U)
`
`OFDM
`
`FOR WIRELESS
`
`MULTIMEDIA
`
`COMMUNICATIONS
`
`RICHARD VAN NEE
`
`RAMJEE PRASAD
`
`2 O I
`
`.—
`
`< 0 Z 3 2 Z O U _
`
`I < Z O U
`
`
`
`QERRRRBSLa
`399351
`Q99Ln
`
`
`)
`
`n:
`
`LLI
`
`D.
`
`..I
`4;
`U)
`
`0:
`
`LI]
`
`2 Z 3
`
`EX. 1017
`Ericsson v. Intellectual Ventures
`Page 1 of 24
`
`

`

`.K.
`
`
`
`OFDM Wireless Multimedia
`'
`Communications
`
`Richard van Nee
`Ramjee Prasad
`
`
`
`
`
`EH
`
`
`
`
`
`
`Artech House
`
`Boston 0 London
`
`EX. 1017
`Ericsson v. Intellectual Ventures
`Page 2 of 24
`
`

`

`
`
`
`
`
`To my W
`
`(1}
`
`
`
`TK5103.2.N44 2000
`621/3845—-dc21
`
`99052312
`' CIP
`
`I.
`
`British Library Cataloguing in Publication Data
`Nee, Richard van
`OFDM Wireless multimedia communications. —— (Artech House
`universal personal communications library)
`LWireless communications systems 2.. Multimedia systems
`I. Title II. Prasad, Ramjee
`6213,82
`
`ISBN 0-89006—530—6
`
`Cover design by Igor Vladrnan
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`© 2000 Richatd van Nee and Ramjee Prasad
`
`All rights reserved. Printed and bound in the United States ofAmerica. No part ofthis book
`
`.
`
`i
`
`may be reproduced or utilized in any form or by any means, electronic or mechanical, includ—
`ing photocopying, recording, or by any information storage and retrieval system, Without per—
`mission in writing from the authors.
`All terms mentioned in this book that are known to be trademarks or service marks have
`
`been appropriately capitalized. Artech House cannot attest to the accuracy of this informa—
`tion. Use of a term in this book should not be regarded as affecting the validity of any trade—
`mark or service mark.
`
`International Standard Book'Number: 0—89006-530—6
`Library of Congress Catalog Card Number: 99—052312
`
`10987654321
`
`
`Library of Congress Cataloging—in—Publication Data
`Nee, Richard van.
`OFDM for wireless multimedia communications / Richard van Nee, Ramjee Prasad
`p. cm. — (Artech House universal personal communications library)
`Includes bibliographical references and index.
`ISBN 0—89006-530—6 (alk. paper)
`1. Wireless communications systems. 2. Multimedia systems. 3. Multiplexing.
`Prasad, Ramjee. II. Title III. Series
`
`
`
`
`
`EX. 1017
`Ericsson v. Intellectual Ventures
`Page 3 of 24
`
`

`

`
`
`Contents
`
`Preface
`
`Acknowledgments
`
`Chapter 1
`
`Introduction
`
`1.1
`
`1.2
`
`Standardization and Frequency Bands
`
`Multimedia Communications
`
`1.2.1 The Need for High Data Rates
`
`1.2.2
`
`Services and Applications
`
`1.2.3 Antennas and Batteries
`
`1.2.4 Safety Considerations
`
`1.2.5 ATM-Based Wireless (Mobile) Broadband
`Multimedia Systems
`
`1.3
`
`Multipath Propagation
`
`1.3.1 Multipath Channel Models
`
`1.3.2 Delay Spread Values
`
`Time Variation of the Channel
`
`History of OFDM
`
`Preview of the Book
`
`1.4
`
`1.5
`
`1.6
`
`References
`
`Chapter 2
`
`OFDM Basics
`
`2.1
`
`2.2
`
`2.3
`
`2.4
`
`2.5
`
`2.6
`
`2.7
`
`Introduction
`
`Generation of Subcarriers using the [EFT
`
`Guard Time and Cyclic Extension
`
`Windowing
`
`Choice of OFDM Parameters
`
`OFDM Signal Processing
`
`Implementation Complexity of OFDM Versus
`Single Carrier Modulation
`
`vii
`
`xiii
`
`xvii
`
`1
`
`4
`
`7
`
`8
`
`9
`
`9
`
`10
`
`12
`
`15
`
`16
`
`17
`
`19
`
`20
`
`24
`
`25
`
`33
`
`33
`
`33
`
`39
`
`42
`
`46
`
`47
`
`48
`
`Page 4 of 24
`
`EX. 1017
`Ericsson v. Intellectual Ventures
`Page 4 of 24
`
`

