AN INTERNATIONAL PUBLICATION
`
`CONTENTS
`pages 1-96
`
`ANTENNAS
`Accurate modelling of anti(cid:173)
`resonant dipole antennas
`using the method of moments
`D.H. Werner and
`R.J. Allard (USA)
`Dual-polarised uniplanar conical(cid:173)
`beam antennas for HIPERLAN
`E.M. Ibrahim, N.J. McEwan,
`R.A. Abd-Aihameed and
`P.S. Excell (United Kingdom)
`
`CIRCUIT THEORY & DESIGN
`Analogue CMOS high-frequency
`continuous wavelet transform
`circuit
`E.W. Justh and F.J. Kub (USA)
`Apparent power transducer for
`three-phase three-wire system
`S. KusuiandM.Kogane
`(Japan)
`Efficient and fast iterative reweighted
`least-squares nonrecursive filters
`Yue-Dar Jou, Chaur-Heh Hsieh
`and Chung-Ming Kuo (Taiwan)
`Input switch configuration
`suitable for rail-to-rail
`operation of switched
`opamp circuits
`M. Dessouky and
`A. Kaiser (France)
`Unified model of PWM switch
`including inductor in DCM
`Sung-Soo Hong (Korea)
`
`COMMUNICATIONS & SIGNAL
`PROCESSING
`Adaptive multiwavelet prefilter
`Yang Xinxing and
`Jiao Licheng (China)
`Decision feedback equalisation
`of coded 1-0 OPSK in mobile
`radio environments
`A. Adinoyi, S. AI-Semari
`and A. Zerguine (Saudi Arabia)
`Detection algorithm and initial
`laboratory results using
`V-BLAST space-time
`communication architecture
`G.D. Golden, C.J. Foschini,
`R.A. Valenzuela and
`P.W. Wolniansky (USA)
`
`page
`
`2
`
`4
`
`5
`
`7
`
`8
`
`10
`
`11
`
`13
`
`14
`
`7th January 1999 Vol. 35 No. 1
`
`Efficient complexity reduction
`technique in trellis decoding
`algorithm
`Sooyoung Kim Shin
`and Soo In Lee (Korea)
`Extended complex RBF and
`its application to M-OAM
`in presence of co-channel
`interference
`Ki Yong Lee and
`Souhwan Jung (Korea)
`Fair queueing algorithm with
`rate independent delay for
`ATM networks
`S. Ho, S. Chan and
`K.T. Ko (Hong Kong)
`Integrated space-time equaliser
`for DS/CDMA receiver with
`unequal reduced lengths
`V.D. Pham and T.B. Vu
`(Australia)
`Investigation of sensor failure
`with respect to ambiguities
`in linear arrays
`V. Lefkaditis and A. Manikas
`(United Kingdom)
`Learning algorithms for minimum
`cost, delay bounded multicast
`routing in dynamic environments
`J. Reeve, P. Mars and
`T. Hodgkinson (United Kingdom)
`Multiple target tracking using
`constrained MAP data association
`Hong Jeong and
`Jeong-Ho Park (Korea)
`Passband flattening and
`broader>lng techniques for
`high spectral efficiency
`wavelength demultiplexers
`E.G. Churin and
`P. Bayvel (United Kingdom)
`Performance of CDMA/PRMA
`protocol for Nakagami-m
`frequency selective fading
`channel
`R.P.F. Hoefel and
`C. de Almeida (Brazi{)
`Tbit/s switching scheme
`for ATM/WDM networks
`J. Nir, I. Elhanany
`and D. Sa dot (lsraen
`
`page
`
`16
`
`17
`
`19
`
`20
`
`22
`
`24
`
`25
`
`27
`
`28
`
`30
`
`(continued on back cover}
`
`Hughes, Exh. 1014, p. 1
`
`

`

`CONTENTS
`(continued from front cover)
`
`ELECTROMAGNETIC WAVES
`Electromagnetic penetration into
`20 multiple slotted rectangular
`cavity: TE-wave
`H.H. Park and H.J. Eom (Korea)
`
`IMAGE PROCESSING
`Encoding edge blocks by partial
`blocks of codevectors in vector
`quantisation
`Hui-Hsun Huang, Cheng-Wen Ko
`and Chien-Ping Wu (Taiwan)
`Technique for accurate correspondence
`estimation in object borders and
`occluded image regions
`E. Izquierdo M. (United Kingdom)
`
`INFORMATION THEORY
`Analysis of turbo codes with
`asymmetric modulation
`Young Min Choi and
`Pil Jeong Lee (Korea)
`Improved group signature
`scheme based on discrete
`logarithm problem
`Yuh-Min Tseng and
`Jinn-Ke Jan (Taiwan)
`Low density parity check codes with
