2004-03-15
`
`IEEE C802.16-04/51r1
`
`Project
`
`Title
`
`Date
`Submitted
`
`Source(s)
`
`Re:
`
`Abstract
`
`Purpose
`
`Notice
`
`Release
`
`Patent
`Policy and
`Procedures
`
`IEEE 802.16 Broadband Wireless Access Working Group <http://ieee802.org/16>
`
`AAS enhancements for 1x Scalable PHY
`
`John.Liebetreu@intel.com
`mailto:dbranlund@beamreachnetworks.com
` lkotecha@beamreachnetwors.com
` mwebb@beamreachnetworks.com
`
`2004-03-15
`
`Intel
`BeamReach Networks
`Alvarion
`Proxim
`Arraycomm(?)
`Navini Networks
`
`Dale Branlund
`Lalit Kotecha
`Mike Webb
`BeamReach Networks
`755, N Mathilda Ave,
`Sunnyvale CA 94086
`IEEE P802.16-REVd/D3-2003
`
`This contribution introduces AAS enhancements for proposed 1x Scalable PHY as an optional
`feature
`
`Adopt into P802.16d/D4 draft.
`This document has been prepared to assist IEEE 802.16. It is offered as a basis for discussion and is not binding on
`the contributing individual(s) or organization(s). The material in this document is subject to change in form and
`content after further study. The contributor(s) reserve(s) the right to add, amend or withdraw material contained
`herein.
`The contributor grants a free, irrevocable license to the IEEE to incorporate material contained in this contribution,
`and any modifications thereof, in the creation of an IEEE Standards publication; to copyright in the IEEE’s name
`any IEEE Standards publication even though it may include portions of this contribution; and at the IEEE’s sole
`discretion to permit others to reproduce in whole or in part the resulting IEEE Standards publication. The
`contributor also acknowledges and accepts that this contribution may be made public by IEEE 802.16.
`The contributor is familiar with the IEEE 802.16 Patent Policy and Procedures
`<http://ieee802.org/16/ipr/patents/policy.html>, including the statement "IEEE standards may
`include the known use of patent(s), including patent applications, provided the IEEE receives
`assurance from the patent holder or applicant with respect to patents essential for compliance
`with both mandatory and optional portions of the standard." Early disclosure to the Working
`Group of patent information that might be relevant to the standard is essential to reduce the
`possibility for delays in the development process and increase the likelihood that the draft
`publication will be approved
`for publication. Please notify
`the Chair
`<mailto:chair@wirelessman.org> as early as possible, in written or electronic form, if patented
`technology (or technology under patent application) might be incorporated into a draft standard
`being developed within the IEEE 802.16 Working Group. The Chair will disclose this
`notification via the IEEE 802.16 web site <http://ieee802.org/16/ipr/patents/notices>.
`
`1
`
`ERIC-1017
`Ericsson v. IV
`Page 1 of 16
`
`

`
`2004-03-15
`
`IEEE C802.16-04/51r1
`
`1
`
`Introduction
`
`AAS can extend cell coverage by improving the system link budget. Link budget gain is
`achieved through beamforming. Beamforming coherently combines the RF wavefront
`received from multiple antennas in an adaptive combiner while increasing the order of
`the diversity combining mechanism. At the same time, AAS can increase base station
`capacity by enabling the use of higher order modulation through interference reduction
`and by enabling M-fold spectral reuse within the cell. To gain these benefits, the adaptive
`arrays must be trained using a known set of training symbols. Furthermore, robust
`signaling methods are needed to page subscriber stations when the adaptive arrays
`provided no beamforming gain. These training, paging, and initialization signals are
`collectively called AAS control signals in this document.
`
`The current OFDMA standard is silent on the definition of these signals. To ensure
`compatibility across different base stations and SSs, the control signals must be defined.
`Accordingly, a compact set of AAS control signals, compatible with the 1x Scalable
`OFDMA PHY, is proposed in this submission. The use of these controls is only required
`for systems using the optional AAS mode. Non-AAS systems are not required to use
`these signals, and therefore bear no inefficiency.
`
`2
`
` Problem Definition
`
`2.1 Broadcast Control Messages and Range
`
`Coherent beamforming with a base station antenna array can effectively increase the
`transmission range of the uni-cast channels, since there exists an optimum beamforming
`solution to serve the intended SS, but it cannot directly increase the range of broadcast
`messages on broadcast channels – most crucially, broadcast MAP bursts do not enjoy the
`extended range. An SS who cannot receive the current DL-MAP is cut-off from
`receiving other downlink traffic intended for it even though enough link budget is
`available. The same problem occurs on the uplink – any SS that cannot receive the
`broadcast UL-MAP will not be able to transmit, even though the base station can use
`coherent combining gain to close the link.
`The present OFDMA standard patches this problem in the AAS mode be redefining
`several of the broadcast messages, in particular the MAP messages, to be received as a
`series of private uni-cast messages. However, the large increase in overhead, the increase
`in latency and the inability to send uni-cast messages to portable or inactive SSs were not
`considered adequately.
`
`2.2
`
` Interference on Control Messages
`
`AAS system that employ adaptive arrays for the purpose of increasing base station
`capacity do so by aggressive reuse of frequency – often by re-using frequencies within
`
`2
`
`ERIC-1017
`Page 2 of 16
`
`

