ey
`be .\e
`me_@at: If fi,
`
`Johan Skold APPLE 1016
`
`for Mobile Broadband
`\
`_
`
`SalePyslellaiteia
`Stefan Parkvall
`
`APPLE 1016
`
`1
`
`

`

`4G LTE/LTE-Advanced
`for Mobile Broadband
`
`Erik Dahlman, Stefan Parkvall, and
`Johan Sköld
`
`(cid:33)(cid:45)(cid:51)(cid:52)(cid:37)(cid:50)(cid:36)(cid:33)(cid:45)(cid:0)(cid:115)(cid:0)(cid:34)(cid:47)(cid:51)(cid:52)(cid:47)(cid:46)(cid:0)(cid:115)(cid:0)(cid:40)(cid:37)(cid:41)(cid:36)(cid:37)(cid:44)(cid:34)(cid:37)(cid:50)(cid:39)(cid:0)(cid:115)(cid:0)(cid:44)(cid:47)(cid:46)(cid:36)(cid:47)(cid:46)(cid:0)(cid:115)(cid:0)(cid:46)(cid:37)(cid:55)(cid:0)(cid:57)(cid:47)(cid:50)(cid:43)(cid:0)(cid:115)(cid:0)(cid:47)(cid:56)(cid:38)(cid:47)(cid:50)(cid:36)
`(cid:48)(cid:33)(cid:50)(cid:41)(cid:51)(cid:0)(cid:115)(cid:0)(cid:51)(cid:33)(cid:46)(cid:0)(cid:36)(cid:41)(cid:37)(cid:39)(cid:47)(cid:0)(cid:115)(cid:0)(cid:51)(cid:33)(cid:46)(cid:0)(cid:38)(cid:50)(cid:33)(cid:46)(cid:35)(cid:41)(cid:51)(cid:35)(cid:47)(cid:0)(cid:115)(cid:0)(cid:51)(cid:41)(cid:46)(cid:39)(cid:33)(cid:48)(cid:47)(cid:50)(cid:37)(cid:0)(cid:115)(cid:0)(cid:51)(cid:57)(cid:36)(cid:46)(cid:37)(cid:57)(cid:0)(cid:115)(cid:0)(cid:52)(cid:47)(cid:43)(cid:57)(cid:47)
`Academic Press is an imprint of Elsevier
`
`2
`
`

`

`Academic Press is an imprint of Elsevier
`The Boulevard, Langford Lane, Kidlington, Oxford, OX5 1GB, UK
`30 Corporate Drive, Suite 400, Burlington, MA 01803, USA
`
`First published 2011
`
`Copyright © 2011 Erik Dahlman, Stefan Parkvall & Johan Sköld. Published by Elsevier Ltd. All rights reserved
`
`The rights of Erik Dahlman, Stefan Parkvall & Johan Sköld to be identified as the authors of this work has been
`asserted in accordance with the Copyright, Designs and Patents Act 1988.
`
`No part of this publication may be reproduced or transmitted in any form or by any means, electronic or
`mechanical, including photocopying, recording, or any information storage and retrieval system, without
`permission in writing from the publisher. Details on how to seek permission, further information about the
`Publisher’s permissions policies and our arrangement with organizations such as the Copyright Clearance
`Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions
`
`This book and the individual contributions contained in it are protected under copyright by the Publisher
`(other than as may be noted herein).
`
`Notices
`Knowledge and best practice in this field are constantly changing. As new research and experience broaden our
`understanding, changes in research methods, professional practices, or medical treatment may become necessary.
`
`Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using
`any information, methods, compounds, or experiments described herein. In using such information or methods
`they should be mindful of their own safety and the safety of others, including parties for whom they have a
`professional responsibility.
`
`To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability
`for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or
`from any use or operation of any methods, products, instructions, or ideas contained in the material herein.
`
`British Library Cataloguing-in-Publication Data
`A catalogue record for this book is available from the British Library
`
`Library of Congress Control Number: 2011921244
`
`ISBN: 978-0-12-385489-6
`
`For information on all Academic Press publications
`visit our website at www.elsevierdirect.com
`
`Typeset by MPS Limited, a Macmillan Company, Chennai, India
`www.macmillansolutions.com
`
`Printed and bound in the UK
`
`11 12 13 14 10 9 8 7 6 5 4 3 2 1
`
`3
`
`