`

`
`
`viii
`
`References
`
`5 1
`
`
`
`Coding and Modulation
`Introduction
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Coherent and Differential Detection
`95
`Chapter 5
`5.1
`Introduction
`
`
` Coherent Detection
`5.2
`
`
`
`Chapter 4
`
`Synchronization
`
`4. 1
`
`4.2
`
`4.3
`
`4.4
`
`4.5
`
`4.6
`
`4.7
`
`References
`
`Introduction
`
`Sensitivity to Phase Noise
`
`Sensitivity to Frequency Offset
`
`Sensitivity to Timing Errors
`
`Synchronization using the Cyclic Extension
`
`.
`
`Synchronization using Special Training Symbols
`
`Optimum Timing in the Presence of Multipath
`
`Chapter 3
`3. 1
`
`3.2
`
`Forward Error Correction Coding
`
`I
`
`3.2.1 Block Codes
`
`3.2.2 Convolutional Codes
`
`3.2.3 Concatenated Codes
`
`Interleaving
`
`Quadrature Amplitude Modulation
`
`Coded Modulation
`
`3.3
`
`3.4
`
`3.5
`
`References
`
`53
`53
`
`54
`
`54
`
`55
`
`58
`
`59
`
`60
`
`62
`
`70
`
`73
`
`73
`
`74
`
`77
`
`78
`
`80
`
`86
`
`88
`
`92
`
`
`
` 5.2.1 Two Dimensional Channel Estimators
`
`
`5.2.3 Special Training Symbols
`104
` 5.2.4 Decision Directed Channel Estimation
` Differential Detection
`
`5.3. 1
` Differential Detection in the Time Domain
`
`5.2.2 One Dimensional Channel Estimators
`
`103
`
`EX. 1017
`Ericsson v. Intellectual Ventures
`Page 5 of 24
`
`

`

`—_Tf
`.
`
`li
`
`i
`
`F—_.________——
`
`J
`
`5.3.2 Differential Detection in the Frequency Domain
`
`51
`
`53
`
`53
`
`54
`54
`
`55
`
`58
`
`59
`
`60
`
`62
`
`70
`
`73
`
`73
`
`74
`77
`78
`80
`
`86
`
`83
`
`92
`
`95
`
`95
`95
`96
`103
`104
`106
`107
`107
`
`.
`
`.
`J
`
`J
`
`i
`
`I
`
`,
`
`3
`
`J
`
`J
`1
`J
`
`1
`
`
`
`
`i
`
`J
`I
`‘
`J
`1
`i
`J
`J
`
`5.3.3 Differential Amplitude and Phase Shift Keying
`
`References
`
`Chapter 6
`6.1
`
`The Peak Power Problem
`Introduction
`
`6.2
`
`6.3
`
`6.4
`
`6.5
`
`Distribution of the Peak—to—Average Power Ratio
`
`,
`
`Clipping and Peak Windowing
`
`6.3.1 Required Backoff with a Non—Ideal Power Amplifier
`
`6.3.2 Coding and Scrambling
`
`Peak Cancellation
`
`PAP Reduction Codes
`
`6.5.1 Generating Complementary Codes
`
`.
`
`6.5.2 Minimum Distance of Complementary Codes
`
`6.5.3 Maximum Likelihood Decoding of Complementary Codes
`6.5.4 Suboptimum Decoding of Complementary Codes
`6.5.5 Large Code Lengths
`SYMBOL Scrambling
`
`145
`147
`150
`150V
`
`6.6
`
`References
`
`Chapter 7
`
`Basics of CDMA
`
`7.1
`
`7.2
`
`7.3
`
`7.4
`
`Introduction
`
`CDMA: Past, Present, and Future
`
`CDMA Concepts
`7.3.1 Pure CDMA
`Basic DS-CDMA Elements
`7.4.1 RAKE Receiver
`7.4.2 Power Control
`7.4.3 Soft Handover
`7.4.4 Interfrequency Handover
`7.4.5 Multiuser Detection
`
`,
`
`~
`
`112
`
`115
`
`1 17
`
`119
`1 19
`
`120
`
`123
`
`127
`
`130
`
`131
`
`138
`
`141
`
`144
`
`153
`
`155
`
`155
`
`156
`
`157
`161
`171
`171
`172
`173
`175
`175
`
`Page 6 of 24
`
`EX. 1017
`Ericsson v. Intellectual Ventures
`Page 6 of 24
`
`

