Signalling for Coordinated Multi-Point Transmission and Reception
(CoMP)
[0001] This application claims the benefit of
U.S. Provisional Application No. 61/955,559, entitled "Signaling Considerations for
Inter-eNB CoMP," filed on March 19, 2014,
U.S. Provisional Application No. 61/991,055, entitled "Signaling Considerations for
NAICS," filed on May 9, 2014,
U.S. Provisional Application No. 61/991,323, entitled "Signaling Considerations for
NAICS," filed on May 9, 2014,
U.S. Provisional Application No. 62/034,724, entitled "X2 Signaling for Inter-eNB CoMP,"
filed on August 7, 2014,
U.S. Provisional Application No. 62/034,885, entitled "X2 Signaling for Inter-eNB CoMP,"
filed on August 8, 2014,
U.S. Provisional Application No. 62/055,381, entitled "Signalling for Inter-eNB CoMP,"
filed on September 25, 2014, and
U.S. Provisional Application No. 62/056,095, entitled "Signalling for Inter-eNB CoMP,"
filed on September 26, 2014,
the contents of all of which are incorporated herein by reference.
BACKGROUND OF THE INVENTION
[0002] The present invention relates to coordinated multi-point transmission and reception
(CoMP) in wireless or mobile communications and, more particularly, to inter-eNB
(E-UTRAN NodeB or eNodeB) CoMP with Network Assisted Interference Cancellation
and Suppression (NAICS) and/or non-ideal backhaul (NIB).
[0003] The CoMP schemes that were discussed during the 3rd Generation Partnership
Project (3GPP) Release 11 CoMP standardization assumed the availability of an ideal
backhaul connecting the transmission points in each cluster. This assumption allowed for
coordination within the cluster based on the instantaneous channel state information (CSI)
reported by the users to those transmission points. Unfortunately, such schemes are far
from being suitable when faced with a non-ideal backhaul that has a high latency. To guide
the design of schemes that are appropriate for the NIB scenario, the following agreement
was reached during 3GPP RANI (Radio Access Network Working Group 1 or Radio Layer
1) Meeting #74:
[0004] For each evaluated scheme, information relating to a transmission to/from a serving
node in a given subframe should be categorized into two groups:
- Group 1 information: information which is considered valid for a period longer than the
backhaul delay, which may therefore be provided from a different node(s) from the serving
node; and
- Group 2 information: information which is considered valid for a period shorter than the
backhaul delay, which must therefore be derived by the serving node.
[0005] The types of information may include for example:
- CSI,
- Allocated power per resource (including muting),
- User equipment (UE) selection,
- Precoding selection (including the number of transmit layers),
- Modulation and coding scheme (MCS) selection,
- Hybrid automatic repeat request (HARQ) process number, and
- Transmission point (TP) selection.
[0006] Transmission layers are sometimes called "transmit layers" or "layers." The number
of transmission layers is known as "transmission rank" or "rank." A codebook is a set of
precoding matrices or precoders. A precoding matrix is also known as a codeword.
[0007] Reference
[0008] [1] H. Zhang, L. Venturino, N. Prasad, P. Li, S. Rangarajan, X. Wang,"Weighted
Sum-Rate Maximization in Multi-Cell Networks via Coordinated Scheduling and Discrete
Power Control", IEEE Journal on Selected Areas in Communications, 29(6): pp. 1214-1224,
201 1.
[0009] [2] Rl-141816, "LS on Inter-eNB CoMP for LTE," RANI, March 31-April 4 2014.
[0010] [3] R3-141487, "CHANGE REQUEST," March 31-April 4 2014.
[001 1] [4] Rl-141206, "Signaling Considerations for Inter-eNB CoMP", NEC, March 31st
to April 4th, 2014.
BRIEF SUMMARY OF THE INVENTION
[0012] An objective of the present invention is to provide a suitable scheme for CoMP
operation.
[0013] An aspect of the present invention includes, in a wireless communications system
including a first transmission point and a second transmission point, a wireless
communications method implemented in the first transmission point supporting coordinated
multi-point transmission and reception (CoMP). The wireless communications method
comprises transmitting to the second transmission point one or more CoMP hypothesis sets,
and transmitting to the second transmission point a benefit metric corresponding to each
CoMP hypothesis set, wherein the benefit metric can be a negative value.
[0014] Another aspect of the present invention includes, in a wireless communications
system including a first transmission point and a second transmission point, a wireless
communications method implemented in the second transmission point supporting
coordinated multi-point transmission and reception (CoMP). The wireless communications
method comprises receiving from the first transmission point one or more CoMP
hypothesis sets, and receiving from the first transmission point a benefit metric
corresponding to each CoMP hypothesis set, wherein the benefit metric can be a negative
value.
