposed [] where, similar to the selective partial update (SPU- ) algorithm [], only taps corresponding to the M largest absolute values of the product between input signal and filter coefficients are selected for adaptation. An optimal step-size for P has been derived in [3] and employing an approximate μ-law function, it has been shown that the proposed segment P (SP) outperforms the P algorithm. Adaptive algorithms deployed for echo cancellation are expected to track any variations in the impulse response in order to achieve sufficient level of echo cancellation. The aim of this work is to develop an adaptive algorithm which utilizes the sparseness of an impulse response for NEC application. This is achieved by incorporating the sparseness measure [4][5] into such that the proportionate term in the update is allocated a higher weighting for increasing sparse systems. The result of this work is the sparseness-controlled (SC-) which is robust to the sparseness nature of the impulse response. This paper is organized as follows: In Section II, we review adaptive algorithms for sparse system identification. We then review and illustrate the sparseness of an impulse response in Section III while in Section IV we propose the SC- algorithm and discuss the computational complexity relative to the algorithm. Section V evaluates the performance of SC- while Section VI concludes our work /6/$. 375

2 II. SPARSE SYSTEM IDENTIFICATION We review the P and algorithms which have been proposed for sparse system identification. With reference to Fig., we first define filter coefficients and tap-input vector as ĥ(n) = [ĥ(n) ĥ(n)...ĥ (n)] T, () x(n) = [x(n) x(n )...x(n +)] T, () where is the adaptive filter length and the superscript T is defined as the transposition operator. The adaptive filter will model the unknown impulse response h(n) using the near-end signal y(n) =x T (n)h(n)+w(n), (3) where w(n) is the observation noise. For simplicity, we shall temporarily ignore the effects of w(n) in the description of algorithms. A. General formulation Sparse adaptive algorithms such as presented in [7][8] for network echo cancellation can be generalized using the following set of equations: T e(n) = y(n) x (n)ĥ(n ), (4) μq(n )x(n)e(n) ĥ(n) = ĥ(n ) + x T (n)q(n )x(n)+δ, (5) where μ is the adaptive step-size and δ is the regularization parameter. The diagonal control matrix Q(n ) = diag{q (n )... q (n )} (6) determines the step-size gain for each filter coefficient and is dependent on the specific algorithm as described below. B. The P algorithm The P algorithm [7] assigns higher step-sizes for coefficients with higher magnitude using a control matrix Q(n). Elements of this control matrix for P can be expressed as κ l (n) q l (n) = i= κ i(n), (7) { κ l (n) = max ρ } max{γ, ĥ(n)... ĥ (n) }, ĥl(n) (8) with l =,,..., being the tap-indices. The parameter γ, with a typical value of., prevents ĥl(n) from stalling during initialization stage where ĥ() = while ρ prevents coefficients from stalling when they are much smaller than the largest coefficient. The regularization parameter δ in (6) for P is taken as δ P = δ /, (9) where δ = σ x is the variance of the input signal [8]. It can be seen that for q l =, l, P is equivalent to. C. The algorithm An enhancement of P is the algorithm [8] which is a combination of P and with the relative significance of each controlled by a factor α. Elements of the control matrix Q(n) for are given by q l (n) = α ( + α) ĥl(n) +, l =,...,, () ĥ(n) + ɛ where ɛ is a small value and. is the l -norm operator. It can be seen from the second term of () that the magnitude of the estimated taps are normalized by the l -norm of ĥ. This shows that the weighting on the step-size for is dependent only on the relative scaling of the filter coefficients as oppose to their absolute values. Results presented in [8][6] have shown that good choices of α values are, -.5 and The regularization parameter δ in (6) for should be taken [8] as δ = α δ. () This choice of regularization ensures that the algorithm achieves the same steady-state normalized misalignment compared to that of the algorithm for the same stepsize, i.e., μ = μ. It can be seen that is equivalent to when α = while for α close to, behaves like P. III. SPARSENESS MEASURE The sparseness of an impulse response can be quantified by the sparseness measure [4][5] [ h(n) ξ(n) = ], h(n) () where h(n) is the l -norm of h(n). By considering impulse responses with various degrees of sparseness it can be shown that ξ(n). A. Sparseness of a delta function The sparseness of a length impulse defined by { ±v, n = n, h(n) =, n,n n, (3) where n defines the location of the impulse, can be computed from () by first considering h(n) = h l (n) = v, l= h(n) = h l (n) = v l= from which substituting into () gives ξ(n) =. 376

3 B. Sparseness of a signal with constant magnitude For h(n) of length having a constant magnitude ±v, we have h(n) = v = v, l= h(n) = v = v. l= Substituting these equations into (), the sparseness for h(n) is then ξ(n) =. C. Variation of ξ with sparseness Sparseness of impulse responses for NEC can be studied by generating synthetic impulses using random sequences. This can be achieved by first defining a vector u = [ p e /ψ e /ψ... e (u )/ψ] T, (4) where the p leading zeros models the length of the bulk delay and u = p is the length of the decaying window while ψ Z + is the decay constant which may be time-varying. For brevity of notation, we ignore the dependence of ψ on time index n. Defining a u vector b as a zero mean white Gaussian noise (WGN) sequence with variance σ b,the synthetic impulse response can then be expressed as B u u = diag{b}, [ ] p h(n) = p p u u + p, (5) u p B u u where the vector p ensures elements in the inactive region are small but non-zero and is an independent zero mean WGN sequence with variance σp. Figure shows an example set of impulse responses generated using (5) with σb =.87, σ p =.45 4, = 5 and p =3. Impulse responses with various degrees of sparseness are generated using decaying constants (a) ψ = 8, (b) ψ =, (c) ψ = 5 and (d) ψ = giving (a) ξ(n) =.845, (b)ξ(n) =.84, (c)ξ(n) =.674 and (d) ξ(n) =.57 respectively. Hence it can be seen that the sparseness of the impulse response increases with ξ(n). IV. THE SC- AGORITHM We propose the sparseness-controlled (SC- ) algorithm by taking into account the sparseness of an impulse response. As explained in Section I, it is known that the takes into account the slow convergence of the inactive region for the P algorithm. As a result, the fast convergence of is achieved by combining an term ( ( α)/() ) ( with the proportionate term (+α)( ĥ l (n) )/( ĥ +ɛ) ) where the relative significance of each term is controlled by the parameter α. As can be seen, the parameter α is chosen a priori for. The proposed SC- algorithm further improves the performance of by emphasizing the proportionate (a) (c) (b) (d) Fig.. Impulse responses controlled using (a) ψ =8,(b)ψ =,(c)ψ = 5 and (d) ψ = giving sparseness measure (a) ξ =.845, (b) ξ =.84, (c) ξ =.674 and (d) ξ =.57. term if the impulse response is significantly sparse. For relatively less sparse impulse responses, SC- allocates a higher weighting to the term. Similar to α, incorporation of ξ(n) into the algorithm can be achieved using two terms +ξ(n), for the proportionate term, and ξ(n), for the term in (). Direct incorporation of ξ(n) into () however is not feasible since (i) the impulse response h(n) is unknown and (ii) ξ(n) can be close to unity. The approach to (i) is to estimate ξ(n) using filter estimates obtained for each iteration giving [ ξ(n) = ĥ(n) ]. (6) ĥ(n) For n =,thel -norm of the filter coefficients ĥ(n) = and hence to prevent division by a small number or zero, this estimation can be computed for n. Forn<elements of the control matrix q l (n) is computed using (). Although it is highly unlikely that ξ(n) =, a high estimate of ξ(n) will introduce significant coefficient noise during adaptation since the effective step-size is increased. Hence to address (ii), we reduce the effect of ξ(n) on q l (n) using [ +.5 ξ(n)]/ and [.5 ξ(n)]/ for the proportionate and terms respectively. The computation of q l (n) for the SC- algorithm can then be expressed as q l (n) = [ ].5 ξ(n) α + [ ] +.5 ξ(n) ( + α) ĥ l (n) ĥ + ɛ (7) for n and l =,...,. The proposed SC- algorithm can be described by ()-(6) and (6), (7). 377