`

`11.4 Uplink L1/L2 Control Signaling
`
`229
`
`As mentioned earlier, uplink L1/L2 control signaling includes hybrid-ARQ acknowledgements,
`channel-state reports, and scheduling requests. Different combinations of these types of messages are
`possible as described later, but to explain the structure for these cases it is beneficial to discuss sepa-
`rate transmission of each of the types first, starting with the hybrid-ARQ acknowledgement and the
`scheduling request. There are three formats defined for PUCCH, formats 1, 2, and 3, each capable of
`carrying a different number of bits.
`
`11.4.1.1 PUCCH Format 1
`Hybrid-ARQ acknowledgements are used to acknowledge the reception of one (or two in the case of
`spatial multiplexing) transport blocks on the DL-SCH. The hybrid-ARQ acknowledgement is only
`transmitted when the terminal correctly received control signaling related to DL-SCH transmission
`intended for this terminal on one of the PDCCHs. In the case of downlink carrier aggregation, there
`can be multiple simultaneous DL-SCHs for a single terminal, one per downlink component carrier and,
`consequently, multiple acknowledgement bits need to be conveyed in the uplink. This is described fur-
`ther below once the non-carrier-aggregated case has been treated.
`If no valid DL-SCH-related control signaling is detected, then nothing is transmitted on the
`PUCCH (i.e. DTX). Apart from not unnecessarily occupying PUCCH resources that can be used for
`other purposes, this allows the eNodeB to perform three-state detection, ACK, NAK, or DTX, on the
`PUCCH received. Three-state detection is useful as NAK and DTX may need to be treated differently.
`In the case of NAK, retransmission of additional parity bits is useful for incremental redundancy, while
`for DTX the terminal has most likely missed the initial transmission of systematic bits and a better
`alternative than transmitting additional parity bits is to retransmit the systematic bits.
`Scheduling requests are used to request resources for uplink data transmission. Obviously, a
`scheduling request should only be transmitted when the terminal is requesting resources, otherwise
`the terminal should be silent to save battery resources and not create unnecessary interference. Hence,
`unlike hybrid-ARQ acknowledgements, no explicit information bit is transmitted by the scheduling
`request; instead the information is conveyed by the presence (or absence) of energy on the corre-
`sponding PUCCH. However, the scheduling request, although used for a completely different pur-
`pose, shares the same PUCCH format as the hybrid-ARQ acknowledgement. This format is referred
`to as PUCCH format 1 in the specifications.14
`PUCCH format 1 uses the same structure in the two slots of a subframe, as illustrated in Figure
`11.23. For transmission of a hybrid-ARQ acknowledgement, the single hybrid-ARQ acknowledge-
`ment bit related to one downlink component carrier is used to generate a BPSK symbol (in the case
`of downlink spatial multiplexing the two acknowledgement bits are used to generate a QPSK sym-
`bol). For a scheduling request, the same constellation point as for a negative acknowledgement is
`used. The modulation symbol is then used to generate the signal to be transmitted in each of the two
`PUCCH slots.
`There are seven OFDM symbols per slot for a normal cyclic prefix (six in the case of an extended
`cyclic prefix). In each of those seven OFDM symbols, a length-12 sequence, obtained by phase rota-
`tion of the cell-specific sequence as described earlier, is transmitted. Three of the symbols are used
`
`14 There are actually three variants in the LTE specifications, formats 1, 1A, and 1B, used for transmission of scheduling
`requests and one or two hybrid-ARQ acknowledgements respectively. However, for simplicity, they are all referred to as
`format 1 herein.
`
`4
`
`

`

`One/two bits hybrid-ARQ acknowledgement
`
`BPSK/
`QPSK
`
`One BPSK/QPSK symbol
`
`0w
`
`1w
`
`2w
`
`3w
`
`0w
`
`1w
`
`2w
`
`Same processing as first slot
`
`IFFT
`
`IFFT
`
`IFFT
`
`IFFT
`
`IFFT
`
`IFFT
`
`IFFT
`
`IFFT
`
`IFFT
`
`IFFT
`
`IFFT
`
`IFFT
`
`IFFT
`
`IFFT
`
`Length-12 phase-rotated sequence
`(varying per symbol)
`Length-4 sequence [
`w …
`0
`Length-3 sequence [
`w …
`0
`
`]
`]
`
`w
`3
`
`w
`2
`
`FIGURE 11.23
`
`PUCCH format 1 (normal cyclic prefix).
`
`1 ms subframe
`
`5
`
`