`

`X
`
`
`
`References
`
`Chapter 8
`
`Multi — Carrier CDMA
`
`)
`
`8. 1
`
`8.2
`
`8.3
`
`8.4
`
`8.5
`
`Introduction
`
`Channel Model
`
`DS-CDMA and MC-CDMA Systems
`
`8.3.1 DS—CDMA System
`
`8.3.2 MC—CDMA System
`
`,
`
`MC—CDMA System Design
`
`BEP LOWER Bound
`
`8.5.1 DS—CDMA System
`
`8.5.2 MC—CDMA System
`
`8.5.3 BEP Lower Bound Equivalence
`
`8.6
`
`Numerical Results
`
`8.6.1 MC-CDMA System Design
`
`8.6.2 Down — Link BEP Performance
`
`8.6.3 Up — Link BER Performance
`
`
`
`
`8.7
`
`Conclusions
`
`Appendix 8A
`
`References
`
`Chapter 9
`
`Orthogonal Frequency Division Multiple Access
`
`9. 1
`
`9.2
`
`9.3
`
`9.4
`
`Introduction
`
`Frequency Hopping OFDMA
`
`Differences between OFDMA and MC—CDMA
`
`OFDMA System Description
`9.4.1 Channel Coding
`9.4.2 Modulation
`
`9.4.3 Time and Frequency Synchronization
`
`9.4.4
`
`Initial Modulation Timing Synchronization
`
`9.4.5
`
`Initial Frequency Offset Synchronization
`
`176
`
`179
`
`179
`
`180
`
`182
`
`182
`
`185
`
`189
`
`194
`
`194
`
`195
`
`196
`
`197
`
`197
`
`199
`
`203
`
`206
`
`208
`
`209
`
`213
`
`213
`
`213
`
`215
`
`217
`' 220
`220
`
`221
`
`221
`
`222
`
`EX. 1017
`Ericsson v. Intellectual Ventures
`Page 7 of 24
`
`

`

`
`
`9.4.6
`
`Synchronization Accuracy
`
`9.4.7
`
`Power Control
`
`9.4.8 Random Frequency Hopping Operation
`
`9.4.9 Dynamic Channel Allocation (Fast DCA)
`
`9.4.10 Dynamic Channel Allocation ( Simple DCA )
`
`9.4.11 Capacity of OFDMA
`
`9.5
`
`Conclusions
`
`References
`
`Chapter 10
`
`Applications of OFDM
`
`10. 1
`
`10.2
`
`10.3
`
`10.4
`
`Introduction
`
`Digital Audio Broadcasting
`
`Terrestrial Digital Video Broadcasting
`
`Magic WAND
`
`10.4.1 Magic WAND Physical Layer
`
`10.4.2 Coding
`
`10.4.3 Simulated Error Probabilities
`
`10.4.4 Effects of Clipping
`
`10.4.5 Magic WAND Medium Access Control Layer
`
`10.5
`
`IEEE 802.11, HIPERLAN/Z, and MMAC Wireless LAN Standards
`
`10.5.1 OFDM Parameters
`
`10.5.2 Channelization
`
`10.5.3 OFDM Signal Processing
`
`10.5.4 Training
`
`10.5.5 Differences between IEEE 802.11, HIPERLAN/Z
`and MAC
`
`10.5.6 Simulation Results
`
`References
`
`About the Authors
`
`Index
`
`222
`
`223
`
`224
`
`225
`
`227
`
`227
`
`227
`
`228
`
`229
`
`229
`
`229
`
`231
`
`233
`
`234
`
`236
`
`236
`
`237
`
`238
`
`241
`
`243
`
`244
`
`245
`
`246
`
`249
`
`250
`
`252
`
`255
`
`257
`
`176
`
`179
`
`179
`
`180
`
`182
`
`182
`
`185
`
`189
`
`194
`
`194
`
`195
`
`196
`
`197
`
`197
`
`199
`
`203
`
`206
`
`208
`
`209
`
`213
`
`213
`
`213
`
`215
`
`. 217
`220
`
`220
`
`221
`
`221
`
`222
`
`
`
`Page 8 of 24
`
`EX. 1017
`Ericsson v. Intellectual Ventures
`Page 8 of 24
`
`

`

`9.1
`
`
`INTRODUCTION
`
`
`
`The previous chapter described some ways in which OFDM could be used both as a
`modulation scheme and as part of the multiple access technique, by applying a
`spreading code in the frequency domain. In this chapter, a variation on this theme is
`described, namely orthogonal
`frequency division multiple access
`(OFDMA).
`In
`OFDMA, multiple access is realized by providing each user with a fraction of the
`available number of subcarriers. In this way, it is equal to ordinary frequency division
`multiple access (FDMA); however, OFDMA avoids the relatively large guard bands
`that are necessary in FDMA to separate different users. An example of an OFDMA
`time-frequency grid is shown in Figure 9.1, where seven users a to g each use a certain
`fraction—Which may be different for each user—0f the available subcarriers. This
`particular example in fact is a mixture of OFDMA and Time Division Multiple Access
`(TDMA), because each user only transmits in one out of every four timeslots, which
`may contain one or several OFDM symbols.
`
`
`
`9.2
`FREQUENCY-HOPPING OFDMA
`In the previous example of OFDMA, every user had a fixed set of subcarriers. It is a
`relatively easy change to allow hopping of the subcarriers per timeslot, as depicted in
`Figure 9.2. Allowing hopping with different hopping patterns for each user actually
`transforms the OFDMA system in a frequency—hopping CDMA system. This has the
`benefit of an increased frequency diversity, because each user uses all of the available
`bandwidth, as well as the interference averaging benefit that is common for all CDMA
`variants. By using forward—error correction coding over multiple hops, the system can
`correct for subcarriers in deep fades or subcarriers that are interfered by other users.
`Because the interference and fading characteristics change for every hop, the system
`
`
`
`
`
`
`
`
`
`Orthogonal Frequency Division Multiple Access
`
`CHAPTER 9
`
`
`
`
`
`EX. 1017
`Ericsson v. Intellectual Ventures
`Page 9 of 24
`
`