4 SC SC SC Fig. 3. Relative convergence of, and SC- using WGN input. Impulse response Fig. (b) is used. [α =.75, μ = μ =.3, μ SC =.7, SNR = 3 db]. A. Computational complexity We consider the computational complexity of the proposed SC- algorithm in terms of the number of multiplications required relative to the algorithm for each time iteration. The requires O(4) multiplications per iteration [9][6]. The additional complexity of the SC- algorithm compared to arises from the computation of ξ(n) and q l (n) given by (6) and (7). Given that /( ) can be computed off-line, and since ĥ(n) computation is already available from (), the remaining computation of (6) requires only + multiplications. For (7), 5 multiplications are required in addition to that for the. Hence, the complexity for the SC- is O(5). V. SIMUATION RESUTS The performance of SC- is compared with and in the context of system identification for network echo cancellation. This performance can be quantified using the normalized misalignment defined by η(n) = h(n) ĥ(n) h(n). (8) We assumed throughout our simulation that the length of the adaptive filter is equivalent to that of the unknown system. Any effects of undermodelling will introduce a bias to the normalized misalignment [7]. A. Simulations using WGN input In this first experiment we used impulse response as shown in Fig. (b) with = 5. A zero mean white Gaussian noise (WGN) sequence is used as the input signal while another zero mean WGN sequence w(n), as illustrated in Fig., is added to achieve a signal-to-noise ratio (SNR) of 3 db. The stepsizes for and are μ = μ = Fig. 4. Relative convergence of, and SC- using WGN input with echo path change at.5 s. Impulse response is changed from Fig. (a) to (b). [α =.75, μ = μ =.3, μ SC =.7, SNR = 3 db]. while the step-size for SC- is adjusted in order to achieve the same steady-state normalized misalignment as that for and. This corresponds to μ SC =.7. Similar to [6], we have used α =.75 for and SC-. Figure 3 shows convergence result where we note that SC- achieves improved normalized misalignment of approximately 7 db compared to that of the during initial convergence. Both and SC- achieve a higher rate of convergence than the as expected. In the second experiment, we used impulse responses as shown in Fig. (a) and (b) with = 5. As before, a zero mean WGN sequence is used as the input signal while an SNR of 3 db is used. The step-sizes are μ = μ =.3 and μ SC =.7 with α =.75 for and SC-. Figure 4 shows convergence result when an echo path change is introduced midway through the simulation where we changed the impulse response from a sparse to one which is relatively less sparse as shown in Fig. (a) to (b). We note from this result that SC- achieves improved normalized misalignment of approximately 6 db compared to that of the during initial convergence. After the echo path change at.5 s, a 3 db improvement in normalized misalignment can be seen for SC- compared to that of. As before, both and SC- achieve a higher rate of convergence (during initial phase and after echo path change) than the as expected. Figure 5 shows results when the initial impulse response is less sparse (Fig. (b)) compared to that after the echo path change (Fig. (a)). Step-size and α parameters are the same used for the previous experiment while the SNR is also at 3 db. From Fig. 5, the SC- algorithm achieves the highest rate of convergence with an approximately 7 and 3 db improvement in normalized misalignment over that for the during convergence before and after echo path 378

5 3 SC SC Fig. 5. Relative convergence of, and SC- using WGN input with echo path change at.5 s. Impulse response is changed from Fig. (b) to (a). [α =.75, μ = μ =.3, μ SC =.7, SNR = 3 db]. Speech SC 6 SC Fig. 6. Relative convergence of, and SC- using speech input with echo path change midway through simulation. Impulse response is changed from Fig. (a) to (b). [α =.75, μ = μ =.3, μ SC =.7, SNR = 3 db]. change respectively. B. Simulation using speech input Figure 6 shows results using speech input from a male speaker. The signal comprises 6 sentences from the same male speaker. In this experiment, a sparse impulse response is changed to one that is less sparse as shown in Fig. (a) to (b). As before, we have used μ = μ =.3, μ SC =.7in order for the algorithms to achieve the same steady-state normalized misalignment. We have also used α =.75 for the and SC- algorithms. An SNR of 3 db is used for this experiment. As before, the proposed SC- algorithm achieves the highest rate of convergence. Compared to the algorithm, SC- achieves an improvement of approximately.9 db normalized misalignment during initial convergence. VI. CONCUSION We have proposed the SC- algorithm for network echo cancellation. This algorithm estimates the sparseness of an impulse response and allocates a higher weighting to the proportionate term in the gain matrix for a relatively more sparse impulse response compared to one which is less sparse. Simulation results presented showed approximately.9 to 7 db improvement in convergence for the SC- algorithm compared to. REFERENCES [] S. Haykin, Adaptive Filter Theory, 4th ed., ser. Information and System Science. Prentice Hall,. [] M. M. Sondhi and D. A. Berkley, Silencing echoes on the telephone network, in Proc. IEEE, vol. 68, Aug. 98, pp [3] J. Radecki, Z. Zilic, and K. Radecka, Echo cancellation in IP networks, in Proc. Fourty-Fifth Midwest Symposium on Circuits and Systems, vol.,, pp. 9. [4] M. Boujida and J.-M. Boucher, Higher order statistics applied to wavelet identification of marine seismic signals, in Proc. Eur. Signal Process. Conf., 996. [5] Y.-F. Cheng and D. M. Etter, Analysis of an adaptive technique for modeling sparse systems, IEEE Trans on Acoustics, Speech, and Signal Processing, vol. 37, no., pp , Feb [6] E. A. Robinson and T. S. Durrani, Geophysical Signal Processing. Englewood Cliffs, NJ: Prentice-Hall, 986. [7] D.. Duttweiler, Proportionate normalized least mean square adaptation in echo cancellers, IEEE Trans. Speech Audio Processing, vol. 8, no. 5, pp , Sep.. [8] J. Benesty and S.. Gay, An improved P algorithm, in Proc. IEEE Int. Conf. Acoustics Speech Signal Processing, vol.,, pp [9] J. Cui, P. A. Naylor, and D. T. Brown, An improved algorithm for echo cancellation in packet-switched networks, in Proc. IEEE Int. Conf. on Signal Processing, vol. 4, May 4, pp [] R. Ahmad, A. W. Khong, and P. A. Naylor, Proportionate frequency domain adaptive algorithms for blind channel identification, in Proc. IEEE Int. Conf. Acoustics Speech Signal Processing, vol. 5, May 6, pp. V9 V3. [] H. Deng and M. Doroslovacki, New sparse adaptive algorithms using partial update, in Proc. IEEE Int. Conf. on Signal Processing, vol., May 4, pp [] K. Dogancay and O. Tanrikulu, Adaptive filtering algorithms with selective partial updates, IEEE Trans. Circuits Syst. II, vol. 48, no. 8, pp , Aug.. [3] H. Deng and M. Doroslovacki, Improving convergence of the P algorithm for sparse impulse response identification, IEEE Signal Processing ett., vol., no. 3, pp. 8 84, Mar. 5. [4] P. O. Hoyer, Non-negative matrix factorization with sparseness constraints, Journal of Machine earning Research, vol. 5, pp , Nov. 4. [5] J. Benesty, Y. A. Huang, J. Chen, and P. A. Naylor, Adaptive algorithms for the identification of sparse impulse responses, in Selected methods for acoustic echo and noise control, E. Hänsler and G. Schmidt, Eds. Springer, 6, ch. 5, pp [6] A. W. H. Khong, J. Benesty, and P. A. Naylor, An improved proportionate multi-delay block adaptive filter for packet-switched network echo cancellation, in Proc. European Signal Processing Conference, Sep. 5. [7] C. Breining, Control of a hands-free telephone set, Signal Processing, vol. 6, pp. 3 43,