`

`11.4 Uplink L1/L2 Control Signaling
`
`231
`
`as reference signals to enable channel estimation by the eNodeB and the remaining four15 are modu-
`lated by the BPSK/QPSK symbols described earlier. In principle, the BPSK/QPSK modulation sym-
`bol could directly modulate the rotated length-12 sequence used to differentiate terminals transmitting
`on the same time–frequency resource. However, this would result in unnecessarily low capacity on the
`PUCCH. Therefore, the BPSK/QPSK symbol is multiplied by a length-4 orthogonal cover sequence.16
`Multiple terminals may transmit on the same time–frequency resource using the same phase-rotated
`sequence and be separated through different orthogonal covers. To be able to estimate the channels
`for the respective terminals, the reference signals also employ an orthogonal cover sequence, with the
`only difference being the length of the sequence – three for the case of a normal cyclic prefix. Thus,
`since each cell-specific sequence can be used for up to 3 ⭈ 12 ⫽ 36 different terminals (assuming all
`12 rotations are available; typically at most six of them are used), there is a threefold improvement in
`the PUCCH capacity compared to the case of no cover sequence. The cover sequences are three Walsh
`sequences of length 4 for the data part and three DFT sequences of length 3 for the reference signals
`(for an extended cyclic prefix, the reference-signal cover sequence is of length 2).
`A PUCCH format 1 resource, used for either a hybrid-ARQ acknowledgement or a scheduling
`request, is represented by a single scalar resource index. From the index, the phase rotation and the
`orthogonal cover sequence are derived.
`The use of a phase rotation of a cell-specific sequence together with orthogonal sequences as
`described earlier provides orthogonality between different terminals in the same cell transmitting
`PUCCH on the same set of resource blocks. Hence, in the ideal case, there will be no intra-cell inter-
`ference, which helps improve the performance. However, there will typically be inter-cell interference
`for the PUCCH as the different sequences used in neighboring cells are non-orthogonal. To rand-
`omize the inter-cell interference, the phase rotation of the sequence used in a cell varies on a symbol-
`by-symbol basis in a slot according to a hopping pattern derived from the physical-layer cell identity.
`On top of this, slot-level hopping is applied to the orthogonal cover and phase rotation to further
`randomize the interference. This is exemplified in Figure 11.24 assuming normal cyclic prefix and
`six of 12 rotations used for each cover sequence. To the phase rotation given by the cell-specific hop-
`ping a slot-specific offset is added. In cell A, a terminal is transmitting on PUCCH resource number
`3, which in this example corresponds to using the (phase rotation, cover sequence) combination
`(6, 0) in the first slot and (11, 1) in the second slot of this particular subframe. PUCCH resource number
`11, used by another terminal in cell A transmitting in the same subframe, corresponds to (11, 1) and
`(8, 2) in the first and second slots respectively of the subframe. In another cell the PUCCH resource
`numbers are mapped to different sets (rotation, cover sequence) in the slots. This helps to randomize the
`inter-cell interference.
`For an extended cyclic prefix, the same structure as in Figure 11.23 is used, with the difference
`being the number of reference symbols in each slot. In this case, the six OFDM symbols in each slot
`are divided such that the two middle symbols are used for reference signals and the remaining four
`symbols used for the information. Thus, the length of the orthogonal sequence used to spread the
`
`15 The number of symbols used for reference signals and the acknowledgement is a trade-off between channel-estimation
`accuracy and energy in the information part; three symbols for reference symbols and four symbols for the acknowledge-
`ment have been found to be a good compromise.
`16 In the case of simultaneous SRS and PUCCH transmissions in the same subframe, a length-3 sequence is used, thereby
`making the last OFDM symbol in the subframe available for the sounding reference signal.
`
`6
`
`