`
`2004-03-15
`
`IEEE C802.16-04/51r1
`
`the cell several times. In such an RF environment, the control messages are buried by
`interference, not only from interference generated by adjacent cells, but by interference
`generated from multiple users within the same cell. Thus, it becomes imperative to
`protect control signaling that opens and closes data flows between various SS and the
`serving base station from this interference. This implies that control signaling be
`structured to enable interference mitigation using either in time, frequency, spatial and/or
`coding dimensions.
`
`2.3
`
` Proposed Solution
`
`The proposed solution introduces low overhead control symbols and signaling that
`can be overlaid onto the 1x scalable PHY framing structure. This control signaling is
`specifically designed for the AAS mode and may be selectively removed in non-AAS
`modes. Specially, the control signaling is designed so that base stations that employ
`adaptive antenna arrays can use spatial or spatial/spectral filtering to isolate this critical
`signaling and maintain the link budget advantages described above. Reliance on
`extended uni-cast maps is reduced.
`
`3
`
` PHY Control Signaling Overview Solution
`
`The following paragraphs provide an overview of the physical layer control
`signaling supporting the optional AAS mode. The signaling mechanisms described
`herein have been rationalized and integrated with the 1x scalable frame structure. The
`control signaling consists of special symbols modulating OFDMA carriers within the 1x
`scalable bin and sub-channel structure. The AAS symbol structure minimizes overhead
`thus maintaining high airlink efficiency. The handshaking mechanisms described below
`provide reliable, low latency airlink control in co-channel interference environments.
`
`3.1 TDD Framing
`
`In the informative text that follows, the target AAS system uses time division
`duplexing (TDD). The 1x scalable frame layout uses a frame time of 5 milliseconds and
`48 OFDMA symbols per frame. The frame contains 84 bins x 48 symbols slots. For
`clarity throughout this document, a new term “partition” is used. A partition is defined as
`1 bin by 48 symbol slots. It is assumed for illustration purposes that 33 symbols are
`allocated to the forward link and 15 symbols are allocated to the reverse link resulting in
`2 to 1 asymmetry (provisioned) in the forward and reverse link rates. An AAS sub-
`channel is defined as six consecutive bins in time defined by a contiguous area of 1 bin x
`6 symbol slots in length. Mandatory CC coding and optional BTC or CTC FEC is
`supported by this frame structure. Optional 2x spreading or SFC is used on the access
`channel for improved control channel reliability.
`
`3
`
`ERIC-1017
`Page 3 of 16
`
`

`
`2004-03-15
`
`IEEE C802.16-04/51r1
`
`3.2 Reverse Link Signals
`
`A reverse link partition in the TDD frame is shown in Figure 1 for one of 84 partitions.
`The reverse link in this example provides 15 symbol slots and is organized as two AAS
`sub-channels. One of the 2 AAS sub-channels contains one AAS reverse link control
`signals transmitted once every multi-frame. A multi-frame is 1, 2, or 4 frames. Non-AAS
`systems do not send this AAS control signal.
`
`There are two physical layer control signals for the reverse link. The first is a reverse
`link initialization (RLI) signal, which allows a SS to send an AAS training signal to the
`base for a given sub-channel. The RLI provides the time-bandwidth product necessary to
`adapt up to 12 antennas at the base station. The RLI signal occurs at the beginning of the
`reverse link frame as shown in Figure 1 and is sent alternately every frame, every other
`frame or every fourth frame as provisioned by the “multi-frame parameter”. Map and
`traffic data are sent after the RLI in the first sub-channel and in subsequent sub-channels
`thereafter also shown in Figure 1. The RLI occupies a maximum of 8 bins by 8 tones (9
`tones with pilot) per bin providing 64 QPSK symbols for base station training.
`
`
`CQI, Ack / Nack . Pwr Headroom (QPSK, ? rate) CQI, Ack / Nack . Pwr Headroom (QPSK, ? rate)
`
`Ranging Ranging
`
`
`
`RLI (64 Symbols) RLI
`
`
`
`1 Subchannel 1 Subchannel
`
`
`
`2 Subchannels 2 Subchannels
`
`
`
`1 BTC FEC Block or 3 RS - CC FEC Blocks per 2 Frames 1 BTC FEC Block or 3 RS - CC FEC Blocks per 2 Frames
`
`
`
`15 Symbols/Frame, 2 Frames Shown 15 Symbols/Frame, 2 Frames Shown
`
`Reverse Link AAS Frame Structure Showing RLI Signaling
`Figure 1
`The second control signal is the reverse link access (RLA) signal. The SS uses the RLA
`to inform the base that it has information to send on the uplink. The reverse link access
`partition is identical to the traffic partition shown in Figure 1. SSs use the RLA signal
`mechanism for sending supervisory messages such as bandwidth requests and signaling
`for initial ranging. The base in turn, with coordination through its scheduling
`mechanism, sets up traffic sub-channels using forward link control signaling, either an
`FLI or FLA as described below.
`
`At least one access partition is allocated in the TDD frame for network entry and ranging,
`bandwidth request, and auxiliary SICH communications. The access partition, shown in
`Figure 2, occupies the first bin location in the frame structure. A second partition that
`occupies the last bin location may be paired with
`
`4
`
`ERIC-1017
`Page 4 of 16
`
`