`

`214
`
`
`performance depends on the average received signal power and interference, rather than
`on the worst case fading and interference power.
`
`Figure 9.1 Example of the time-frequency grid with seven OFDMA users, a to g, which all have a fixed
`set of subcarriers every four timeslots.
`
`
`
`
`
`
`
`
`
`
`
`Frequency
`Frequency
`
`
`
`
`
`Figure 9.2 Example of the time—frequency grid with three hopping users, a, b and c, which all have one
`hop every four time slots.
`
`
`
`A major advantage of frequency-hopping CDMA systems over direct-sequence
`or multicarrier CDMA‘systems is that it
`is relatively easy to eliminate intracell .
`interference by using orthogonal hopping patterns within a cell. An example of such an
`orthogonal hopping set is depicted in Figure 9.3. For N subcarriers; it is always possible
`to construct N orthogonal-hopping patterns. Some useful construction rules for
`generating hopping patterns can be found in [1].
`‘
`
`.;
`
`EX. 1017
`Ericsson v. Intellectual Ventures
`Page 10 of 24
`
`

`

`
`
`
`
`g
`
`i
`i
`(i
`
`‘
`
`l ltl|l
`
`‘
`1;
`"
`
`I
`
`X
`'
`iii
`
`4
`
`a
`i
`‘
`
`‘1
`
`‘
`I
`,
`
`1‘
`l;
`‘
`
`x
`4'
`;
`
`‘
`3
`
`
`
`
`
`Page 11 of 24
`
`215
`
`
`er and interference, rather than
`
`Frequency
`
`Figure 9.3 Example of six orthogonal hopping patterns with six different hopping frequencies.
`
`9.3
`
`DIFFERENCES BETWEEN OFDMA AND MC-CDMA
`
`'
`
`.
`
`.users, atog,whichallhaveafixed
`
`
`
`lsers, a, b and c, which all have one
`
`' systems over direct—sequence
`y easy to eliminate intracell ‘
`a cell. An-example Of such an
`lbcarriers, 1t 15 always 130351131e
`156ml construction rules for
`
`‘
`
`J
`
`7 The main differences between OFDMA and the MC-CDMA techniques dlscussed in
`Chapter 8 is that users within the same cell use a distinct set of subcarriers, while in
`MC—CDMA, all users use all subcarriers simultaneously. To distinguish different users,
`orthogonal or near—orthogonal spreading codes are used in MC-CDMA. Because of
`code distortion by multipath fading channels, however, MC-CDMA loses
`its
`orthogonality in the uplink even for a single cell. This makes rather complicated
`equalization techniques necessary, which introduce a loss in SNR performance and
`diminish the complexity advantage of OFDM over single-carrier techniques. OFDMA
`does not have this disadvantage, because in a single cell, all users have different
`subcarriers,
`thereby eliminating the possibility of
`intersymbol or
`intercarrier
`interference. Hence, OFDMA does not suffer from intracell interference, provided that
`the effects of frequency and timing offsets between users are kept at a sufficiently low
`level. This is a major advantage of OFDMA compared with MC-CDMA and DS-
`CDMA, because in those systems,
`intracell
`interference is the main source of
`interference. A typical ratio of intracell and intercell interference is 0.55 [2]. Because
`the system capacity is inversely proportional to the total amount of interference power,
`a capacity gain of2.8 can be achieved by eliminating all intracell interference, so that is
`the maximum capacity gain of OFDMA over DS—CDMA and MC—CDMA networks.
`The main advantages of using CDMA in general or MC—CDMA in particular is
`interference averaging. In CDMA, the interference consists of a much larger number of
`interfering signals than the interference in a non-CDMA system. Each interfering signal
`'
`is subject to independent fading, caused by shadowing and multipath effects. For both
`7
`the CDMA and the non—CDMA system, an outage occurs when the total interference
`,
`power (after despreading for the CDMA system) exceeds some maximum value. In a
`:
`non-CDMA system, the interference usually consists of a single or a few cochannel
`;
`interferers. Because of fading, the interference power is fluctuating over a large range,
`'''w.‘_5 '':*«II-mum:-
`
`
`
`
`
`l
`»,
`fl
`1
`ti
`,1
`l
`J
`[l i
`l h
`[l
`3'
`ii
`51
`
`l
`l
`
`EX. 1017
`Ericsson v. Intellectual Ventures
`Page 11 of 24
`
`