`

`13.2 Scheduling and Rate Adaptation
`
`277
`
`scheduling grant received in downlink subframe 0 applies to one or both of the uplink subframes
`4 and 7, depending on which of the bits in the uplink index are set.
`Similarly to the downlink case, the uplink scheduler can exploit information about channel condi-
`tions, buffer status, and priorities of the different data flows, and, if some form of interference coordina-
`tion is employed, the interference situation in neighboring cells. Channel-dependent scheduling, which
`typically is used for the downlink, can be used for the uplink as well. In the uplink, estimates of the
`channel quality can be obtained from the use of uplink channel sounding, as described in Chapter 11.
`For scenarios where the overhead from channel sounding is too costly, or when the variations in the
`channel are too rapid to be tracked, for example at high terminal speeds, uplink diversity can be used
`instead. The use of frequency hopping as discussed in Chapter 11 is one example of obtaining diversity
`in the uplink.
`Inter-cell interference coordination can be used in the uplink for similar reasons as in the down-
`link by exchanging information between neighboring cells, as discussed in Section 13.3.
`
`13.2.2.1 Uplink Priority Handling
`Multiple logical channels of different priorities can be multiplexed into the same transport block
`using the same MAC multiplexing functionality as in the downlink (described in Chapter 8).
`However, unlike the downlink case, where the prioritization is under control of the scheduler and up
`to the implementation, the uplink multiplexing is done according to a set of well-defined rules in the
`terminal as a scheduling grant applies to a specific uplink carrier of a terminal, not to a specific radio
`bearer within the terminal. Using radio-bearer-specific scheduling grants would increase the control
`signaling overhead in the downlink and hence per-terminal scheduling is used in LTE.
`The simplest multiplexing rule would be to serve logical channels in strict priority order.
`However, this may result in starvation of lower-priority channels; all resources would be given to the
`high-priority channel until its transmission buffer is empty. Typically, an operator would instead like
`to provide at least some throughput for low-priority services as well. Therefore, for each logical chan-
`nel in an LTE terminal, a prioritized data rate is configured in addition to the priority value. The logi-
`cal channels are then served in decreasing priority order up to their prioritized data rate, which avoids
`starvation as long as the scheduled data rate is at least as large as the sum of the prioritized data rates.
`Beyond the prioritized data rates, channels are served in strict priority order until the grant is fully
`exploited or the buffer is empty. This is illustrated in Figure 13.4.
`
`13.2.2.2 Scheduling Requests and Buffer Status Reports
`The scheduler needs knowledge about the amount of data awaiting transmission from the terminals to
`assign the proper amount of uplink resources. Obviously, there is no need to provide uplink resources
`to a terminal with no data to transmit as this would only result in the terminal performing padding to
`
`Prioritized
`data rate
`
`Scheduled
`data rate
`
`LCH 1
`
`LCH 2
`
`Transmitted
`
`LCH 1
`
`LCH 2
`
`Transmitted
`
`LCH 1
`
`LCH 2
`
`Transmitted
`
`FIGURE 13.4
`
`Prioritization of two logical channels for three different uplink grants.
`
`7
`
`