A SPARSENESS CONTROLLED PROPORTIONATE ALGORITHM FOR ACOUSTIC ECHO CANCELLATION

A SPARSENESS CONTROLLED PROPORTIONATE ALGORITHM FOR ACOUSTIC ECHO CANCELLATION 6th European Signal Processing Conference (EUSIPCO 28), Lausanne, Switzerland, August 25-29, 28, copyright by EURASIP A SPARSENESS CONTROLLED PROPORTIONATE ALGORITHM FOR ACOUSTIC ECHO CANCELLATION Pradeep

More information

Sparseness-Controlled Affine Projection Algorithm for Echo Cancelation

Sparseness-Controlled Affine Projection Algorithm for Echo Cancelation Sparseness-Controlled Affine Projection Algorithm for Echo Cancelation ei iao and Andy W. H. Khong E-mail: liao38@e.ntu.edu.sg E-mail: andykhong@ntu.edu.sg Nanyang Technological University, Singapore Abstract

More information

A SPARSENESS CONTROLLED PROPORTIONATE ALGORITHM FOR ACOUSTIC ECHO CANCELLATION

A SPARSENESS CONTROLLED PROPORTIONATE ALGORITHM FOR ACOUSTIC ECHO CANCELLATION A SPARSENESS CONTROLLED PROPORTIONATE ALGORITHM FOR ACOUSTIC ECHO CANCELLATION Pradeep Loganathan, Andy W.H. Khong, Patrick A. Naylor pradeep.loganathan@ic.ac.uk, andykhong@ntu.edu.sg, p.naylorg@ic.ac.uk

More information

A low intricacy variable step-size partial update adaptive algorithm for Acoustic Echo Cancellation USNRao

A low intricacy variable step-size partial update adaptive algorithm for Acoustic Echo Cancellation USNRao ISSN: 77-3754 International Journal of Engineering and Innovative echnology (IJEI Volume 1, Issue, February 1 A low intricacy variable step-size partial update adaptive algorithm for Acoustic Echo Cancellation

More information

System Identification and Adaptive Filtering in the Short-Time Fourier Transform Domain

System Identification and Adaptive Filtering in the Short-Time Fourier Transform Domain System Identification and Adaptive Filtering in the Short-Time Fourier Transform Domain Electrical Engineering Department Technion - Israel Institute of Technology Supervised by: Prof. Israel Cohen Outline

More information

Samira A. Mahdi University of Babylon/College of Science/Physics Dept. Iraq/Babylon

Samira A. Mahdi University of Babylon/College of Science/Physics Dept. Iraq/Babylon Echo Cancelation Using Least Mean Square (LMS) Algorithm Samira A. Mahdi University of Babylon/College of Science/Physics Dept. Iraq/Babylon Abstract The aim of this work is to investigate methods for

More information

Department of Electrical and Electronic Engineering

Department of Electrical and Electronic Engineering Imperial College London Department of Electrical and Electronic Engineering Final Year Project Report 27 Project Title: Student: Course: Adaptive Echo Cancellation Pradeep Loganathan ISE4 Project Supervisor:

More information

Research Article Frequency-Domain Adaptive Algorithm for Network Echo Cancellation in VoIP

Research Article Frequency-Domain Adaptive Algorithm for Network Echo Cancellation in VoIP Hindawi Publishing Corporation EURASIP Journal on Audio, Speech, and Music Processing Volume 28, Article ID 15696, 9 pages doi:1.1155/28/15696 Research Article Frequency-Domain Adaptive Algorithm for Network

More information

Adaptive filters for sparse system identification

Adaptive filters for sparse system identification Scholars' Mine Doctoral Dissertations Student heses and Dissertations Spring 2016 Adaptive filters for sparse system identification Jianming Liu Follow this and additional works at: http://scholarsmine.mst.edu/doctoral_dissertations

More information

MITIGATING UNCORRELATED PERIODIC DISTURBANCE IN NARROWBAND ACTIVE NOISE CONTROL SYSTEMS

MITIGATING UNCORRELATED PERIODIC DISTURBANCE IN NARROWBAND ACTIVE NOISE CONTROL SYSTEMS 17th European Signal Processing Conference (EUSIPCO 29) Glasgow, Scotland, August 24-28, 29 MITIGATING UNCORRELATED PERIODIC DISTURBANCE IN NARROWBAND ACTIVE NOISE CONTROL SYSTEMS Muhammad Tahir AKHTAR

More information

Research Letter An Algorithm to Generate Representations of System Identification Errors

Research Letter An Algorithm to Generate Representations of System Identification Errors Research Letters in Signal Processing Volume 008, Article ID 5991, 4 pages doi:10.1155/008/5991 Research Letter An Algorithm to Generate Representations of System Identification Errors Wancheng Zhang and

More information

Chapter 2 Fundamentals of Adaptive Filter Theory

Chapter 2 Fundamentals of Adaptive Filter Theory Chapter 2 Fundamentals of Adaptive Filter Theory In this chapter we will treat some fundamentals of the adaptive filtering theory highlighting the system identification problem We will introduce a signal

More information

STATISTICAL MODELLING OF MULTICHANNEL BLIND SYSTEM IDENTIFICATION ERRORS. Felicia Lim, Patrick A. Naylor

STATISTICAL MODELLING OF MULTICHANNEL BLIND SYSTEM IDENTIFICATION ERRORS. Felicia Lim, Patrick A. Naylor STTISTICL MODELLING OF MULTICHNNEL BLIND SYSTEM IDENTIFICTION ERRORS Felicia Lim, Patrick. Naylor Dept. of Electrical and Electronic Engineering, Imperial College London, UK {felicia.lim6, p.naylor}@imperial.ac.uk

More information

Adaptive Sparse System Identification Using Wavelets

Adaptive Sparse System Identification Using Wavelets 656 IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II: ANALOG AND DIGITAL SIGNAL PROCESSING, VOL. 49, NO. 10, OCTOBER 2002 Adaptive Sparse System Identification Using Wavelets K. C. Ho, Senior Member, IEEE,

More information

A Flexible ICA-Based Method for AEC Without Requiring Double-Talk Detection

A Flexible ICA-Based Method for AEC Without Requiring Double-Talk Detection APSIPA ASC 2011 Xi an A Flexible ICA-Based Method for AEC Without Requiring Double-Talk Detection Marko Kanadi, Muhammad Tahir Akhtar, Wataru Mitsuhashi Department of Information and Communication Engineering,

More information

Research Article Efficient Multichannel NLMS Implementation for Acoustic Echo Cancellation