`

`
`
`ll llll
`
`216
`
`so a large fading margin has to be taken into account, which reduces system capacity.
`In a CDMA system, the interference is a sum of a large number of interfering signals.
`Because all these signals fade independently, the fluctuation in the total interference
`power is much less than the power fluctuation of a single interfering signal. Hence, in a
`CDMA system the fading margin can be significantly smaller than the margin for a
`non—CDMA system. This improvement in margin largely determines the capacity gain
`of a CDMA system.
`
`In OFDMA, interference averaging is obtained by having different hopping
`patterns Within each cell. The hopping sequences are constructed in such a way that two
`users in different cells interfere with each other only during a small fraction of all hops.
`In a heavily loaded system, many hops will interfere, but the interference will be
`different for each hop. Hence, by forward—error correction across several hops, the
`OFDMA performance will be limited by the average amount of interference rather than
`the worst case interference. An additional advantage of OFDMA over DS—CDMA and
`MC—CDMA is that there are some relatively simple ways to reduce the amount of
`intercell interference. For instance, the receiver can estimate the signal quality of each
`hop and use this information to give heavily interfered hops a lower weight in the
`decoding process.
`
`Another important feature of CDMA is the possibility to perform soft handover
`by transmitting two signals from different base stations simultaneously on the same
`channel to one mobile terminal. Combining the signals from different base stations
`gives a diversity gain that significantly reduces the fading margin, because the
`probability that two base stations are in a fade is much smaller than the probability that
`one base station is in a fade. Less fading means that less power has to be transmitted,
`and hence less interference is generated, which gives an improvement in the capacity of
`the system. A nice feature of CDMA soft handover is that it has no impact on the
`complexity of the mobile terminal; as far as the mobile terminal is concerned, the
`overlapping signals of different base stations have the same effect as overlapping
`signals caused by multipath propagation.
`
`For OFDMA systems, two basic soft handover methods exist, applicable to both
`the uplink and the downlink.
`(base station—to-mobile, mobile—to—base station). A
`requirement for both methods is that the transmissions from and to the base stations are
`synchronized such that the delay differences at the two base stations are well within the
`guard time of the OFDM symbols.
`
`The first technique is to use the same set of subcarriers and the same hopping
`sequence in two cells to connect to two base stations. Hence, in the downlink the
`mobile receives a sum of two signals with identical data content. The mobile is not able
`to distinguish between the two base stations; the effect of soft handover is similar to
`that of adding extra multipath components, increasing the diversity gain. This type of
`soft handover is similar to soft handover in DS—CDMA networks.
`
`The second way for soft handoff is to use different sets of subcarriers in two
`cells. In contrast to the first method, in the downlink the mobile has to distinguish noW
`
`Page 12 of 24
`
`'W—q ....m-ml...”-
`
`l
`
`—.—
`
`l
`
`"w....0..._WM..-..-...—...-._._J...m-i..__m....
`
`
`
`
`“mm-mm“w..«an...nm-vmn
`
`.m-r,y»._..-..
`
`éa l
`
`EX. 1017
`Ericsson v. Intellectual Ventures
`Page 12 of 24
`
`

`

`
`
`
`
`between the two base stations. It has to demodulate the signals from the two base
`stations separately, after which they can be combined, preferably by using maximal
`ratio combining. This type of soft handover is similar to the one that could be used in a
`non-CDMA network.
`
`217
`
`"
`
`
`
`
`
`
`
`bandslot of 100 kHz
`
`
`
`:;—_
`
`
`
`iich reduces system capacity.
`number of interfering signals.
`ition in the total interference
`
`interfering signal. Hence, in a
`maller than the margin for 3.
`determines the capacity gain
`
`by having different hopping
`tructed in such a way that two
`1g a small fraction of all hops.
`but the interference will be
`
`tion across several hops, the
`nut of interference rather than
`)FDMA over DS—CDMA and
`
`ays to reduce the amount of
`rate the signal quality of each
`. hops a lower weight in the
`
`lility to perform soft handover
`. simultaneously on the same
`. from different base stations
`fading margin, because the
`Jaller than the probability that
`'. power has to be transmitted,
`nprovement in the capacity of
`that it has no impact on the
`le terminal is concerned, the
`
`: same effect as overlapping
`
`:thods exist, applicable to both
`, mobile-to-base station). A
`)m and to the base stations are
`lse stations are Well within the
`
`:arriers and the same hopping
`Hence, in the downlink the
`:ontent. The mobile is not able
`of soft handover is similar to .
`
`1e diversity gain. This type of
`
`'ent sets of subcarriers in two
`
`mobile has to distinguish now ‘j‘
`
`
`
`
`
`
`
`
`
`Advantages of the second method over the first—in the downlink—are an
`
`
`increased SNR gain because of receiver diversity, and more freedom for the base
`
`
`stations to allocate available subcarriers. In the first method, base stations are forced to
`
`
`use the same subcarriers. A main advantage of the first method is its simpler
`
`
`
`implementation; no additional hardware is needed, only some extra protocol features to
`
`
`connect to two base stations simultaneously. The second method does require extra
`
`
`hardware, because it has to demodulate an extra set of subcarriers. Further, it has to
`
`
`perform extra processing for the maximal ratio combining of the signals from the
`
`
`.
`different base stations.
`
`
`
`
`9.4
`OFDMA SYSTEM DESCRIPTION
`
`
`
`
`As an example of an OFDMA system, this section gives a description of a system that
`
`was proposed for the European UMTS [3, 4]. Table 9.1 summarizes the parameters and
`
`
`key technical characteristics of this OFDMA air interface.
`
`
`Table 9.1
`
`
`
`
`
`
`Parameters of the Proposed OFDMA system
`
`
`
`
`
`
`
`
`
`
`
`
`
`OOUJ
`-_:_
`
`
`
`
`H
`bandslot and 1 time slot (= 1 s mbol)
`
`
`
`
`
`
`4time slots and 1 bandslot
`Modulation block
`
`
`
`
`
`
`
`Figure 9.4 illustrates the time-frequency grid of the OFDMA system. The
`resources (time and frequency) are allocated based on the type of services and
`
`
`operational environment. The number of timeslots and bandslots per user is variable to
`
`
`realize variable data rates. The smallest data rate is obtained for one bandslot of 24
`
`
`
`
`
`
`
`
`
`subcarriers per timeslot of 288.46 us.
`
`
`
`
`
`
`
`
`EX. 1017
`Ericsson v. Intellectual Ventures
`Page 13 of 24
`
`