`semi-random parity check matrix
`Li Ping, W.K. Leung (Hong Kong)
`and Nam Phamdo (USA)
`Non-binary convolutional
`codes for turbo coding
`C. Berrou and M. Jezequel (France)
`
`INTEGRATED OPTOELECTRONICS
`46GHz bandwidth monolithic
`lnP /lnGaAs pin/SHBT photoreceiver
`D. Huber, M. Bitter, T. Morf,
`C. Bergamaschi, H. Melchior
`and H. Jackel (Switzerland)
`
`LASERS
`1.5J.!m lnGaAIAs-strained MOW ridge-
`waveguide laser diodes with hot-
`carrier injection suppression structure
`H. Fukano, Y. Noguchi
`and S. Kondo (Japan)
`9.5W CW output power from high
`brightness 980 nm lnGaAs/ AlGa As
`tapered laser arrays
`F.J. Wilson, J.J. Lewandowski,
`B.K. Nayar, D.J. Robbins,
`P.J. Williams, N. Carr and
`F.O. Robson (United Kingdom)
`Investigation of data transmission
`characteristics of polarisation-
`controlled 850nm GaAs-based
`VCSELs grown on (311)8 substrates
`H. Uenohara, K. Tateno,
`T. Kagawa, Y. Ohiso,
`H. Tsuda, T. Kurokawa
`and C. Amano (Japan)
`Low current and highly reliable
`operation at 80°C of 650nm
`5mW LOs for DVD applications
`M. Ohya, H. Fujii,
`K. Doi and K. Endo (Japan)
`Modelocked distributed
`Bragg reflector laser
`H. Fan, N.K. Dutta, U. Koren,
`C. H. Chen and A.B. Piccirilli (USA)
`
`page
`
`31
`
`32
`
`34
`
`35
`
`37
`
`38
`
`39
`
`40
`
`41
`
`43
`
`45
`
`46
`
`48
`
`Near room-temperature continuous-
`wave operation of electrically
`pumped 1.55J.!m vertical cavity
`lasers with lnGaAsP/InP bottom
`mirror
`S. Rapp, F. Salomonsson,
`J. Sentell (Sweden),
`I. Sagnes, H. Moussa,
`C. Meriadec, R. Raj (France),
`K. Streubel and
`M. Hammar (Sweden)
`Record high characteristic
`temperature (7;, = 122K) of 1.55J.!m
`strain-compensated AIGalnAs/
`AIGalnAs MOW lasers with
`AlAs/ AllnAs multiquantum barrier
`N. Ohnoki, G. Okazaki,
`F. Koyama and K. lga (Japan)
`Red light generation by sum frequency
`mixing of Er/Yb fibre amplifier
`output in OPM LiNb03
`D.L. Hart, L. Goldberg
`and W.K. Burns (USA)
`
`MICROWAVE GUIDES &
`COMPONENTS
`Lumped DC-50GHz amplifier
`using lnP/InGaAs HBTs
`A. Huber, D. Huber,
`C. Bergamaschi, T. Morf
`and H. Jackel (Switzerland)
`RF tunable attenuator and modulator
`using high Tc superconducting filter
`Lu Jian, Tan Chin Yaw, C.K. Ong
`and Chew Siou Teck (Singapore)
`
`NEURAL NETWORKS
`Compact building blocks for
`artificial neural networks
`M. Melendez-Rodrfguez and
`J. Silva-Martinez (Mexico)
`
`OPTICAL COMMUNICATIONS
`40Gbit/s single channel dispersion
`managed pulse propagation in
`standard fibre over 509 km
`S.B. Alleston, P. Harper,
`I.S. Penketh, I. Bennion and
`N.J. Doran (United Kingdom)
`All-optical 2R regeneration based
`on interferometric structure
`incorporating semiconductor
`optical amplifiers
`D. 1/\(olfson, P.B. Hansen,
`A. K1och and K.E. Stubkjaer
`(Denmark)
`Demonstration of time interweaved
`photonic four-channel WDM
`sampler for hybrid analogue-
`digital converter
`J.U. Kang and R.D. Esman (USA)
`Desi.gn of short dispersion decreasing
`f1bre for enhanced compression of
`higher-order soliton pulses around
`1550nm
`M.D. Pelusi, Y. Matsui
`and A. Suzuki (Japan)
`Experimental measurement of
`group velocity dispersion in
`photonic crystal fibre
`M.J. Gander, R. McBride,
`J.D.C. Jones, D. Mogilevtsev
`T.A. Birks, J.C. Knight and
`'
`P.St.J. Russell (United Kingdom)
`
`page
`
`49
`
`51
`
`52
`
`53
`
`55
`
`56
`
`57
`
`59
`
`60
`
`61
`
`63
`
`(continued on inside back cover)
`
`Hughes, Exh. 1014, p. 2
`
`

`

`. --
`
`':t:--~.:-:"'"··~··.-..~.)0::·~::·-~:.-·-"·---=-·- ~~ ~;..-;--"'~""~ ·.:.- .~;v.,:.~,·~-~·· ~.=-.,..'":'~-.