Research Article Efficient Multichannel NLMS Implementation for Acoustic Echo Cancellation Hindawi Publishing Corporation EURASIP Journal on Audio, Speech, and Music Processing Volume 27, Article ID 78439, 6 pages doi:1.1155/27/78439 Research Article Efficient Multichannel NLMS Implementation

More information

Adaptive MMSE Equalizer with Optimum Tap-length and Decision Delay

Adaptive MMSE Equalizer with Optimum Tap-length and Decision Delay Adaptive MMSE Equalizer with Optimum Tap-length and Decision Delay Yu Gong, Xia Hong and Khalid F. Abu-Salim School of Systems Engineering The University of Reading, Reading RG6 6AY, UK E-mail: {y.gong,x.hong,k.f.abusalem}@reading.ac.uk

More information

Sparse Least Mean Square Algorithm for Estimation of Truncated Volterra Kernels

Sparse Least Mean Square Algorithm for Estimation of Truncated Volterra Kernels Sparse Least Mean Square Algorithm for Estimation of Truncated Volterra Kernels Bijit Kumar Das 1, Mrityunjoy Chakraborty 2 Department of Electronics and Electrical Communication Engineering Indian Institute

More information

COMPLEX PROPORTIONATE-TYPE NORMALIZED LEAST MEAN SQUARE ALGORITHMS

COMPLEX PROPORTIONATE-TYPE NORMALIZED LEAST MEAN SQUARE ALGORITHMS COMPEX PROPORTIONATE-TYPE NORMAIZED EAST MEAN SQUARE AGORITHMS Kevin T. Wagner Naval Research aboratory Radar Division Washington, DC 375, USA Miloš I. Doroslovački The George Washington University Department

More information

Convergence Evaluation of a Random Step-Size NLMS Adaptive Algorithm in System Identification and Channel Equalization

Convergence Evaluation of a Random Step-Size NLMS Adaptive Algorithm in System Identification and Channel Equalization Convergence Evaluation of a Random Step-Size NLMS Adaptive Algorithm in System Identification and Channel Equalization 1 Shihab Jimaa Khalifa University of Science, Technology and Research (KUSTAR) Faculty

More information

FAST IMPLEMENTATION OF A SUBBAND ADAPTIVE ALGORITHM FOR ACOUSTIC ECHO CANCELLATION

FAST IMPLEMENTATION OF A SUBBAND ADAPTIVE ALGORITHM FOR ACOUSTIC ECHO CANCELLATION Journal of ELECTRICAL ENGINEERING, VOL. 55, NO. 5-6, 24, 113 121 FAST IMPLEMENTATION OF A SUBBAND ADAPTIVE ALGORITHM FOR ACOUSTIC ECHO CANCELLATION Khaled Mayyas The block subband adaptive algorithm in

More information

Reduced-cost combination of adaptive filters for acoustic echo cancellation

Reduced-cost combination of adaptive filters for acoustic echo cancellation Reduced-cost combination of adaptive filters for acoustic echo cancellation Luis A. Azpicueta-Ruiz and Jerónimo Arenas-García Dept. Signal Theory and Communications, Universidad Carlos III de Madrid Leganés,

More information

A DISSERTATION SUBMITTED TO THE FACULTY OF THE GRADUATE SCHOOL OF THE UNIVERSITY OF MINNESOTA BY. Jie Yang

A DISSERTATION SUBMITTED TO THE FACULTY OF THE GRADUATE SCHOOL OF THE UNIVERSITY OF MINNESOTA BY. Jie Yang Adaptive Filter Design for Sparse Signal Estimation A DISSERTATION SUBMITTED TO THE FACULTY OF THE GRADUATE SCHOOL OF THE UNIVERSITY OF MINNESOTA BY Jie Yang IN PARTIAL FULFILLMENT OF THE REQUIREMENTS

More information

A Robust Zero-point Attraction LMS Algorithm on Near Sparse System Identification

A Robust Zero-point Attraction LMS Algorithm on Near Sparse System Identification A Robust Zero-point Attraction LMS Algorithm on Near Sparse System Identification Jian Jin, Qing Qu, and Yuantao Gu Received Feb. 2012; accepted Feb. 2013. This article will appear in IET Signal Processing.

More information

Acoustic Echo Cancellation Using a Vector-Space-Based Adaptive Filtering Algorithm

Acoustic Echo Cancellation Using a Vector-Space-Based Adaptive Filtering Algorithm Acoustic Echo Cancellation Using a Vector-Space-Based Adaptive Filtering Algorithm Yu Tsao, Member IEEE, Shih-Hau Fang*, Senior Member IEEE, and Yao Shiao Abstract A novel vector-space-based adaptive filtering

More information

An overview on optimized NLMS algorithms for acoustic echo cancellation

An overview on optimized NLMS algorithms for acoustic echo cancellation Paleologu et al. EURASIP Journal on Advances in Signal Processing (015) 015:97 DOI 10.1186/s13634-015-083-1 REVIEW Open Access An overview on optimized NLMS algorithms for acoustic echo cancellation Constantin

More information

Adaptive Filter Theory

Adaptive Filter Theory 0 Adaptive Filter heory Sung Ho Cho Hanyang University Seoul, Korea (Office) +8--0-0390 (Mobile) +8-10-541-5178 dragon@hanyang.ac.kr able of Contents 1 Wiener Filters Gradient Search by Steepest Descent

More information

NOISE ROBUST RELATIVE TRANSFER FUNCTION ESTIMATION. M. Schwab, P. Noll, and T. Sikora. Technical University Berlin, Germany Communication System Group

NOISE ROBUST RELATIVE TRANSFER FUNCTION ESTIMATION. M. Schwab, P. Noll, and T. Sikora. Technical University Berlin, Germany Communication System Group NOISE ROBUST RELATIVE TRANSFER FUNCTION ESTIMATION M. Schwab, P. Noll, and T. Sikora Technical University Berlin, Germany Communication System Group Einsteinufer 17, 1557 Berlin (Germany) {schwab noll

More information

Submitted to Electronics Letters. Indexing terms: Signal Processing, Adaptive Filters. The Combined LMS/F Algorithm Shao-Jen Lim and John G. Harris Co

Submitted to Electronics Letters. Indexing terms: Signal Processing, Adaptive Filters. The Combined LMS/F Algorithm Shao-Jen Lim and John G. Harris Co Submitted to Electronics Letters. Indexing terms: Signal Processing, Adaptive Filters. The Combined LMS/F Algorithm Shao-Jen Lim and John G. Harris Computational Neuro-Engineering Laboratory University

More information

Integrated Direct Sub-band Adaptive Volterra Filter and Its Application to Identification of Loudspeaker Nonlinearity

Integrated Direct Sub-band Adaptive Volterra Filter and Its Application to Identification of Loudspeaker Nonlinearity Integrated Direct Sub-band Adaptive Volterra Filter and Its Application to Identification of Loudspeaker Nonlinearity Satoshi Kinoshita Department of Electrical and Electronic Engineering, Kansai University

More information

NONLINEAR ECHO CANCELLATION FOR HANDS-FREE SPEAKERPHONES. Bryan S. Nollett and Douglas L. Jones

NONLINEAR ECHO CANCELLATION FOR HANDS-FREE SPEAKERPHONES. Bryan S. Nollett and Douglas L. Jones NONLINEAR ECHO CANCELLATION FOR HANDS-FREE SPEAKERPHONES Bryan S. Nollett and Douglas L. Jones Coordinated Science Laboratory University of Illinois at Urbana-Champaign 1308 W. Main St. Urbana, IL 61801