`
`2004-03-15
`
`IEEE C802.16-04/51r1
`
`
`
`FrequencyFrequency
`
`
`
`TimeTime
`
`
`Antennas Antennas
`
`(1 shown here)(1 shown here)
`
`Access Sub-channel (Net Entry/BW Request sub-channel )Access Sub-channel (Net Entry/BW Request sub-channel )
`
`
`5ms5ms
`
`
`
`DownlinkDownlink
`
`
`
`UplinkUplink
`
`
`
`TTGTTG
`
`
`
`RTGRTG
`
`...
`...
`
`
`
` ... ...
`
`
`
` ... ...
`
`
`
` ... ...
`
`...
`...
`
`...
`...
`
`
`
` ... ...
`
`
`
` ... ...
`
`...
`...
`
`
`
` ... ...
`
`
`
`Band 0Band 0
`
`
`
`Band 1Band 1
`
`
`
`Band 2Band 2
`
`
`
`Band 3Band 3
`
`...
`...
`...
`...
`
`
`
`Payload Sub-channelsPayload Sub-channels
`
`
`
`Band N-1Band N-1
`
`
`
`Band NBand N
`
`
`
`PreamblePreamble
`
`
`
`Ranging, CQICH, ACKRanging, CQICH, ACK
`
`
`
`Access Sub-channel (maybe be paired for diversity and SINR enhancement)Access Sub-channel (maybe be paired for diversity and SINR enhancement)
`
`Figure 2
`
`1x Scalable Frame Layout
`
`the first to improve reliability and SINR through diversity combining methods. Either
`simple 2x spreading or space-frequency coding (SFC) maybe used as the diversity
`combining method. The partitions are spaced at the extremes of the RF channel to
`maximize the spectral diversity and may be power boosted.
`
`3.3
`
` Access Sub-Channels
`
`At least one access partition is required for each 5 MHz channel. In addition, sectorized
`base stations provision at least one access partition per sector. For the case where the RF
`band has been divided into sub-bands, at least one access partition is provisioned per sub-
`band.
`
`The access partition is contention based. If collisions occur, SSs use a random back-off
`algorithm to randomize retry timing. By using the coding methods described latter in this
`document, AAS base stations are able to spatially separate subscriber stations thus
`minimizing contention, and linearly increasing the number of logical access partitions in
`proportion to the number of spatially processed antennas.
`
`3.4
`
` Forward Link Signals
`
`The forward link partition is shown in Figure 3 for one of 84 bins. The forward
`link partition in this example provides 33 or 32 symbol slots and is organized as five
`
`5
`
`ERIC-1017
`Page 5 of 16
`
`