`

`218
`
`
`
`Subcarfier
`
`Spadng:
`t1fiEkHz
`
`
`Frequency
`
`
`
`fimefim
`288.46ps
`
`fime
`
`Figure 9.4 Time—frequency grid.
`
`The following summary shows some advantages of the proposed OFDMA
`
`system:
`
`0 Use of frequency—hopping OFDMA for interference averaging and frequency
`diversity;
`.
`
`o Time-division duplex MAC with dynamic channel allocation used for unpaired
`spectrum allocations, asymmetrical services, and unlicensed usage;
`
`0 Straightforward and efficient high bit rate support by allocating more subcarriers
`and/or timeslots;
`
`0
`
`small guard band requirements at approximately 100 kHz;
`
`0 No frequency planning option available; effective re-use factor of 1;
`
`o GSM backwards compatibility; and
`
`0 Minimum bandwidth requirements for system deployment only 1.6 MHz (or less)
`and deployment possible in steps of IOOkHz.
`
`
`
`Figure 9.5 shows the TDMA frame structure. Each frame is of length 4.615 ms,
`which is divided into 4 subframes of length 1.1534 ms. A sub-frame contains 4 time-
`slots of duration 288.46 us. The timeslot contains a guard period, power control
`information and data. Every OFDM symbol is mapped onto one time—slot. The structure
`of an OFDMA symbol is depicted in Figure 9.6.
`
`EX. 1017
`Ericsson v. Intellectual Ventures
`Page 14 of 24
`
`

`

`w
`
`219
`
`
`
`
`4.615ms irgme.I ‘
`‘
`4 frames (13.46ms) Interleave
`_,,
`
`4 lrames (18.46msi lnterleave
`,
`
`
`
`l
`
`
`
`
`
`
`
`
`
`
` 288.46ius] time s
`<———+
`
`
`
`
`
`
`s of the proposed OFDMA
`
`Time Alignment
`
`Guard period
`
`:e averaging and frequency
`
`allocation used for unpaired
`
`I allocating more subcarriers
`
`ment only 1.6 MHz (or less)
`
`[frame is of length 4.615 1115,
`\ sub—frame contains 4 time—
`
`;uard period, power control
`0 one time-slot. The structure
`
`
`
`Figure 9.5 Frame (TDMA) Structure.
`
`Modulation Period (283.46us)
`
` Effective Modulation Period (240115)
`
`Figure 9.6 OFDM modulation burst.
`
`'
`
`The whole system frequency band is divided into small blocks (bandslots) with
`a fixed number of subcarriers. To maintain compatibility with GSM, a ,100—kHz
`bandslot is chosen that consists of 24 subcarriers. Therefore, the subcarrier spacing is
`100/24 = 4.167 kHz. Figure 9.7 shows the OFDMA frequency structure.
`
`the two subcarriers at the edge of the bandslot are left
`In each bandslot,
`unmodulated to relax receiver blocking requirements. In addition, the interference of
`two adjacent blocks of subcarriers
`is
`reduced, which may occur when their
`orthogonality is compromised because of nonlinear PA effects. Adjacent bandslots can
`be concatenated to allow transmission of wideband services.
`
`vi
`
`I
`
`.
`
`‘1'
`l
`
`i
`
`.j
`l
`
`;
`
`i
`
`ii
`
`i
`i
`"
`9 i
`
`i
`
`‘
`
`~
`p
`
`
`
`EX. 1017
`Ericsson v. Intellectual Ventures
`Page 15 of 24
`
`