More information

Alpha-Stable Distributions in Signal Processing of Audio Signals

Alpha-Stable Distributions in Signal Processing of Audio Signals Alpha-Stable Distributions in Signal Processing of Audio Signals Preben Kidmose, Department of Mathematical Modelling, Section for Digital Signal Processing, Technical University of Denmark, Building 3,

More information

KNOWN approaches for improving the performance of

KNOWN approaches for improving the performance of IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II: EXPRESS BRIEFS, VOL. 58, NO. 8, AUGUST 2011 537 Robust Quasi-Newton Adaptive Filtering Algorithms Md. Zulfiquar Ali Bhotto, Student Member, IEEE, and Andreas

More information

A STATE-SPACE PARTITIONED-BLOCK ADAPTIVE FILTER FOR ECHO CANCELLATION USING INTER-BAND CORRELATIONS IN THE KALMAN GAIN COMPUTATION

A STATE-SPACE PARTITIONED-BLOCK ADAPTIVE FILTER FOR ECHO CANCELLATION USING INTER-BAND CORRELATIONS IN THE KALMAN GAIN COMPUTATION A STATE-SPACE PARTITIONED-BLOCK ADAPTIVE FILTER FOR ECHO CANCELLATION USING INTER-BAND CORRELATIONS IN THE KALAN GAIN COPUTATION aría Luis Valero, Edwin abande and Emanuël A. P. Habets International Audio

More information

System Identification in the Short-Time Fourier Transform Domain

System Identification in the Short-Time Fourier Transform Domain System Identification in the Short-Time Fourier Transform Domain Electrical Engineering Department Technion - Israel Institute of Technology Supervised by: Prof. Israel Cohen Outline Representation Identification

More information

Image Compression using DPCM with LMS Algorithm

Image Compression using DPCM with LMS Algorithm Image Compression using DPCM with LMS Algorithm Reenu Sharma, Abhay Khedkar SRCEM, Banmore -----------------------------------------------------------------****---------------------------------------------------------------

More information

An Adaptive Sensor Array Using an Affine Combination of Two Filters

An Adaptive Sensor Array Using an Affine Combination of Two Filters An Adaptive Sensor Array Using an Affine Combination of Two Filters Tõnu Trump Tallinn University of Technology Department of Radio and Telecommunication Engineering Ehitajate tee 5, 19086 Tallinn Estonia

More information

Blind Deconvolution of Ultrasonic Signals Using High-Order Spectral Analysis and Wavelets

Blind Deconvolution of Ultrasonic Signals Using High-Order Spectral Analysis and Wavelets Blind Deconvolution of Ultrasonic Signals Using High-Order Spectral Analysis and Wavelets Roberto H. Herrera, Eduardo Moreno, Héctor Calas, and Rubén Orozco 3 University of Cienfuegos, Cuatro Caminos,

More information

A Derivation of the Steady-State MSE of RLS: Stationary and Nonstationary Cases

A Derivation of the Steady-State MSE of RLS: Stationary and Nonstationary Cases A Derivation of the Steady-State MSE of RLS: Stationary and Nonstationary Cases Phil Schniter Nov. 0, 001 Abstract In this report we combine the approach of Yousef and Sayed [1] with that of Rupp and Sayed

More information

ADaptive signal processing algorithms have been widely

ADaptive signal processing algorithms have been widely Variable Step-Size Affine Projection Algorithm with a Weighted and Regularized Projection Matrix Tao Dai, Andy Adler, and Behnam Shahrrava Abstract This paper presents a forgetting factor scheme for variable

More information

GAUSSIANIZATION METHOD FOR IDENTIFICATION OF MEMORYLESS NONLINEAR AUDIO SYSTEMS

GAUSSIANIZATION METHOD FOR IDENTIFICATION OF MEMORYLESS NONLINEAR AUDIO SYSTEMS GAUSSIANIATION METHOD FOR IDENTIFICATION OF MEMORYLESS NONLINEAR AUDIO SYSTEMS I. Marrakchi-Mezghani (1),G. Mahé (2), M. Jaïdane-Saïdane (1), S. Djaziri-Larbi (1), M. Turki-Hadj Alouane (1) (1) Unité Signaux

More information

Title without the persistently exciting c. works must be obtained from the IEE

Title without the persistently exciting c.   works must be obtained from the IEE Title Exact convergence analysis of adapt without the persistently exciting c Author(s) Sakai, H; Yang, JM; Oka, T Citation IEEE TRANSACTIONS ON SIGNAL 55(5): 2077-2083 PROCESS Issue Date 2007-05 URL http://hdl.handle.net/2433/50544

More information

A Multichannel Acoustic Echo Canceler Double-Talk Detector Based on a Normalized Cross-Correlation Matrix

A Multichannel Acoustic Echo Canceler Double-Talk Detector Based on a Normalized Cross-Correlation Matrix Acoustic Echo and Noise Control A Multichannel Acoustic Echo Canceler Double-Talk Detector Based on a Normalized Cross-Correlation Matrix JACOB BENESTY Bell Laboratories, Lucent Technologies, 700 Mountain

More information

Acoustic MIMO Signal Processing

Acoustic MIMO Signal Processing Yiteng Huang Jacob Benesty Jingdong Chen Acoustic MIMO Signal Processing With 71 Figures Ö Springer Contents 1 Introduction 1 1.1 Acoustic MIMO Signal Processing 1 1.2 Organization of the Book 4 Part I

More information

A METHOD OF ADAPTATION BETWEEN STEEPEST- DESCENT AND NEWTON S ALGORITHM FOR MULTI- CHANNEL ACTIVE CONTROL OF TONAL NOISE AND VIBRATION

A METHOD OF ADAPTATION BETWEEN STEEPEST- DESCENT AND NEWTON S ALGORITHM FOR MULTI- CHANNEL ACTIVE CONTROL OF TONAL NOISE AND VIBRATION A METHOD OF ADAPTATION BETWEEN STEEPEST- DESCENT AND NEWTON S ALGORITHM FOR MULTI- CHANNEL ACTIVE CONTROL OF TONAL NOISE AND VIBRATION Jordan Cheer and Stephen Daley Institute of Sound and Vibration Research,

More information

IMPLEMENTATION OF SIGNAL POWER ESTIMATION METHODS

IMPLEMENTATION OF SIGNAL POWER ESTIMATION METHODS IMPLEMENTATION OF SIGNAL POWER ESTIMATION METHODS Sei-Yeu Cheng and Joseph B. Evans Telecommunications & Information Sciences Laboratory Department of Electrical Engineering & Computer Science University

More information

IN real-world adaptive filtering applications, severe impairments

IN real-world adaptive filtering applications, severe impairments 1878 IEEE TRANSACTIONS ON SIGNAL PROCESSING, VOL. 56, NO. 5, MAY 2008 A New Robust Variable Step-Size NLMS Algorithm Leonardo Rey Vega, Hernán Rey, Jacob Benesty, Senior Member, IEEE, and Sara Tressens

More information

2262 IEEE TRANSACTIONS ON SIGNAL PROCESSING, VOL. 47, NO. 8, AUGUST A General Class of Nonlinear Normalized Adaptive Filtering Algorithms

2262 IEEE TRANSACTIONS ON SIGNAL PROCESSING, VOL. 47, NO. 8, AUGUST A General Class of Nonlinear Normalized Adaptive Filtering Algorithms 2262 IEEE TRANSACTIONS ON SIGNAL PROCESSING, VOL. 47, NO. 8, AUGUST 1999 A General Class of Nonlinear Normalized Adaptive Filtering Algorithms Sudhakar Kalluri, Member, IEEE, and Gonzalo R. Arce, Senior