`
`... ~.;.?;..::::;---·~~-__,.,.,.,.,..~ • .,. :--~-- r....,..~-..-4..~·· .._~ ~~.-~-"-""'"··:o:;-,..•-::;:-.::.---..-.l!':-•,-,-.~-- -~-·-·:.,~·;,_-.,,~~~· ~· r
`
`CONTENTS
`
`(continued from back cover)
`
`Experimental observation of soliton
`robustness to polarisation
`dispersion pulse broadening
`B. Bakhshi, J. Hansryd,
`P.A. Andrekson, J. Brentel,
`E. Kolltveit, B.-E. Olsson
`and M. Karlsson (Sweden)
`Frequency downconverter for high-
`capacity fibre grating based
`beamformers for phased arrays
`K.E. Alameh (Australia)
`Helical WDM ring network architecture
`H. Obara and K. Aida (Japan)
`Optimisation of dispersion-induced
`power penalty mitigation in
`millimetre-wave fibre optic links
`J.M. Fuster, J. Marti, J.L. Corral,
`V. Polo and F. Ramos (Spain)
`Polarisation-independent all-optical
`circulating shift register based
`on self-phase modulation of
`semiconductor optical amplifier
`Hyuek Jae Lee and
`Hae Geun Kim (Korea)
`Polarisation independent all-optical
`demultiplexing using four wave
`mixing in dispersion shifted fibre
`R. Calvani, F. Cisternino,
`R. Girardi and E. Riccardi (Italy)
`Scaling limitations in full-mesh WDM
`ring networks using arrayed-
`waveguide grating OADMs
`J.J.O. Pires (Portugan.
`M. O'Mahony, N. Parnis
`and E. Jones (United Kingdom)
`Single-wavelength 40Gbit/s soliton
`field transmission experiment
`over 400km of installed fibre
`E. Kolltveit, P.A. Andrekson,
`J. Brentel, B.E. Olsson,
`B. Bakhshi, J. Hansryd,
`P.O. Hedekvist, M. Karlsson,
`H. Sunnerud and J. Li (Sweden)
`Ultra-high speed soliton transmission
`in presence of polarisation mode
`dispersion using in-line
`synchronous modulation
`A. Sahara, H. Kubota
`and M. Nakazawa (Japan)
`
`OPTICAL FIBRES & SENSORS
`Injection moulded plastic
`ferrules for singlemode
`optical fibre connections
`S. Yanagi, H. Sato, Y. Shuto,
`M. Ohno and S. Tohno (Japan)
`
`page
`
`65
`
`66
`
`67
`
`69
`
`70
`
`72
`
`73
`
`75
`
`76
`
`78
`
`Low-cost microlens array for
`long-period grating fabrication
`S.Y. Liu, H.Y. Tam and
`M.S. Demokan (Hong Kong)
`Widely tunable long-
`period fibre gratings
`A.A. Abramov, A. Hale,
`R.S. Windeler and
`T.A. Strasser (USA)
`
`OPTOELECTRONICS
`Large-signal compression-current
`measurements in high-power
`microwave pin photodiodes
`K.J . Williams and
`R.D. Esman (USA)
`Numerical solution of time-
`dependent coupled-wave
`equations using split-step
`algorithm
`Byoung-Sung Kim and
`Youngchul Chung (Korea)
`
`SEMICONDUCTOR DEVICES &
`MATERIALS
`High-fT n-MODFETs fabricated
`on Si/SiGe heterostructures
`grown by UHV-CVD
`S.J. Koester, J.O. Chu
`and R.A. Groves (USA)
`
`ULTRASONICS
`Algorithm for robot position
`tracking using ultrasonics
`W.T. Kuang and A.S. Morris
`(United Kingdom)
`Higher-order time-varying allpass
`filters for signal decorrelation
`in stereophonic acoustic echo
`cancellation
`N. Tangsangiumvisai,
`J.A. Chambers and
`A.G. Constantinides
`(United Kingdom)
`Monitoring of crystallisation
`phenomena by ultrasound
`J.S. Tebbutt, T. Marshall and
`R.E. Challis (United Kingdom)
`
`Errata
`
`page
`79
`
`81
`
`82
`
`84
`
`86
`
`87
`
`88
`
`90
`
`91
`
`Hughes, Exh. 1014, p. 3
`
`

`

`ELECTRONICS
`
`T H E
`
`INSTITUTION
`
`0 F ELECTRICAL
`
`ENGINEERS
`
`7 JANUARY 1999
`ELLEAK 35 (1)
`
`VOLUME35
`1 -96
`
`NUMBER1
`ISSN 0013-5194
`
`ELECTRONICS LETTERS (ISSN 0013-5194) is published every other week except for one issue in December (total of 25 issues). 1999 annual
`subscription price £630.00. Single copy £26.00. Available also in a concurrent microfiche edition at the same subscription rate as for the printed
`edition. Air mail service direct to subscnbers an additional £62.00.
`
`All subscnption inqu1ries and orders should be sent to lEE Publication Sales, PO Box 96, Stevenage SG1 2SD, United Kingdom.
`Contributions should be in accordance with the Guide to Authors, which is usually to be found on the inside back cover or the last page of each issue,
`and should be addressed as indicated in the Guide.