`
`2004-03-15
`
`IEEE C802.16-04/51r1
`
`AAS sub-channels. One of the 5 AAS sub-channels contains three forward link control
`signals once every multi-frame.
`
`There are three types of AAS control signals used by the forward link. The first is the
`forward link initiation (FLI) signal. The FLI signals to the SS to initiate communications
`on traffic sub-channels. This “paging” and “link initiation” signal is shown for the
`downlink frame structure shown in Figure 3 and has coding unique to a SS. One or two
`FLI signals are provisioned per AAS signaling sub-channel in every other or every forth
`frame. Each FLI signal modulates 16 tones (1 bin x 2 symbol times) with 16 QPSK
`symbols. The FLI provides 12 dB of processing gain to signal subscriber stations
`through all antennas without directed beam steering knowledge.
`
`
`PreamblePreamble
`
`Preamble #2Preamble #2
`
`
`
`One of two 32 Symbol Frames ShownOne of two 32 Symbol Frames Shown
`
`
`2 FLI (16 Symbols)2 FLI (16 Symbols)
`
`1 FLT (16 Symbols)1 FLT (16 Symbols)
`
`Use every other frameUse every other frame
`
`
`
`4 or 5 Subchannels4 or 5 Subchannels
`
`
`
`3 BTC FEC or 9 RS CC FEC Blocks every 2 Frames3 BTC FEC or 9 RS CC FEC Blocks every 2 Frames
`
`Forward Link Frame Structure showing FLI and FLT
`Figure 3
`The forward link training (FLT) signal occupies the 2 bins located after the two FLI
`signals. The FLT transmits a known training sequence unique to the SS so that an SS can
`estimate and update the vector channel response. The FLT is sent in TDD systems with
`full beamforming gain. Multiple SSs may be trained on the same sub-channel during the
`same time slot.
`
`The third PHY layer control signal is the forward link access (FLA) signal. The base
`uses the FLA signal followed by the user code number (identifies which RLI and FLI
`codes to recognize) and map data to direct SSs to start traffic flows. Flows start by
`transmitting RLI signals in the specified sub-channels. The FLA is transmitted with full
`beamforming gain and interference cancellation. Moreover, since the FLA is sent in
`response to an RLA, an estimate of channel quality derived from the RLA is available at
`the base. Thus, the FLA frame may be used to convey initial modulation burst
`parameters in the uplink. Similar information is conveyed in the RLA message. In this
`case, initial channel quality parameters are derived from the forward link synchronization
`(FLS) preamble.
`
`6
`
`ERIC-1017
`Page 6 of 16
`
`

`
`2004-03-15
`
`IEEE C802.16-04/51r1
`
`3.5
`
`Initial Ranging, the RLA and FLA
`
`If the RU is not yet registered with the base station and hence, does not know the proper
`timing for reverse link transmissions, it randomly chooses a ranging access code, sends a
`RLA message, detects a FLA response from the base, then adjusts its delay and transmit
`power based iteratively until an FLA is detected with maximum strength. This process is
`repeated until the best delay and transmit power have been identified. Once this has been
`accomplished, other mechanisms manage the transmit window time. The RU uses the
`average power level derived from forward link preamble measurements to set its initial
`transmit power level during initial ranging.
`
`3.6 Forward Link Synchronization Preamble, the FLS
`
`The base sends forward link synchronization (FLS) preambles that the SS uses for
`synchronization in time and frequency with the base. In addition, the SS also computes
`the average received signal strength of FLS preambles to determine the path loss between
`the base and SS. As shown in Figure 2, each forward link frame has FLS signals in the
`first and second time slot. Multiple FLS bursts from adjacent frames maybe used to
`increase synchronization accuracy.
`The FLS symbols are spread in frequency by K times, where 4 ≤ K ≤ 8, to provide a
`robust means to increase the time-bandwidth product of the signal and to remove
`competing interference from other base stations FLS preambles. The FLS can be used in
`cellularized layouts with a frequency re-use of 1 and maintain the requisite accuracy to
`support 64 QAM constellations. The FLS is unique to a base station to within a
`conservative 12 to 1 frequency reuse pattern.
`
`3.7 PHY Layer Control Signal Sequencing
`
`Having defined the control signaling above, the controlling sequences can now be
`described. The AAS physical layer is controlled via the signaling sequences described
`below. Table 1 provides a list of sequence actions keyed to the sequence diagram shown
`in Figure 4. For the first case, we consider a base station initiated data flows.
`
`7
`
`ERIC-1017
`Page 7 of 16
`
`