More information

NEW STEIGLITZ-McBRIDE ADAPTIVE LATTICE NOTCH FILTERS

NEW STEIGLITZ-McBRIDE ADAPTIVE LATTICE NOTCH FILTERS NEW STEIGLITZ-McBRIDE ADAPTIVE LATTICE NOTCH FILTERS J.E. COUSSEAU, J.P. SCOPPA and P.D. DOÑATE CONICET- Departamento de Ingeniería Eléctrica y Computadoras Universidad Nacional del Sur Av. Alem 253, 8000

More information

New Insights Into the Stereophonic Acoustic Echo Cancellation Problem and an Adaptive Nonlinearity Solution

New Insights Into the Stereophonic Acoustic Echo Cancellation Problem and an Adaptive Nonlinearity Solution IEEE TRANSACTIONS ON SPEECH AND AUDIO PROCESSING, VOL 10, NO 5, JULY 2002 257 New Insights Into the Stereophonic Acoustic Echo Cancellation Problem and an Adaptive Nonlinearity Solution Tomas Gänsler,

More information

IS NEGATIVE STEP SIZE LMS ALGORITHM STABLE OPERATION POSSIBLE?

IS NEGATIVE STEP SIZE LMS ALGORITHM STABLE OPERATION POSSIBLE? IS NEGATIVE STEP SIZE LMS ALGORITHM STABLE OPERATION POSSIBLE? Dariusz Bismor Institute of Automatic Control, Silesian University of Technology, ul. Akademicka 16, 44-100 Gliwice, Poland, e-mail: Dariusz.Bismor@polsl.pl

More information

COMPLEX WAVELET TRANSFORM IN SIGNAL AND IMAGE ANALYSIS

COMPLEX WAVELET TRANSFORM IN SIGNAL AND IMAGE ANALYSIS COMPLEX WAVELET TRANSFORM IN SIGNAL AND IMAGE ANALYSIS MUSOKO VICTOR, PROCHÁZKA ALEŠ Institute of Chemical Technology, Department of Computing and Control Engineering Technická 905, 66 8 Prague 6, Cech

More information

Blind Deconvolution via Maximum Kurtosis Adaptive Filtering

Blind Deconvolution via Maximum Kurtosis Adaptive Filtering Blind Deconvolution via Maximum Kurtosis Adaptive Filtering Deborah Pereg Doron Benzvi The Jerusalem College of Engineering Jerusalem, Israel doronb@jce.ac.il, deborahpe@post.jce.ac.il ABSTRACT In this

More information

SINGLE-CHANNEL SPEECH PRESENCE PROBABILITY ESTIMATION USING INTER-FRAME AND INTER-BAND CORRELATIONS

SINGLE-CHANNEL SPEECH PRESENCE PROBABILITY ESTIMATION USING INTER-FRAME AND INTER-BAND CORRELATIONS 204 IEEE International Conference on Acoustic, Speech and Signal Processing (ICASSP) SINGLE-CHANNEL SPEECH PRESENCE PROBABILITY ESTIMATION USING INTER-FRAME AND INTER-BAND CORRELATIONS Hajar Momeni,2,,

More information

BLOCK-BASED MULTICHANNEL TRANSFORM-DOMAIN ADAPTIVE FILTERING

BLOCK-BASED MULTICHANNEL TRANSFORM-DOMAIN ADAPTIVE FILTERING BLOCK-BASED MULTICHANNEL TRANSFORM-DOMAIN ADAPTIVE FILTERING Sascha Spors, Herbert Buchner, and Karim Helwani Deutsche Telekom Laboratories, Technische Universität Berlin, Ernst-Reuter-Platz 7, 10587 Berlin,

More information

EIGENFILTERS FOR SIGNAL CANCELLATION. Sunil Bharitkar and Chris Kyriakakis

EIGENFILTERS FOR SIGNAL CANCELLATION. Sunil Bharitkar and Chris Kyriakakis EIGENFILTERS FOR SIGNAL CANCELLATION Sunil Bharitkar and Chris Kyriakakis Immersive Audio Laboratory University of Southern California Los Angeles. CA 9. USA Phone:+1-13-7- Fax:+1-13-7-51, Email:ckyriak@imsc.edu.edu,bharitka@sipi.usc.edu

More information

Blind Deconvolution by Modified Bussgang Algorithm

Blind Deconvolution by Modified Bussgang Algorithm Reprinted from ISCAS'99, International Symposium on Circuits and Systems, Orlando, Florida, May 3-June 2, 1999 Blind Deconvolution by Modified Bussgang Algorithm Simone Fiori, Aurelio Uncini, Francesco

More information

System Identification in the Short-Time Fourier Transform Domain

System Identification in the Short-Time Fourier Transform Domain System Identification in the Short-Time Fourier Transform Domain Yekutiel Avargel System Identification in the Short-Time Fourier Transform Domain Research Thesis As Partial Fulfillment of the Requirements

More information

Assesment of the efficiency of the LMS algorithm based on spectral information

Assesment of the efficiency of the LMS algorithm based on spectral information Assesment of the efficiency of the algorithm based on spectral information (Invited Paper) Aaron Flores and Bernard Widrow ISL, Department of Electrical Engineering, Stanford University, Stanford CA, USA

More information

NSLMS: a Proportional Weight Algorithm for Sparse Adaptive Filters

NSLMS: a Proportional Weight Algorithm for Sparse Adaptive Filters NSLMS: a Proportional Weight Algorithm for Sparse Adaptive Filters R. K. Martin and C. R. Johnson, Jr. School of Electrical Engineering Cornell University Ithaca, NY 14853 {frodo,johnson}@ece.cornell.edu

More information

BILINEAR forms were addressed in the literature in

BILINEAR forms were addressed in the literature in IEEE SIGNAL PROCESSING LEERS, VOL. 24, NO. 5, MAY 2017 653 On the Identification of Bilinear Forms With the Wiener Filter Jacob Benesty, Constantin Paleologu, Member, IEEE, and Silviu Ciochină, Senior

More information

Bayesian Regularization and Nonnegative Deconvolution for Room Impulse Response Estimation

Bayesian Regularization and Nonnegative Deconvolution for Room Impulse Response Estimation IEEE TRANSACTIONS ON SIGNAL PROCESSING, VOL. 54, NO. 3, MARCH 2006 839 Bayesian Regularization and Nonnegative Deconvolution for Room Impulse Response Estimation Yuanqing Lin and Daniel D. Lee Abstract

More information

GREEDY RLS FOR SPARSE FILTERS

GREEDY RLS FOR SPARSE FILTERS 18th European Signal Processing Conference (EUSIPCO-2010) Aalborg, Denmark, August 23-27, 2010 GREEDY RLS FOR SPARSE FILTERS Bogdan Dumitrescu, Ioan Tăbuş Department of Signal Processing Tampere University

More information

Dominant Pole Localization of FxLMS Adaptation Process in Active Noise Control

Dominant Pole Localization of FxLMS Adaptation Process in Active Noise Control APSIPA ASC 20 Xi an Dominant Pole Localization of FxLMS Adaptation Process in Active Noise Control Iman Tabatabaei Ardekani, Waleed H. Abdulla The University of Auckland, Private Bag 9209, Auckland, New