`All other correspondence, including advertisements, should be sent to the lEE Publishing Department.
`
`The advantages of rapid publication given by Electronics Letters are nullified if subscribers are subject to excessive delays in the receipt of the
`JOurnal. The Executive Editor would be grateful to learn of any difficulties subscribers may experience.
`Electronics Letters Online
`
`Electronics Letters is available online through the lEE Online Journals service on the lEE's World-Wide Web server. Articles are stored in PDF format
`and may be viewed and printed using the Adobe® Acrobat® Reader. Features include: browsable table of contents pages for each journal issue,
`cross-journal searching of bibliographic records (inc. authors, titles, abstracts). and the ability to print articles locally. The issues are added in
`advance of print publication. For further information, please look at http://ioj.iee.org.uk/ or contact the Marketing Department at Stevenage.
`Copyright and copying
`
`This publication is copyright under the Berne Convention and the Universal Copyright Convention. All rights reserved. Apart from any copying under
`the UK Copyright, Designs and Patents Act 1988, Part 1, Section 38, whereby a single copy of an article may be supplied, under certain conditions,
`for the purposes of research or private study, by a library of a class prescribed by The Copyright (Librarians and Archivists) (Copying of Copyright
`Material) Regulations 1989: Sl 1989/1212, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by
`any means without the prior permission of the copyright owners. Permission is, however, not required to copy abstracts of individual contributions on
`condition that a full reference to the source is shown.
`Single copies may be made for the purpose of research or private study.
`
`Authorisation to photocopy items for internal or personal use, or for the internal or personal use of specific clients, is granted by the Institution of
`Electrical Eng1neers for libraries and other users registered with the Copyright Clearance Center Transactional Reporting Service, provided that the
`base fee of $10.00 per copy is paid directly to the Copyright Clearance Center, Inc., 21 Congress Street, Salem, MA 01970, USA. 00135194/97
`$10.00. This authorisation does not extend to other kinds of copying, such as copying for general distribution, for advertising or promotional
`purposes, for creating new collective works or for resale.
`Multiple copying of the content of this publication without permission is always illegal.
`The lEE is not as a body responsible for the opinions expressed by individual authors in Electronics Letters.
`The lEE is a member of the Assoc1allon of Learned & Professional Society Publishers.
`
`The Institution of Electrical Eng1neers
`Savoy Place
`London WC2R OBL
`United Kingdom
`
`Telephone
`
`Facsimile
`
`t44 (0)1 71 2401871
`
`t44 (0)1 71 240 7735
`
`http://www.iee.org.uk
`
`Publishing Department:
`lEE
`Michael Faraday House
`Six Hills Way
`Stevenage SG1 2AY
`United Kingdom
`Telephone
`Telex
`Facsimile:
`Sales
`Editorial
`Advertising
`Marketing
`E-mail:
`
`t44 (0) 1438 313311
`825578 IEES1V G
`
`+44 (0) 1438 742792
`+44 (OJ 1438 742849
`+44 (0) 1438 318361
`+44 (0) 1438 742840
`
`Secretary of the lEE:
`J. C. Williams, OBE, PhD, FEng, FlEE
`
`© 1999: The Institution of Electrical
`Engineers
`
`Editorial staff:
`Gill Wheeler (Managing Edrtor)
`Joanne Davies
`Rachel Smyly
`
`HONORARY EDITORS
`Sir Eric Ash, CBE, FEng, FRS
`Prof. Peter Clarricoats, CBE, FEng, FRS
`Prof. Chris Toumazou
`
`Typeset in England by
`
`Pindar pic
`Ryedale Building
`First Roor
`60 Piccadilly
`YorkY019WU
`
`Printed m UK by
`A. Mclay & Co. Umited
`Longwood Drive
`Forest Farm
`Cardiff CF4 7ZB
`
`Hughes, Exh. 1014, p. 4
`
`

`

`References
`
`2
`
`'A \'LSI tmplememation of
`EDW\RDS. R.T. and t \l \\l~lli·RGIIS. <;
`the continuous \\avclct transform·. Proc. IEEE Int. Symp. Circuits
`Systems, 1996. Vol. 4. pp. :168 3 71
`'Time-domain
`\IOREIR ,\-T,\1\.iAYO. 0.. and Pl~rr>A or GY\E7 J:
`analog wavelet tran,form 111 real-time·. Proc. IEEF lnt Symp.
`Circuits Systems. 1995. Vol. 3. pp. 1640 1643
`'Analog VLSI
`r·D\\,\RDS T, and SIIA\1\.f\ s.:
`U'< J .•
`Io;l. \\-H.
`implementations of audllory wavelet transforms usmg switched(cid:173)
`capacitor circuits'. IEEE 1i·wl.l. Circuits Syst I 1994, 41. (9). pp.
`572 583
`4 EDW \RD . R.T. and GOD I RI·Y. \I D
`• \n analog wavelet transform
`chip'. Proc. IEEE Int. C'onf
`'e ural
`etworks. 1993. Vol. \ pp.