`
`2004-03-15
`
`IEEE C802.16-04/51r1
`
`Table 1
`
`Base Initiated Data Flows
`
`The base station uses the assigned SS access
`code to open sub-channel(s) to a SS:
`1. Base station sends the FLI of the SS being
`addressed in the intended sub-channel(s).
`2. SS looks for its assigned FLI in all sub-channels. When it
`receives a FLI in a sub-channel , it starts transmitting its
`RLI in the next reverse link time slot, followed by
`data in the sub-channel.
`3. When base station receives the RLI, it performs the
`necessary training for both RL and FL
`directions. A beam is formed and the link is
`established.
`4. Base station transmits FLT in forward link time slot
`and user data in the subsequent sub-channel
`5. The (RLI+Data, FLT+Data) exchange continues as
`long as the sub-channel is open. A field in the FL frame
`header lets the base station tell he SS to maintain or
`close a partition.
`6. When told to close a sub-channel , SS stops transmitting
`RLI+Data, and turns on FLI detect for that sub-channel.
`
`
`
`The diagram on the right side of Figure 4 also illustrates the SS initiated connection. In
`this case an RLA at step 0 is sent to the base station. The control sequence then is
`identical to the base initiated connection. The base station has the option of sending an
`FLA at step 1 instead of the FLI(s) if burst parameters need to be updated.
`
`8
`
`ERIC-1017
`Page 8 of 16
`
`

`
`2004-03-15
`
`IEEE C802.16-04/51r1
`
`
`
`SSSS
`
`
`
`BaseBase
`
`
`
`SSSS
`
`
`
`BaseBase
`
`
`
`11
`
`34
`34
`
`
`
`RLARLA
`
`
`
`F L IF L I
`
`
`
`RLI + DataRLI + Data
`
`
`
`F LT + D a t aF LT + D a t a
`
`
`RLI + DataRLI + Data
`
`55
`
`F LT + D a t aF LT + D a t a
`
`
`
`66
`
`
`SS initiatedSS initiated
`
`connectionconnection
`
`
`
`00
`
`
`
`22
`
`
`
`11
`
`34
`34
`
`
`
`22
`
`
`
`F L IF L I
`
`
`
`RLI + DataRLI + Data
`
`
`
`F LT + D a t aF LT + D a t a
`
`
`RLI + DataRLI + Data
`
`55
`
`F LT + D a t aF LT + D a t a
`
`
`
`66
`
`
`Base initiatedBase initiated
`
`connectionconnection
`
`Figure 4
`
`PHY Control Signal Sequence Diagrams
`
`3.8 Granularity
`
`In the illustrated multi-frame structure, a SS is allocated a continuous set of AAS sub-
`channels spanning 2 frames (10 msec). The following table tabulates the granularity of
`bandwidth allocation in this scenario with forward and reverse link asymmetry parameter
`set to 50%.
`
`Table 2 Bandwidth Granularity with AAS
`Modulation Scheme Bytes/Sub-Channel
`B y t e s / 1 0 m s e c
`asymmetry)
`36
`54
`72
`108
`108
`144
`162
`
`QPSK _
`QPSK _
`16QAM _
`16QAM _
`64QAM _
`64QAM 2/3
`64QAM _
`
`6
`9
`12
`18
`18
`24
`27
`
`( 5 0 %
`
`Note
`
`9
`
`ERIC-1017
`Page 9 of 16
`
`

`
`2004-03-15
`
`IEEE C802.16-04/51r1
`
`3.9
`
`Information Elements
`
`Add following information element in SICH for Forward and Reverse link framing in
`AAS Mode
`Item
`
`Note
`
`Size
`
`AAS Frame_Structure
`
`2 bits
`
`00 - DL: 32 Symbols, UL:
`15 Symbols
`
`01- DL: 24 Symbols, UL:
`24 Symbols
`
`02 - UL/DL frames are not
`separated and dynamically
`allocated (only non-AAS
`system)
`
`03- Reserved
`
`In this example, the location of access sub-channel is shown on 2 extreme frequencies. It
`is proposed to make it configurable parameter, which will be broadcasted in SICH. This
`allows BS select the location and number of access sub-channels.
`
`Item
`Number of Access Subchannels
`Location of Access SubChan1 Group
`Location of Access SubChan2 Group
`
`Size
`2 bit
`8 bits
`8 bits
`
`Note
`
`Location in frequency domain
`Location in frequency domain
`
`The reverse link frame in Figure 1 shows 8 bins used for reverse link training and a
`multi-frame of 2. Following information element is sent along with SICH for other
`combinations.
`
`Item
`No. Reverse Link Training
`Bins
`Number of FLI/frame
`
`Size
`2 bits
`
`1 bits
`
`AAS Training Periodicity
`
`2 bits
`
`Comment
`Bins used for reverse link
`training 2, 4, 6, 8
`0 one FLI/frame
`1 two FLIs/frame
`1, higher mobility,
`2, fixed, portable, mobile
`3, fixed wireless
`4, reserved
`
`10
`
`ERIC-1017
`Page 10 of 16
`
`