More information

LEAST Mean Squares Algorithm (LMS), introduced by

LEAST Mean Squares Algorithm (LMS), introduced by 1 Hard Threshold Least Mean Squares Algorithm Lampros Flokas and Petros Maragos arxiv:1608.0118v1 [cs.sy] 3 Aug 016 Abstract This work presents a new variation of the commonly used Least Mean Squares Algorithm

More information

Construction of low complexity Array based Quasi Cyclic Low density parity check (QC-LDPC) codes with low error floor

Construction of low complexity Array based Quasi Cyclic Low density parity check (QC-LDPC) codes with low error floor Construction of low complexity Array based Quasi Cyclic Low density parity check (QC-LDPC) codes with low error floor Pravin Salunkhe, Prof D.P Rathod Department of Electrical Engineering, Veermata Jijabai

More information

TECHNICAL MEMORANDUM

TECHNICAL MEMORANDUM Bell Laboratories subject: Proposals for Dynamic Resource Allocation for Network Echo Cancellation Work Project No. 31152-273 File Case 38794-43 date: March 9, 1999 from: Tomas Gaensler Lund University

More information

Performance Analysis of Norm Constraint Least Mean Square Algorithm

Performance Analysis of Norm Constraint Least Mean Square Algorithm IEEE TRANSACTIONS ON SIGNAL PROCESSING, VOL. 60, NO. 5, MAY 2012 2223 Performance Analysis of Norm Constraint Least Mean Square Algorithm Guolong Su, Jian Jin, Yuantao Gu, Member, IEEE, and Jian Wang Abstract

More information

DIRECTION ESTIMATION BASED ON SOUND INTENSITY VECTORS. Sakari Tervo

DIRECTION ESTIMATION BASED ON SOUND INTENSITY VECTORS. Sakari Tervo 7th European Signal Processing Conference (EUSIPCO 9) Glasgow, Scotland, August 4-8, 9 DIRECTION ESTIMATION BASED ON SOUND INTENSITY VECTORS Sakari Tervo Helsinki University of Technology Department of

More information

EFFECTS OF ILL-CONDITIONED DATA ON LEAST SQUARES ADAPTIVE FILTERS. Gary A. Ybarra and S.T. Alexander

EFFECTS OF ILL-CONDITIONED DATA ON LEAST SQUARES ADAPTIVE FILTERS. Gary A. Ybarra and S.T. Alexander EFFECTS OF ILL-CONDITIONED DATA ON LEAST SQUARES ADAPTIVE FILTERS Gary A. Ybarra and S.T. Alexander Center for Communications and Signal Processing Electrical and Computer Engineering Department North

More information

COMPARISON OF MULTIPLE-MICROPHONE AND SINGLE-LOUDSPEAKER ADAPTIVE FEEDBACK/ECHO CANCELLATION SYSTEMS

COMPARISON OF MULTIPLE-MICROPHONE AND SINGLE-LOUDSPEAKER ADAPTIVE FEEDBACK/ECHO CANCELLATION SYSTEMS 19th European Signal rocessing Conference EUSICO 211) Barcelona, Spain, August 29 - September 2, 211 COMARISON OF MULTILE-MICROHONE AND SINGLE-LOUDSEAKER ADATIVE FEEDBACK/ECHO CANCELLATION SYSTEMS Meng

More information

Variable, Step-Size, Block Normalized, Least Mean, Square Adaptive Filter: A Unied Framework

Variable, Step-Size, Block Normalized, Least Mean, Square Adaptive Filter: A Unied Framework Scientia Iranica, Vol. 15, No. 2, pp 195{202 c Sharif University of Technology, April 2008 Research Note Variable, Step-Size, Block Normalized, Least Mean, Square Adaptive Filter: A Unied Framework M.

More information

Optimal Speech Enhancement Under Signal Presence Uncertainty Using Log-Spectral Amplitude Estimator

Optimal Speech Enhancement Under Signal Presence Uncertainty Using Log-Spectral Amplitude Estimator 1 Optimal Speech Enhancement Under Signal Presence Uncertainty Using Log-Spectral Amplitude Estimator Israel Cohen Lamar Signal Processing Ltd. P.O.Box 573, Yokneam Ilit 20692, Israel E-mail: icohen@lamar.co.il

More information

A Local Model of Relative Transfer Functions Involving Sparsity

A Local Model of Relative Transfer Functions Involving Sparsity A Local Model of Relative Transfer Functions Involving Sparsity Zbyněk Koldovský 1, Jakub Janský 1, and Francesco Nesta 2 1 Faculty of Mechatronics, Informatics and Interdisciplinary Studies Technical

More information

Co-Prime Arrays and Difference Set Analysis

Co-Prime Arrays and Difference Set Analysis 7 5th European Signal Processing Conference (EUSIPCO Co-Prime Arrays and Difference Set Analysis Usham V. Dias and Seshan Srirangarajan Department of Electrical Engineering Bharti School of Telecommunication

More information

Mode Selection for Multi-Antenna Broadcast Channels

Mode Selection for Multi-Antenna Broadcast Channels Mode Selection for Multi-Antenna Broadcast Channels Gill November 22, 2011 Gill (University of Delaware) November 22, 2011 1 / 25 Part I Mode Selection for MISO BC with Perfect/Imperfect CSI [1]-[3] Gill

More information

Optimal and Adaptive Filtering

Optimal and Adaptive Filtering Optimal and Adaptive Filtering Murat Üney M.Uney@ed.ac.uk Institute for Digital Communications (IDCOM) 26/06/2017 Murat Üney (IDCOM) Optimal and Adaptive Filtering 26/06/2017 1 / 69 Table of Contents 1

More information

AdaptiveFilters. GJRE-F Classification : FOR Code:

AdaptiveFilters. GJRE-F Classification : FOR Code: Global Journal of Researches in Engineering: F Electrical and Electronics Engineering Volume 14 Issue 7 Version 1.0 Type: Double Blind Peer Reviewed International Research Journal Publisher: Global Journals

More information

Blind Source Separation with a Time-Varying Mixing Matrix

Blind Source Separation with a Time-Varying Mixing Matrix Blind Source Separation with a Time-Varying Mixing Matrix Marcus R DeYoung and Brian L Evans Department of Electrical and Computer Engineering The University of Texas at Austin 1 University Station, Austin,

More information

IEEE TRANSACTIONS ON SIGNAL PROCESSING, VOL. 57, NO. 10, OCTOBER X/$ IEEE

IEEE TRANSACTIONS ON SIGNAL PROCESSING, VOL. 57, NO. 10, OCTOBER X/$ IEEE IEEE TRANSACTIONS ON SIGNAL PROCESSING, VOL. 57, NO. 10, OCTOBER 2009 3891 Adaptive Nonlinear System Identification in the Short-Time Fourier Transform Domain Yekutiel Avargel, Student Member, IEEE, Israel

More information

A New Subspace Identification Method for Open and Closed Loop Data

A New Subspace Identification Method for Open and Closed Loop Data A New Subspace Identification Method for Open and Closed Loop Data Magnus Jansson July 2005 IR S3 SB 0524 IFAC World Congress 2005 ROYAL INSTITUTE OF TECHNOLOGY Department of Signals, Sensors & Systems

More information

Design and Realization of Quantum based Digital Integrator

Design and Realization of Quantum based Digital Integrator International Journal of Electronic and Electrical Engineering. ISSN 0974-2174, Volume 7, Number 5 (2014), pp. 449-454 International Research Publication House http://www.irphouse.com Design and Realization

More information

Performance analysis and design of FxLMS algorithm in broadband ANC system with online secondary-path modeling