`1247 1251
`t.FSHR. G.: ·A radar waveform processmg utili7ation of \\avekts·.
`Proc.
`IEEE-SP
`lnt Symp. Time-Frequency and Time-Scale
`Analysis. 1994. pp 480 ~li3
`ignal characteristics u ing the
`6 SADOWSK \, J.'
`'lmesllgation of
`contmuous wavelet transfom1'. Johm lfopkim A PL Tech Digest.
`1996. 17. (3). pp. 258 269
`110. K C. PROKOP!W, w. and UIA .. Y T: ·Modulation identificatiOn
`by the wavelet transform· Proc. IEEE Military Communications
`Conlerence (MIL-COM). 1995, Vol. 2. pp. 886 890
`8 M \RTIN, r L.:
`·A BiCMOS 50 M Hz voltage-controlled oscillator
`IEEE Custom Integrated Circuits
`quadrature output ·. Proc
`Conference. 1993. pp. 27.4.1 27.4.4
`Ill CHWALD. A.W. and \I \R II • K.W .. 'H1gh-speed \Oitagc colllrolled
`oscillator with quadrature outputs'. Electron Lei/. 1991. 27. {4).
`pp. 309 310
`10 Kl B. I'.J .. and JUSTH. 1 w: 'Analog CMOS implcmemation of high
`frequency least-mean square error learning circuit'. IEEE 1 Solid(cid:173)
`State Circuits. 1995. 30. ( 12). pp. 1391 1398
`
`5
`
`7
`
`9
`
`capacitor values m the synchronous receiver lowpass filter. a true
`wavelet dilation could be implemented. The half-bandwidths of
`the bandpass filters and thetr centre frequency spacings are about
`650kHz.
`Fig. 3 bows the time-frequenc) plot for a (slowly) frequency(cid:173)
`swept input signal. In Fig. 3. where black represents the peak out(cid:173)
`put voltage for each channel. there is good frequency descrimina(cid:173)
`tion as the input frequency sweeps over the frequency range of the
`filter bank. Fig. 4 shows the time-domain response of a single
`O.SJ..UTI CMOS channel to a IOOmVw 300ns burst at 47MHz.
`
`a
`
`b
`
`Fig. 4 Time-domain response o.fsingle 0.5J1m C.\lOS channel
`a Output . tgnal (200m V 'di\')
`h Input
`ignal (50 111 V 'd iv)
`
`The order of the lowpass filter in the synchronous reccl'o;er
`detennines the bandpass filter characteristic. A single-pole lowpass
`filter design is easily converted to a two-pole design by adding a
`capacitor acros
`the diJTerential outputs of the multiplier that
`dri\'e
`the lowpass filter. The two-pole lowpass ftlter design ha
`much steeper skirts, and hence better separation of signals with
`different frequencies. In Fig. 3. the bandpass ftlter ·hape corre(cid:173)
`sponds to the two-pole lowpass filter de ign.
`Since increasing the number of channels can be u ed to increase
`system perfonnance, size and power di ·sipation constraints arc
`important for a continuous wavelet transform circuit. For ti1e 2J..U11
`de ign, the channel were laid out on a I SO~tm pitch with under
`IOOmW power dissipation per channel and a maximum operatmg
`frequency of SOMHz. For the O.SJ..Ull design. the channels were
`laid out on a S6~m1 pitch with under 40mW power dissipation per
`channel and a maximum operating frequency
`in exce. s of
`IOOMHz. The total size of the 16-channel 2J..U11 chip was 4750J..Un
`by 31 OOm.
`The VCO design is particularly critical, since the VCO needs to
`have a constant, frequency-independent output \ oltage. and also
`needs to be tunable over as large a frequency range a possible.
`Our VCO design uses diode
`to set the output amplitude. triode
`MOSFET resistors to change the oscillation frequency, and bias
`current adjustment slaved to the lliode resistor setting to compen(cid:173)
`sate for ti1e change in loop gain associated with changing the tri(cid:173)
`ode resistor \'alues.
`To prevent drifts in the VCO frequencies with changes in tem(cid:173)
`perature. phase-locked loops and external frequency references can
`be used. as shown in Fig. I. Where the VCO voltage-frequency
`charactetistic is linear. many VCO can be bia ed using a pair of
`PLLs along with a re istive voltage di,ider. The PLLs have been
`succe ·sfull) implemented. but not yet combmed with the continu(cid:173)
`ous wavelet transfom1 circuit. We ba\'e demonstrated a 0.5~1
`CMOS PLL ttmable from 64MHz to 77MHz.
`
`AcknVII'Ie£/~lnenls: This work was supported by the Onice of
`aval Research. The authors would also like to thank Ton)
`Teoli for valuable discussions.
`
`© I EE 1999
`Electronics Lei/en Online So. /9990077
`E.W. Justh and F.J. Kub (,\'aral Rneach Lahoratory, Code 6813, 4555
`SA)
`01'erlook ..ll'e. Sll'. Washington. DC :!0375.