`

`
`
`
`:m I"m CD:
`
`System Band
`
`
`
`
`l
`I
`
`:
`I
`
`I
`
`Guard band
`:
`I 4—»
`
`Guard band
`‘——>
`Band Slot 100kHz
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
` I
`
`
`uard carrierieuard carrier
`(not used! ! snot used)
`
`‘ ’
`OFDM sub—carrler spacing = 4.17kHz
`
`i
`
`Figure 9.7 OFDMA frequency structure.
`
`9.4.1 Channel Coding
`
`Convolutional encoding and soft-decision Viterbi decoding is used for the basic data
`transmission. The objective of this coding is to achieve good quality in the tough
`mobile radio channel. A constraint length of seven is used together with variable
`coderates in the range of 1/4 to 3/4. To achieve very low bit—error rates (e.g. 106) for
`video encoding or data transmission, a concatenated coding scheme is used with an
`inner convolutional code and an outer Reed—Solomon code.
`
`9.4.2 Modulation
`
`The modulation schemes of the OFDMA proposal are QPSK and 8—PSK with
`differential encoding in the frequency domain. An optional coherent mode with 16-
`QAM is available, which uses pilot subcarriers to obtain a channel estimate at the
`
`receiver. For differential encoding, each bandslot contains one known reference
`
`subcarrier value, as depicted in Figure 9.8.
`
`
`
`
`
`
`EX. 1017
`Ericsson v. Intellectual Ventures
`Page 16 of 24
`
`

`

`
`
` Hi
`
`I I
`Guard band
`
`Differential Encoding
`Reference Subcarrier
`
`
`
`
`2 Bandslot = 200kHz
`
`
`Figure 9.8 Reference subcarrier allocation.
`
`
`
`9.4.3 Time and Frequency Synchronization
`
`
`Synchronization is an essential issue for the OFDMA system. The following aspects are
`considered for uplink and downlink:
`initial modulation timing synchronization,
`
`modulation timing tracking, initial frequency offset synchronization, and frequency
`tracking.
`
` 9.4.4
`
`Initial Modulation Timing Synchronization
`
`
`Initial timing synchronization is required to adjust the mobile station’s internal timing
`to the base station’s timeframe. After switching on, the mobile station monitors the
`initial acquisition channel (IACH) and the broadcast channel (BCCH).
`
`
`
`1g is used for the basic data
`: good quality in the tough
`used together with variable
`bit—error rates (e.g. 106) for
`.ing scheme is used with an
`
`re QPSK and 8—PSK with
`nal coherent mode with 16—
`n a channel estimate at the
`tains one known reference
`
`
`
`
`
`
`
`
`
`
`After the mobile has detected the base station’s timing, it sends a random access
`channel (RACH) packet to the base station. The base station measures the time offset
`for the received RACH packet and sends back the necessary timing advance to the
`mobile (similar to GSM). In the frame structure of the OFDMA system, reserved slots
`for reception of RACH packets exist.
`
`
`
`
`
`
`domain, where a time shift is observed as a phase rotation.
`
`Because of the time and frequency structure of the OFDMA system, the timing
`tracking is less critical compared with other OFDM systems where users are interleaved
`in the frequency domain. The base station can measure the position of the received
`OFDM burst within the allocated slot for each mobile station individually and send the
`according timing alignment information back to the mobile station. In addition, timing
`information can be refined and tracked after the transformation in the subcarrier
`
`
` EX. 1017
`
`EX. 1017
`Ericsson v. Intellectual Ventures
`Page 17 of 24
`
`