Performance analysis and design of FxLMS algorithm in broadband ANC system with online secondary-path modeling Title Performance analysis design of FxLMS algorithm in broadb ANC system with online secondary-path modeling Author(s) Chan, SC; Chu, Y Citation IEEE Transactions on Audio, Speech Language Processing,

More information

Independent Component Analysis and Unsupervised Learning

Independent Component Analysis and Unsupervised Learning Independent Component Analysis and Unsupervised Learning Jen-Tzung Chien National Cheng Kung University TABLE OF CONTENTS 1. Independent Component Analysis 2. Case Study I: Speech Recognition Independent

More information

Adaptive sparse algorithms for estimating sparse channels in broadband wireless communications systems

Adaptive sparse algorithms for estimating sparse channels in broadband wireless communications systems Wireless Signal Processing & Networking Workshop: Emerging Wireless Technologies, Sendai, Japan, 28 Oct. 2013. Adaptive sparse algorithms for estimating sparse channels in broadband wireless communications

More information

Riccati difference equations to non linear extended Kalman filter constraints

Riccati difference equations to non linear extended Kalman filter constraints International Journal of Scientific & Engineering Research Volume 3, Issue 12, December-2012 1 Riccati difference equations to non linear extended Kalman filter constraints Abstract Elizabeth.S 1 & Jothilakshmi.R

More information

Transformation Techniques for Real Time High Speed Implementation of Nonlinear Algorithms

Transformation Techniques for Real Time High Speed Implementation of Nonlinear Algorithms International Journal of Electronics and Communication Engineering. ISSN 0974-66 Volume 4, Number (0), pp.83-94 International Research Publication House http://www.irphouse.com Transformation Techniques

More information

IMPROVED MULTI-MICROPHONE NOISE REDUCTION PRESERVING BINAURAL CUES

IMPROVED MULTI-MICROPHONE NOISE REDUCTION PRESERVING BINAURAL CUES IMPROVED MULTI-MICROPHONE NOISE REDUCTION PRESERVING BINAURAL CUES Andreas I. Koutrouvelis Richard C. Hendriks Jesper Jensen Richard Heusdens Circuits and Systems (CAS) Group, Delft University of Technology,

More information

SINGLE CHANNEL SPEECH MUSIC SEPARATION USING NONNEGATIVE MATRIX FACTORIZATION AND SPECTRAL MASKS. Emad M. Grais and Hakan Erdogan

SINGLE CHANNEL SPEECH MUSIC SEPARATION USING NONNEGATIVE MATRIX FACTORIZATION AND SPECTRAL MASKS. Emad M. Grais and Hakan Erdogan SINGLE CHANNEL SPEECH MUSIC SEPARATION USING NONNEGATIVE MATRIX FACTORIZATION AND SPECTRAL MASKS Emad M. Grais and Hakan Erdogan Faculty of Engineering and Natural Sciences, Sabanci University, Orhanli

More information

MANY digital speech communication applications, e.g.,

MANY digital speech communication applications, e.g., 406 IEEE TRANSACTIONS ON AUDIO, SPEECH, AND LANGUAGE PROCESSING, VOL. 15, NO. 2, FEBRUARY 2007 An MMSE Estimator for Speech Enhancement Under a Combined Stochastic Deterministic Speech Model Richard C.

More information

A Fast Algorithm for. Nonstationary Delay Estimation. H. C. So. Department of Electronic Engineering, City University of Hong Kong

A Fast Algorithm for. Nonstationary Delay Estimation. H. C. So. Department of Electronic Engineering, City University of Hong Kong A Fast Algorithm for Nonstationary Delay Estimation H. C. So Department of Electronic Engineering, City University of Hong Kong Tat Chee Avenue, Kowloon, Hong Kong Email : hcso@ee.cityu.edu.hk June 19,

More information

ACTIVE noise control (ANC) ([1], [2]) is an established

ACTIVE noise control (ANC) ([1], [2]) is an established 286 IEEE TRANSACTIONS ON SPEECH AND AUDIO PROCESSING, VOL. 13, NO. 2, MARCH 2005 Convergence Analysis of a Complex LMS Algorithm With Tonal Reference Signals Mrityunjoy Chakraborty, Senior Member, IEEE,

More information

Hybrid Time-Frequency Domain Adaptive Filtering Algorithm for control of the vibration control systems

Hybrid Time-Frequency Domain Adaptive Filtering Algorithm for control of the vibration control systems Hybrid Time-Frequency Domain Adaptive Filtering Algorithm for control of the vibration control systems Martino O. Ajangnay Member, IEE, student member, IEEE Matthew W. Dunnigan Member, IEE Barry W. William

More information

Wavelet Based Image Restoration Using Cross-Band Operators

Wavelet Based Image Restoration Using Cross-Band Operators 1 Wavelet Based Image Restoration Using Cross-Band Operators Erez Cohen Electrical Engineering Department Technion - Israel Institute of Technology Supervised by Prof. Israel Cohen 2 Layout Introduction

More information

Adaptive Stereo Acoustic Echo Cancelation in reverberant environments. Amos Schreibman

Adaptive Stereo Acoustic Echo Cancelation in reverberant environments. Amos Schreibman Adaptive Stereo Acoustic Echo Cancelation in reverberant environments Amos Schreibman Adaptive Stereo Acoustic Echo Cancelation in reverberant environments Research Thesis As Partial Fulfillment of the

More information

HST.582J / 6.555J / J Biomedical Signal and Image Processing Spring 2007

HST.582J / 6.555J / J Biomedical Signal and Image Processing Spring 2007 MIT OpenCourseWare http://ocw.mit.edu HST.582J / 6.555J / 16.456J Biomedical Signal and Image Processing Spring 2007 For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.

More information

On Information Maximization and Blind Signal Deconvolution

On Information Maximization and Blind Signal Deconvolution On Information Maximization and Blind Signal Deconvolution A Röbel Technical University of Berlin, Institute of Communication Sciences email: roebel@kgwtu-berlinde Abstract: In the following paper we investigate

More information

ANALYSIS OF STEADY-STATE EXCESS MEAN-SQUARE-ERROR OF THE LEAST MEAN KURTOSIS ADAPTIVE ALGORITHM

ANALYSIS OF STEADY-STATE EXCESS MEAN-SQUARE-ERROR OF THE LEAST MEAN KURTOSIS ADAPTIVE ALGORITHM ANALYSIS OF SEADY-SAE EXCESS MEAN-SQUARE-ERROR OF HE LEAS MEAN KUROSIS ADAPIVE ALGORIHM Junibakti Sanubari Department of Electronics Engineering, Satya Wacana University Jalan Diponegoro 5 60, Salatiga,

More information

Comparative Performance Analysis of Three Algorithms for Principal Component Analysis

Comparative Performance Analysis of Three Algorithms for Principal Component Analysis 84 R. LANDQVIST, A. MOHAMMED, COMPARATIVE PERFORMANCE ANALYSIS OF THR ALGORITHMS Comparative Performance Analysis of Three Algorithms for Principal Component Analysis Ronnie LANDQVIST, Abbas MOHAMMED Dept.

More information

Independent Component Analysis and Unsupervised Learning. Jen-Tzung Chien

Independent Component Analysis and Unsupervised Learning. Jen-Tzung Chien Independent Component Analysis and Unsupervised Learning Jen-Tzung Chien TABLE OF CONTENTS 1. Independent Component Analysis 2. Case Study I: Speech Recognition Independent voices Nonparametric likelihood

More information