`E-mail. kub!a estd nrl.na\) .mil
`
`20 i\oremher 1998
`
`Apparent power transducer for three-phase
`three-wire system
`
`S. Kusui and M. Koganc
`
`In a newly devdopt.'tl transducer apparent power is measured by
`using a multiplier. in which the AC component of the output Is
`used. l-or a three-phase three·\\ire system. two multipliei> are
`used according to the so-called 'two wattmeter' method. The AC
`output compo'Oents are + 30° and -30 phase-shifted, rcspt.'Clt\el).
`and then the diiTcnmce 1s converted to a DC signal v.-htch
`corresponds to the total appan:nt power.
`
`Imroduc1ion: Measurement of elect1ical apparent power is some(cid:173)
`times necessary in order to take mto account the power factor as
`well as the energy m in the case of large consumers. Vanous
`apparent power meters have been developed [I
`6] u~mg such
`methods a multiplication of the RMS values of the voltage and
`the current. or taking the root of the squared sum of the active
`and reactive powers. An apparent power meter which directly use
`the above defmition is complicated because RMS ACDC com·ert(cid:173)
`ers or a reactive power meter and computer are needed.
`The authors have noticed that the amplitude of the AC compo(cid:173)
`nent of the multiplication of the insantaneous voltage and current
`equals the RMS volt-ampere which ts the apparent power. Thi ·
`idea is applied to an apparent power transducer for the three(cid:173)
`phase three-wire system \\hich is very popular in Japan. The con(cid:173)
`\'entional method needs four multipliers (two for the active power
`transducer and t1-vo for the reactive power transducer) and a calcu(cid:173)
`lator to obtain the root of the squared . um of the active and reac(cid:173)
`tive power . However, the new method needs only two multipliers.
`Furthermore. tf necessary, the active power is easily measured at
`the same time usmg ti1e a me multiplier·. Therefore the contigura(cid:173)
`tion is very . imple and the cost is ver) It)\.\.
`
`Principle and configura/ion: Fig. I show
`the conH~ntional appar(cid:173)
`ent power transducer for a three-phase three-wire system. I. 2 and
`3 arc powerlincs: Lis the load . .11,.1 and ,\In an.: the active power
`transducer whose outputs P, and P2 are ·ummed b) L,. to obtain
`the total power signal P . .\IQ1 and ,\[Q~ are the reactive power
`tran ducers whose outputs Q, and Q- are swnmcd by LQ to obtain
`
`ELECTRONICS LETTERS 7th Januar y 1999 Vol. 35 No. 1
`
`5
`
`Hughes, Exh. 1014, p. 5
`
`

`

`Sccuritl' anah:m: Some possible c~ttack agamst the proposed
`schcm~ are p~escnted bckm
`Attack 1: Although the group authority has th\! knowledge of (r.
`s .. k,). the group signature cannot be forged without the s\!cret k\!~
`x of L'. It is impossible to obtain x, from 1', \\ithout being able to
`solve the discrete logarithm problem. \.1oreover. because the g\!n(cid:173)
`erator a, ::::: gu mod p. k, E z~· and a E z .. . both a and g have
`the same order tf. Therefor\!, forging (R, S) is as difficult as break(cid:173)
`ing ElGamal's chem\! [3]. Since th\! group authonty cannot forge
`the group signawre, forger) by an adversary is even more diffi(cid:173)
`cult. Thus. the impersonation attack can not be successful.
`A tlack 2: The signer {' can be identified if we can obtain y from
`the signature {R, S. h(m). A. B. C. D. £}. Since the receiver does
`not know the (r. s,. /..,)of the group authority. he cannot check the
`equation D8 • .r: · E =I> mod p. Obtaining (r. ~ .. k,) from given
`1 A. B. C. D. £} depends on the discrete logarithm.
`Auack 3: The group authority ma) publish the inf01mation (r,, s1•
`J ,) for the me sage m's s1gnature to enable a verifier to check th_e
`identit) of u~. Thi. does not damage the anonymit) of U,'s pren(cid:173)
`ous group signature because the information (r,. s1 • . 1,) is only
`provided for the 'JX.'Cific group signature { R. S. li(m). 1. B. C. D.
`£}. For different m\!ssages, u will have chosen dilferent random
`ignatures. If an adversat')
`integers a and h to generate group
`wants to obtain a. h aml (r., s,) from given l A. B. C. D. £} ,this IS
`as difficult as solving the d1screte logarithm.
`
`Discussion: The improved scheme preserves the main ments mher(cid:173)
`ent in most of the Lee-Chang scheme. In the case of a later di -
`pute, the group authorit) may publish the infonnation (r 1 , s1, y,)
`to enable a verifier to check the identity of the signer, although
`thi doe not damage the anonymity of the othi!r previous signa(cid:173)
`tures of the signer Meanwhile. the group amhmity need not
`renew any key of the
`igner. The reason is that the information
`(r 1 • s r· y) is only provided for the pecific group signature { R. S,
`h(m), A, B, C. D, £}. Compared to the original scheme, th\!