`

`222
`
`
`In the mobile station, the timing information is obtained and adjusted by the
`above-mentioned correlation algorithm. Accurate timing information is required to
`determine the position of the useful data samples within each burst so the FFT Window
`can be placed correctly. The guard samples relax the requirement for accurate timing
`because the position of the FFT window can be shifted within the guard time without
`performance degradation. Additional timing offset correction can be performed to cope
`with the FFT window misplacements.
`
`Both algorithms can also be combined. The alignment values are calculated
`regularly and reported to the mobile station. Accuracy requirements are relaxed because
`the design of the burst allows some overlapping arrival (another advancing feature of
`the raised cosine pulse—shaping besides the reduction of out-of—band emission). In
`addition, the guard time helps to compensate timing misalignments. The OFDMA burst
`design provides a guard interval at the front and an additional guard interval at the back
`
`‘4
`
`,
`§
`
`i
`
`-;
`
`3
`
`l
`
`‘
`
`.
`
`!
`
`.
`'
`
`i
`
`9.4.5
`
`Initial Frequency Offset Synchronization
`
`After initial timing synchronization of the mobile station, the frequency offset can be
`measured by phase comparison of the (ideally) equal time samples within each burst.
`Equal samples are placed in the guard interval of the OFDM burst. A phase rotation
`indicates a frequency offset. Using this technique, a frequency error up to half the
`subcarrier spacing can be detected. The initial offset, however, can be larger, so it has
`to be detected using the specially designed symbols in the IACH channel.
`
`9.4.6 Synchronization Accuracy
`
`The proposed synchronization, acquisition, and tracking algorithm is independent of the
`modulation scheme (coherent or noncoherent). For coherent 16—QAM reception, further
`processing in the frequency (subcan‘ier domain)
`is possible to improve the
`performance. Frequency domain time tracking (or combined time—domain/frequency—
`domain- tracking algorithms) can be based on observing phase shifts of the known pilots
`within the time-frequency grid‘ on the subcarrier domain.
`
`
`
`In the downlink, only an IACH is multiplexed to allow fast and precise initial
`timing and frequency synchronization. In the actual communication mode, timing and
`frequency tracking can be performed using a correlation-based synchronization
`algorithm. In the uplink, the rough timing offset is detected by the base station by
`measuring the arrival time of the RACH burst. This gives an initial time-advance value
`that is reported back to the mobile. During communication, the arrival time of the burst
`is detected by the base station using the proposed tracking algorithm (same as in the
`downlink) or a tracking algorithm in the frequency (subcarrier) domain, based on the
`detected constellation rotation.
`
`EX. 1017
`Ericsson v. Intellectual Ventures
`Page 18 of 24
`
`

`

`
`
`223
`
`
`
`
`
`
`mm”nqu-ul’mmw-‘vw
`
`of the OFDM symbol—see Figure 9.9—which provides robustness against a timing
`inaccuracy of :10 us.
`
`Modulation Period (288.46ps)
`Effective Modulation Period (240us)
`
`
`
`Optimum receiver FFT window
`
`No degradation
`
`'Delayed' receiver FFT window
`No degradation
`
`‘Early' receiver FFT window No degradation
`Small degradation
` 'Too Early' receiver FFT window
`
`Figure 9.9 OFMDA burst and synchronization requirements.
`
`9.4.7 Power Control
`
`Power control in the uplink removes the unevenness of received signal strength at the
`base station side and decreases the total power to the minimum level required to
`support the specified quality of service (e.g. BER). The accuracy is less critical than
`CDMA because with OFDMA, orthogonality is always provided within one cell. A
`precise power control, however, not only improves the transmission performance but
`also minimizes the interference to other cells and therefore increases overall capacity.
`
`The OFDMA concept uses both closed-loop and open—loop power control.
`Based on quality parameters, measured on a slot-by—slot basis, the power is adjusted in
`the mobile as well as in the base station transmitter. Each receiver measures the quality
`of the received burst (C/I ratio) and transmits in the next burst a request to the opposite
`transmitter to increase, keep, or decrease the power level in steps of 1 dB. For the
`fastest power control mode, one subcarrier is dedicated to carry power control
`information, and the power is then adjusted on a fraIne-by—frame basis (each 1.152 ms).
`Figure 9.10 depicts the power control operation.
`
`btained and adjusted by the
`information is required to
`rch burst so the FFT window
`
`iirement for accurate timing
`ithin the guard time without
`on can be performed to cope
`
`, the frequency offset can be
`e samples within each burst.
`iDM burst. A phase rotation
`:quency error up to half the
`'ever, can be larger, so it has
`
`gorithm is independent of the
`1t 16—QAM reception, further
`possible to improve the
`ined time—domain/frequency—
`ase shifts of the known pilots
`
`allow fast and precise initial
`nunication mode, timing and
`iation—based synchronization
`acted by the base station by
`an initial time-advance value
`
`1, the arrival time of the burst
`ng algorithm (same as in the
`:arrier) domain, based on the
`
`gnment values are calculated
`airements are relaxed because
`
`:another advancing feature of
`of out-of—band emission). In
`
`ignments. The OFDMA burst
`)nal guard interval at the back
`
`
`
`
`
`
`
`
`Page 19 of 24
`
`EX. 1017
`Ericsson v. Intellectual Ventures
`Page 19 of 24
`
`

`

`224
`
`
`Received Data
`
`(solt decision)
`Demodulator
`
`
`.
`Received
`Sinal
`
`
`
`Quamy
`Detect
`
`Receive channel
`
`Power Control Data
`
`Transmit channel Power Control Data
`
`
`
`
`
`
`Transmit
`Si nl
`
`
`
`Received
`Bits
`
`Transmit
`
`Bits
`
`
`
`g MOdUlaiOTTransmitData
`
`
`Power Control Step = -l, 0, +1 [dB]
`Power Control Period = 1.2 [mseclcontrt]
`
`Figure 9.10 Operation of power control in a mobile station.
`
`9.4.8 Random Frequency-Hopping Operation
`
`Frequency hopping is very effective to achieve frequency diversity and interference
`diversity. Frequency diversity is useful to average the frequency—selective channel
`properties (fading dips)