`
`2004-03-15
`
`IEEE C802.16-04/51r1
`
`3.10 Use of PHY Channel Signaling along with existing DL-MAP/UL-MAP
`
`AAS PHY signaling proposed here are for training the SS and BS. Allocation of BW
`is still done using DL-MAP/UL-MAP mechanism currently exists in standards. Also, use
`of mini MAP proposed in 1x Scalable PHY is proposed here in unicast mode to
`communicate to an individual SS.
`
`4 PHY Control Signaling and Coding Structure
`
`The following paragraphs described the details of the AAS control signals.
`
`4.1 RLI and RLA code properties
`
`The RLI and RLA PHY control signals are based upon a compact 64 QPSK symbol
`message constructed from Hadamard sequences. The properties of these signals are as
`follows:
`• Provides a spatial training sequence for up to 12 antennas with the appropriate
`time bandwidth product
`• Provides unique SS identification at the base station. Both signals are detected
`with beamforming gain
`• Provides a fine ranging structure within the symbol modulation
`• 8064 codes are available based on 64 symbols
`• High probability of detection, low false alarm rate consistent with modest cross-
`correlation properties between assigned codes at various code delays
`• The same codes may be re-used multiple times at the base station if sectors or
`sub-bands are used
`• Robust code reuse factor of 4 between base stations. Further code de-correlation
`occurs for distance base stations due to base station to base station range
`differences
`• The base station can separate multiple SS on the access sub-channel using
`different RLAs
`
`4.2 RLI and RLA code construction
`
`Each SS registered to a base is assigned a unique traffic access code (RLI or RLA).
`The access code may be reused from sub-band to sub-band or reused from sector to
`sector. A database in maintained which binds the access code with the SS identification
`number. Thus, within a given sub-band or sector, each SS has its own unique access
`code. There are a maximum of 8064 access codes. The access codes, a = 2016t + c, are
`
`11
`
`ERIC-1017
`Page 11 of 16
`
`

`
`2004-03-15
`
`IEEE C802.16-04/51r1
`
`divided into four equal sets; 0 ≤ t < 4, where t is the base descriptor code. Each set of
`2016 access codes are divided into three types with each type allocated a certain number
`of access codes: there are 2000 traffic access codes, c, for assigned SSs: 0 < c < 1999,
`there are 8 access codes, c, for SS initial registration: 2000 < c < 2007, and there are 8
`access codes, c, for SS initial ranging: 2008 < c < 2015.
`
`RLI and RLA codewords are based on Hadamard basis functions. RLIs are described
`01iip , contains 64 QPSK symbols
`by an access code, a,
`. A RLI codeword,
`0
`8064
`<≤ a
`and has in-phase and quadrature components taken from the columns of a 64 by 64
`Hadamard matrix,
`
`
`
`p
`
`ii
`01
`
`=
`
`
`
`A hF
`1
`
`i
`1
`
`+
`
`
`
`jA hF
`1
`
`i
`
`0
`
`p
`
`ii
`23
`
`=
`
`
`
`A hF
`2
`
`i
`1
`
`+
`
`
`
`jA hF
`2
`
`i
`
`0
`
`,
`
`,
`
`i
`1
`
`≠
`
`i
`0
`
`i
`3
`
`≠
`
`i
`2
`
`0
`
`where,
`h
`h
`and i
`are different columns from the Hadamard matrix, A is an amplitude scaling
`i
`1
`factor and F1 and F2 are toggling matrices. The indices i3, i2, i1 and i0 select a particular
`RLI code. For a given access code, a, the zero-based column indices are,
`i =
`1
`
`)64,
`
`mod(
`a
`(
`a
`
`
`For two given column indices, the access code is,
`(
`)
`i
`
`
`
`i
`
`0
`
`=
`
`mod
`
`64/
`
`i
`++
`1
`
`
`
`a
`
`=
`
`64
`
`mod
`
`i
`+−
`1
`
`0
`
`64,63
`
`+
`
`i
`1
`
`.
`
`)64,1
`
`.
`
`12
`
`ERIC-1017
`Page 12 of 16
`
`