`impro,·ed cheme requires some additional cost in ie1ms of com(cid:173)
`putational time and the
`ize of the group signature. For generating
`a group signature. the signer L may precompute several different
`1 a,. A. B, C. D. Ef usmg (r,. s.) to reduce the real-time computa(cid:173)
`tional time.
`
`Conc!u1iom: Vve hcl\·e proposed an improved group signature
`scheme based on the discrete logarithm. In our impro,·ement, a
`group signature can be opened to reveal the identity of the signer.
`the anonym it) of the other previous signatures signed by thi ·
`group member arc not damaged. Meanwhile. the group authorit)
`also need not renew the keys of the signer. We have demonstrated
`·ome pos ible attacks against the proposed cheme. Undl!r the dif(cid:173)
`ficulty of computing th\! discr\!le logarithm problem. we have
`shm\ n that the improved scheme IS secure against these attacks.
`
`© lEE 1999
`Elcctromcs Letter., Online Vo. 19990071
`Yuh-:\lin Tseng and .Jinn-Ke Jan (Institute ol Applied \fathemann.
`'atiunal Chung H.1ing l..nircrsitl'. Taidumg, Tailran 402. Repuhlic 11(
`China)
`
`28 October JC)98
`
`Jinn-Ke Jan: corresponding author
`
`!.:-mail: jk.janraamath nchu.edu.t\\
`
`References
`
`Signatures·.
`
`Pwc
`
`'2
`
`3
`
`4
`
`c 11 \l \f. D..
`'Group
`and
`HI \'St. 1 :
`E( ROCRl'PT'91. 1992. pp. '257 '265
`<:111.. . 1 . and 1'1 Dl.RSI ~- 1 P: ·New group stgnatur~ schemes·. Proc
`£CROCRYPT'94. 1995. pp. 171 Jl)t
`r L.CiA~I:\1., T
`'A puhlic kc) Ct)pto syswm and a s1gnaturc scheme
`based on discrete logarithms·. lEEr; Trans. Jnf Theon. 1985, 31
`.
`(4), pp. 469 47'2
`11.1, \\.B. and Cll·\ c,, cc 'Efficient group signature scheme bast:d
`on the discrete logarithm, IE/, Proc Comput. D~-<it Tech .. l99X.
`145. (I). pp. 15 18
`5 '\ BfRG. K, and Rl I PI'~ L. RA. '~1~ssage recover~ for signature
`schemes bast:J on the dJscrctc loganthm prohlcm', Desigm, Cot/e.\
`and Cnp/o.f!.raphr. 1996, 7. pp. 61 ~I
`
`Low density parity check codes with semi(cid:173)
`random parity check matrix
`
`Li Ping. W.K. Leung and
`
`am Phamdo
`
`A semi-random approach to low dt:!nsJt) parit) check code dt-sign
`the same perfom1ance as an
`is shown to achieve essentiall)
`extsting method. but mth considerably reduced complexity.
`
`Introduction: Recently. there has been revived interest in the low
`density paril) check (LDPC) codes originally introduced in 1962
`by Gallager ( 1]. ll has been shown that such codes can achieve
`very good performances (>vithin J .5dB of theoretical limits) with
`modest decoding complexil) [2].
`An LDPC code is defined from a randomly generated parity
`check matrix H [2]. For the purpose of encoding, it is nece · ary to
`tran fer H into H . the equivalent systematic fom1 of H. whtch
`can be accomplished by Gau sian elimination. For a rate R ::::: kn
`(k ::::: information length, 11 ::::: coded length). the size of H is (n k) x
`n. When 11 is large. Gaus ian elimination can be costly in tenns of
`both memor) and the operations involved. Besides, a considerable
`amount of memory is required to store H ,,. in the encoder, which
`is not necessarily parse even though H is usually designed so.
`In this Letter, we report a modified approach to LDPC code
`design. We adopt a semi-random technique, i.e. only part of H is
`generated randomly, and the remaining part is deterministic. The
`new method can achieve es enttally the same performance as the
`standard LDPC encoding method with significantly reduel!d com(cid:173)
`plexity.
`
`Proposed approach: For simplicity we will only consider binary
`codes. Decompose the codeword c as c = (p, d]', where p and d
`contain the parity and information bits. respectively. Accordingly.
`we decompose H into H ::::: [H•. W]. Then
`
`(1)
`
`In the proposed method. H• is con tructed in some determini tic
`form. Empirically, we found the following a good chOice (recall
`that H• must be a quare matrix [3]):
`
`1
`
`1
`
`(2)
`
`We adopt the following rule to create Hd. Lett be a preset integer
`constrained by (i) 1 di,idcs n-k and (ii) n-k divide kt. Partition Hd
`(which has n k rows) into 1 equal sub-blocks as
`
`Hd = (H:dl)
`
`Hdt
`
`(3)
`
`In each sub-block Hd ', i::::: I. 2 .. ·t, we random!] create exactly
`one element l per column and kt/(11-k.) ls per row. The partition m
`eqn. 3 is to best increase the recurrence distance of each bit in the
`encoding chain (see below) and, intuitively. reduces the correlation
`durmg the decoding process. The resultant Hd has a column
`we1ght of 1 and a row weight of kt (n k) (the weight of a vector IS
`the numb\!r of l s among its elements).