`

`278
`
`CHAPTER 13 Power Control, Scheduling, and Interference Handling
`
`fill up the granted resources. Hence, as a minimum, the scheduler needs to know whether the terminal
`has data to transmit and should be given a grant. This is known as a scheduling request.
`A scheduling request is a simple flag, raised by the terminal to request uplink resources from the
`uplink scheduler. Since the terminal requesting resources by definition has no PUSCH resource, the
`scheduling request is transmitted on the PUCCH. Each terminal can be assigned a dedicated PUCCH
`scheduling request resource, occurring every nth subframe, as described in Chapter 11. With a dedi-
`cated scheduling-request mechanism, there is no need to provide the identity of the terminal request-
`ing to be scheduled as the identity of the terminal is implicitly known from the resources upon which
`the request is transmitted. When data with higher priority than already existing in the transmit buffers
`arrives at the terminal and the terminal has no grant and hence cannot transmit the data, the terminal
`transmits a scheduling request at the next possible instant, as illustrated in Figure 13.5. Upon recep-
`tion of the request, the scheduler can assign a grant to the terminal. If the terminal does not receive
`a scheduling grant until the next possible scheduling-request instant, then the scheduling request is
`repeated. There is only a single scheduling-request bit, irrespective of the number of uplink compo-
`nent carriers the terminal is capable of. In the case of carrier aggregation, the scheduling request is
`transmitted on the primary component carrier, in line with the general principle of PUCCH transmis-
`sion on the primary component carrier only.
`The use of a single bit for the scheduling request is motivated by the desire to keep the uplink over-
`head small, as a multi-bit scheduling request would come at a higher cost. A consequence of the single-
`bit scheduling request is the limited knowledge at the eNodeB about the buffer situation at the terminal
`when receiving such a request. Different scheduler implementations handle this differently. One pos-
`sibility is to assign a small amount of resources to ensure that the terminal can exploit them efficiently
`without becoming power limited. Once the terminal has started to transmit on the UL-SCH, more
`detailed information about the buffer status and power headroom can be provided through the inband
`MAC control message, as discussed below. Knowledge of the service type may also be used – for exam-
`ple, in the case of voice the uplink resource to grant is preferably the size of a typical voice-over-IP
`package. The scheduler may also exploit, for example, path-loss measurements used for mobility and
`handover decisions to estimate the amount of resources the terminal may efficiently utilize.
`An alternative to a dedicated scheduling-request mechanism would be a contention-based design.
`In such a design, multiple terminals share a common resource and provide their identity as part of
`the request. This is similar to the design of the random access. The number of bits transmitted from
`a terminal as part of a request would in this case be larger, with the correspondingly larger need for
`resources. In contrast, the resources are shared by multiple users. Basically, contention-based designs
`
`Data arrives to terminal,
`triggers scheduling request
`
`SR transmitted Grant received
`
`UL-SCH transmission
`
`n
`
`n+4
`
`SR possibility
`
`SR interval
`
`SR possibility
`
`SR interval
`
`SR possibility
`
`FIGURE 13.5
`
`Scheduling-request transmission.
`
`8
`
`

`

`13.2 Scheduling and Rate Adaptation
`
`279
`
`are suitable for a situation where there are a large number of terminals in the cell and the traffic inten-
`sity, and hence the scheduling intensity, is low. In situations with higher intensities, the collision rate
`between different terminals simultaneously requesting resources would be too high and lead to an
`inefficient design.
`Although the scheduling-request design for LTE relies on dedicated resources, a terminal that has
`not been allocated such resources obviously cannot transmit a scheduling request. Instead, terminals
`without scheduling-request resources configured rely on the random-access mechanism described in
`Chapter 14. In principle, an LTE terminal can therefore be configured to rely on a contention-based
`mechanism if this is advantageous in a specific deployment.
`Terminals that already have a valid grant obviously do not need to request uplink resources.
`However, to allow the scheduler to determine the amount of resources to grant to each terminal in
`future subframes, information about the buffer situation and the power availability is useful, as dis-
`cussed above. This information is provided to the scheduler as part of the uplink transmission through
`MAC control elements (see Chapter 8 for a discussion on MAC control elements and the general
`structure of a MAC header). The LCID field in one of the MAC subheaders is set to a reserved value
`indicating the presence of a buffer status report, as illustrated in Figure 13.6.
`From a scheduling perspective, buffer information for each logical channel is beneficial, although
`this could result in a significant overhead. Logical channels are therefore grouped into logical-channel
`groups and the reporting is done per group. The buffer-size field in a buffer-status report indicates the
`amount of data awaiting transmission across all logical channels in a logical-channel group. A buffer-
`status report represents one or all four logical-channel groups and can be triggered for the following
`reasons:
`
`● Arrival of data with higher priority than currently in the transmission buffer – that is, data in a
`logical-channel group with higher priority than the one currently being transmitted – as this may
`impact the scheduling decision.
`● Change of serving cell, in which case a buffer-status report is useful to provide the new serving
`cell with information about the situation in the terminal.
`● Periodically as controlled by a timer.
`
`Logical channel index = buffer status report
`
`Logical channel index = power headroom report
`
`E LCID
`
`E LCID
`
`E LCID
`
`F
`
`L
`
`subheader
`
`subheader
`
`subheader
`
`Buffer Size
`
`Power Headroom
`
`MAC header
`
`MAC ctrl
`
`MAC ctrl
`
`RLC PDU
`
`RLC PDU
`
`Padding
`
`Transport block (MAC PDU)
`
`FIGURE 13.6
`
`Signaling of buffer status and power-headroom reports.
`
`9
`
`