`
`2004-03-15
`
`IEEE C802.16-04/51r1
`
`4.3 FLI, FLA and FLT code properties
`
`The FLI, FLT and FLA control signals are based upon a compact 16 QPSK tones (8
`tones/symbols, 2 symbols) message constructed from Kronecker products. The
`properties of these signals are as follows:
`• The FLT provides a vector channel training sequence for up to 4 degrees of
`freedom with the appropriate time bandwidth product.
`• The FLT and FLA are directed transmissions and benefit from beamforming
`• The FLI transmission uses random beam diversity principles
`• The FLT, FLA and FLI are uniquely coded and assigned to the SS by the base
`station.
`• 8064 codes are available based on 16 tones (8 tones/symbol, 2 symbols)
`• High probability of detection, low false alarm rate consistent with modest cross-
`correlation properties between assigned codes at various code delays
`• The same codes may be re-used multiple times at base station if sectors or sub-
`bands are used
`• Robust code reuse factor of 4 between base stations. Code de-correlation occurs
`for distance base stations due to base station to base station range differences
`• The FLI does identify which base is sending the FLI via recognition of the base
`descriptor code.
`
`4.4 FLI, FLT and FLA code construction
`
`Each SS registered with a base is assigned a unique link initiation and training code
`(FLI, FLT or FLA). Coding is the same for the FLI, FLT, and FLA.
`The modulation on each tone of a FLI message is QPSK and thus can be represented
`by two bits of information. Each FLI message is described in a compact format by 32
`bits: 16 tones by 2 bits per tone. A table can be used to represent these compact
`codewords. Table 2 lists Matlab that can be used to convert a compact codeword into an
`FLI modulation sequence.
`
`Table 2. Matlab code to generate forward link codewords.
`
`In the FLI codeword directory:
`
`fli_new_codes.m makes the compact codeword and outputs it to
`fli_new_codes_cx_results.m. This takes about 28 hours to find a compatible
`set of codewords.
`fli_new_sort.m orders the codewords so that the best set consists of those with
`a small access code.
`The sorted compact codeword table is
`fli_new_codes_cx_sorted.m
`make_fli_new.m is a matlab routine that returns a specific FLI codeword vector
`from an existing compact codeword table, fli_new_codes_cx_sorted.m
`fli_new_make_c.m converts the compact codeword table into “c” files.
`
`13
`
`ERIC-1017
`Page 13 of 16
`
`

`
`2004-03-15
`
`IEEE C802.16-04/51r1
`
`5 AAS Synchronization with Interference Cancellation
`
`The following paragraphs describe the AAS synchronization preamble properties and
`construction.
`5.1
` Synchronization code properties
`
`The FLS preambles are based upon a compact 32 BPSK symbol modulating a cluster
`of 2 bins by 2 symbol slots. The FLS sequence is constructed by adaptively optimizing
`sequences based on the following properties:
`• The FLS provides a preamble structure permitting SSs to rapidly acquire
`frequency, time, and frame, and multi-frame synchronization with the base station
`• The FLS codewords are selected so as to maximize the probability that the SS will
`lock onto the correct base, at the correct multi-frame sequence, and at the correct
`frequency.
`• The FLS provides a preamble sequence with up to K degrees of freedom to
`enhance SINR and reduce cross-correlation interference via adaptive combining.
`• The FLS is transmitted with constant power so that the SS can estimate path loss
`and reverse link transmit power.
`• The pattern of FLS codewords is unique within a multi-frame and repeats from
`multi-frame to multi-frame.
`• The FLS transmission uses beam diversity principles. Each FLS coding cluster,
`defined as a 2 x 2 cluster of adjacent bins uses a different beam position in the
`same time epoch.
`• 12 unique FLS sequences indexed by base offset code are available. Reducing the
`number to 12 allows rapid handover for mobile users.
`• The same FLS code is re-used multiple times at a base station if sectors or sub-
`bands are used
`• Robust code reuse factor of 12 between base stations.
`
`5.2
`
` FLS code construction
`
`The FLS preamble is a constant modulus BPSK code unique to a given base station.
`The code has nonlinear phase and is uncorrelated to the codes used by the other bases.
`Furthermore, the codeword in the second FLS burst does not resemble a complex scalar
`multiplying the codeword in the first FLS burst. For the 1x scalable configuration, a code
`of length 32 is sufficient. The code is split into two codewords for the two FLS bursts in
`
`14
`
`ERIC-1017
`Page 14 of 16
`
`