`. Based on eqns. I and 2. p ::::: ip 1 can easily be calculated from a
`gn·en d - {d} as
`,.,
`
`:uu1
`
`c u1od :n
`
`( l)
`
`Compared with the standard LDPC code design [2]. the abo\'e
`m~thod has several advantages. First, the encoding process in eqn.
`4 IS much simpler than a full Gaussian elimination. Secondly. a
`random _l-Jd can be singular. which causes additional programmit:g
`comple~tty in realising a pccJfied rate. On the other hand. Hr m
`e~n. 2 IS always non-sin},rular so the new method can realise ~ill)
`gtven rate directly and precisely. Thirdlv, it requires very ltttle
`memory to store W in the encoder if Hd is spar e (this can be
`ensured using small /).
`
`38
`
`ELECTRONICS LETTERS 7th January 7999 Vol. 35 No. 1
`
`Hughes, Exh. 1014, p. 6
`
`

`

`·1
`10
`
`-2
`10
`
`-3
`10
`
`a:
`w
`-4
`aJ 10
`
`-5
`10
`
`-6
`10
`
`decoders passing the result of their work to each other. at each
`iterative step.
`
`X
`systematic
`part
`
`-7
`10 0.6
`
`1.0
`
`1.4
`EJN0,dB
`
`1.8
`
`2.2
`
`951 11
`
`Fig. I Per{onnmzces of LDPC code.\ generated hy .1emi-random parill'
`30000
`check matrixes 11·ith k
`+ R - 1/3
`• R - 112
`.A R
`'113
`
`Simulation S/11((1': Fig. 1 contains the simulated performances of
`the proposed encoding method for various rates (1 /3. 112. 213)
`using 1 = 4. The decoding algorithm follows that in [2). The results
`are e entially the same as those obtained using fully random H.
`
`Conclusion: It has been shown that a ·emi-random approach to
`LDPC code design can achieve essentially the same performance
`as the existing method with considerably reduced complexity.
`
`::?3 ,'\'oremher /998
`
`© lEE 1999
`Electronics Lellers Online Vo · 19990065
`Li Ping anti W.K. Leung (Depaumenr of Elecrronic Engineering, Cay
`Unil'(?rsii.J' of Hong Kong. Hong Kong)
`E-mail; eeliping(cLcityu.etlu.hk
`am Phamtlo (Department of Electrim/ am/ Complller EnginC'ering.
`f(I(C' L ni1·ersiry of .\'e11· l'ork ar S10ny Brook. Srmn· Brook . . \' }' !1794-
`::?350, LSA)
`.
`
`References
`
`GALLAGER. R G.: 'Low density parity check codes·. IRE Trans. In/
`Theory, 1962. IT -8. pp. 21 28
`2 Mad~". DJ c. and -...EAL. R. 1: ·Near Shannon lim1t performance of
`low density parity check codes'. Elecrron. Lell .. 1997. 33. (6). pp.
`457 458
`3 PRO \KIS. J G; 'Digital communications· (McGraw-Hill. 1995)
`I'Ll ERso . v. \\ . and 11 H.oo-.... 1 .1. Jr: 'Error-correcting codes· (1\cllT
`4
`Press. Cambridge. Massachusetts, 1972) 2nd etln.
`
`Non-binary convolutional codes for turbo
`coding
`
`C. Berrou and M. Jezequel
`
`The author~ cons1der the usc of non-binary convolutional codes in
`turbo coding.
`It
`is shown
`that quatemary codes can be
`advantageous.
`both
`from
`p.:rformance
`and
`complexity
`standpoints. but that higher-order codes ma} not bring further
`imprO\cment.
`
`Introduction. Turbo codes are error correcting codes with at least
`two dimensions (i.e. each datum is encoded at least twice). The
`decoding of turbo codes is based on an iterative procedure using
`the concept of extrinsic information. Fig. I gives an example of n
`two-dimensional turbo code built from a parallel concatenation of
`two identical recursi1e systematic convolutwnal (RSC) code 1\ith
`generators 15. 13 (octal notation). The global (non-iterative)
`decoding of such a code is too complex to be emisaged bc'Cause of
`the ver) large number of states induced by the interleaver. An iter(cid:173)
`ative procedure is therefore used, the two codes being decoded
`altcrnatcl)
`in
`theu· own dimenstons and the two associated
`
`i9551J
`Fig. l Tll'o-dime/1.\ivn turho code 1ritlr generawrs 15. J 3
`
`Binary codes \'f'rsus quatemw:1· codes: Fig. 2a represents a block of
`size k encoded by the code of Fig. I. This block i · een as a two(cid:173)
`dimensional ..Jk x "1/k block and for simplicity 1ve consider that the
`i