`
`2004-03-15
`
`IEEE C802.16-04/51r1
`
`a forward link preamble slot. Each length 16 codewords modulates two adjacent bins (a
`bin pair).
`The 32-element vector containing the code is multiplied by a pseudo random complex
`scalar for each of the K spread FLS bin pairs. For each FLS bin pair, the resulting 32
`complex gain elements are split between the consecutive FLS bursts. The FLS of the first
`burst has the first 16 complex elements and the FLS of the second burst has the second 16
`complex elements. The base then transmits the code over the assigned FLS bins pairs.
`The code is received at the SS in its corresponding FLS bin pairs with an unknown
`modulation frequency due to frequency offsets. For initial acquisition, a frequency offset
`is estimated by applying an objective function to frequency shifted and time shifted
`versions of the FLS data. Once the initial search is completed, a tracking frequency
`estimate is obtained by measuring the phase change between bursts.
`
`An AAS base selects random transmit weight vectors for FLS messages for each bin
`pair and spreading location. Each element of each transmit weight vector has the same
`amplitude and a randomly selected phase. The random transmit weight vectors are used
`so that with high probability, at least one of them has a main lobe in the direction of each
`SS. The random number generator use at one base is not be correlated with or have the
`same repeat period as the generator of another bin pair of any base with a different base
`offset code.
`Every base uses a particular set of FLS codewords. The base offset code associated
`with the base forms part of the FLS codewords used by that base. Figure 5 gives an
`example of a hexagonal layout of N cells. The numeral in each cell is its base offset
`code. Figure 6 gives an example of a rectangular layout of N cells. FLS codeword
`sequences, like the base offset codes, may be reused every 12 cells.
`
`6
`
`0
`
`7
`
`10
`
`9
`
`8
`
`3
`
`2
`
`1
`
`7
`
`10
`
`8
`
`11
`
`6
`
`1
`
`5
`
`4
`
`11
`
`6
`
`0
`
`2
`
`3
`
`9
`
`5
`
`1
`
`7
`
`0
`
`4
`
`10
`
`3
`
`2
`
`9
`
`8
`
`7
`
`10
`
`8
`
`11
`
`6
`
`1
`
`5
`
`11
`
`4
`
`0
`
`6
`
`2
`
`3
`
`9
`
`5
`
`1
`
`7
`
`0
`
`4
`
`10
`
`3
`
`2
`
`9
`
`8
`
`7
`
`10
`
`4
`
`11
`
`1
`
`5
`
`4
`
`0
`
`4
`
`7
`
`1
`
`9
`
`8
`
`6
`
`10
`
`3
`
`2
`
`1
`
`5
`
`7
`
`10
`
`8
`
`11
`
`6
`
`4
`
`2
`
`5
`
`11
`
`6
`
`9
`
`7
`
`0
`
`3
`
`1
`
`0
`
`4
`
`10
`
`3
`
`2
`
`9
`
`8
`
`1
`
`5
`
`7
`
`10
`
`8
`
`11
`
`6
`
`9
`
`7
`
`10
`
`4
`
`2
`
`5
`
`11
`
`6
`
`9
`
`7
`
`0
`
`3
`
`1
`
`4
`
`11
`
`0
`
`4
`
`3
`
`10
`
`8
`
`2
`
`1
`
`5
`
`Figure 5. Base offset codes for a re-use factor of 12, hexagonal layout.
`
`15
`
`ERIC-1017
`Page 15 of 16
`
`

`
`2004-03-15
`
`IEEE C802.16-04/51r1
`
`3
`
`2
`
`2
`
`3
`
`2
`
`3
`
`3
`
`2
`
`0
`
`1
`
`1
`
`0
`
`0
`
`1
`
`1
`
`0
`
`3
`
`2
`
`0
`
`1
`
`1
`
`0
`
`0
`
`1
`
`2
`
`0
`
`2
`
`3
`
`2
`
`3
`
`2
`
`2
`
`3
`
`2
`
`3
`
`3
`
`2
`
`0
`
`1
`
`1
`
`0
`
`0
`
`1
`
`1
`
`0
`
`3
`
`2
`
`0
`
`1
`
`1
`
`0
`
`2
`
`3
`
`2
`
`3
`
`1
`
`3
`
`2
`
`2
`
`3
`
`0
`
`1
`
`1
`
`0
`
`2
`
`3
`
`2
`
`3
`
`2
`
`2
`
`3
`
`2
`
`3
`
`3
`
`2
`
`0
`
`1
`
`1
`
`0
`
`1
`
`0
`
`3
`
`2
`
`0
`
`1
`
`0
`
`1
`
`3
`
`1
`
`0
`
`3
`
`2
`
`2
`
`3
`
`2
`
`3
`
`3
`
`2
`
`0
`
`1
`
`1
`
`0
`
`1
`
`0
`
`3
`
`2
`
`0
`
`1
`
`0
`
`1
`
`3
`
`1
`
`0
`
`3
`
`2
`
`2
`
`3
`
`0
`
`1
`
`3
`
`1
`
`0
`
`1
`
`3
`
`1
`
`0
`
`2
`
`0
`
`Base color codes: red , green , and blue .
`
`Figure 6. Base descriptor codes have a repeat factor of 4.
`
` Codes will be generated and provided for the 1x Scalable PHY
`
`FLS Codewords
`
`6 Text to be included in the standard
`
`The AAS control signal overlay is coordinated with the 1X Scalable PHY definition.
`Text will be provided
`
`16
`
`ERIC-1017
`Page 16 of 16