[0015] Still another aspect of the present invention includes a first transmission point
supporting coordinated multi-point transmission and reception (CoMP) and used in a
wireless communications system. The first transmission point comprises a transmitter to
transmit to a second transmission point one or more CoMP hypothesis sets and a benefit
metric corresponding to each CoMP hypothesis set, wherein the benefit metric can be a
negative value.
[0016] Still another aspect of the present invention includes a second transmission point
supporting coordinated multi-point transmission and reception (CoMP) and used in a
wireless communications system. The second transmission point comprises a receiver to
receive from a first transmission point one or more CoMP hypothesis sets and a benefit
metric corresponding to each CoMP hypothesis set, wherein the benefit metric can be a
negative value.
[0017] Still another aspect of the present invention includes a wireless communications
method implemented in a wireless communications system supporting coordinated
multi-point transmission and reception (CoMP). The wireless communications method
comprises transmitting from a first transmission point to a second transmission point one or
more CoMP hypothesis sets, and transmitting from the first transmission point to the
second transmission point a benefit metric corresponding to each CoMP hypothesis set,
wherein the benefit metric can be a negative value.
[0018] Still another aspect of the present invention includes a wireless communications
system supporting coordinated multi-point transmission and reception (CoMP). The
wireless communications system comprises a first transmission point, and a second
transmission point to receive form the first transmission point one or more CoMP
hypothesis sets, wherein the first transmission point transmits to the second transmission
point a benefit metric corresponding to each CoMP hypothesis set, and wherein the benefit
metric can be a negative value.
[0019] Still another aspect of the present invention includes a wireless communications
method implemented in a transmission point (TP) used in a wireless communications
system. The wireless communications method comprises receiving, from another TP,
channel state information (CSI) for a user equipment (UE), and receiving, from said
another TP, user identification for the user equipment, wherein the signaling of the CSI for
the user equipment enables user identification for the user equipment.
[0020] Still another aspect of the present invention includes a wireless communications
method implemented in a transmission point (TP) used in a wireless communications
system. The wireless communications method comprises transmitting, to another TP,
channel state information (CSI) for a user equipment (UE), and transmitting, to said another
TP, user identification for the user equipment, wherein the signaling of the CSI for the user
equipment enables user identification for the user equipment.
[0021] Still another aspect of the present invention includes a transmission point (TP) used
in a wireless communications system. The TP comprises a receiver to receive, from another
TP, channel state information (CSI) for a user equipment (UE) and user identification for
the user equipment, wherein the signaling of the CSI for the user equipment enables user
identification for the user equipment.
[0022] Still another aspect of the present invention includes a transmission point (TP) used
in a wireless communications system. The TP comprises a transmitter to transmit, to
another TP, channel state information (CSI) for a user equipment (UE) and user
identification for the user equipment, wherein the signaling of the CSI for the user
equipment enables user identification for the user equipment.
[0023] Still another aspect of the present invention includes a wireless communications
method implemented in a wireless communications system. The wireless communications
method comprises transmitting, from a transmission point (TP) to another TP, channel state
information (CSI) for a user equipment (UE), and transmitting, from the transmission point
(TP) to said another TP, user identification for the user equipment, wherein the signaling of
the CSI for the user equipment enables user identification for the user equipment.
[0024] Still another aspect of the present invention includes a wireless communications
system comprising a first transmission point (TP), and a second transmission point (TP) to
transmit to the first TP, channel state information (CSI) for a user equipment (UE) and user
identification for the user equipment, wherein the signaling of the CSI for the user
equipment enables user identification for the user equipment.
BRIEF DESCRIPTION OF THE DRAWINGS
[0025] FIG. 1 depicts a block diagram of a CoMP system.
[0026] FIG. 2 depicts a CoMP coordination request under CoMP-NIB implementation.
[0027] FIG. 3(a) depicts an example of centralized CoMP coordination via CoMP
hypothesis and Benefit metric over X2.
[0028] FIG. 3(b) depicts an example of centralized CoMP coordination via CoMP
hypothesis and Benefit metric over X2. Note here that the BM is used to convey the utility
change for a particular resource allocation indicated in the associated CH to the Master
node. The CH sent by the Master node contains the resource allocation decision.
[0029] FIG. 4 depicts an example of distributed CoMP coordination via CoMP hypothesis
and Benefit metric over X2.
[0030] FIG. 5 depicts that if only "gains" can be conveyed via benefit metric, eNB2 may
not obtain the information about the loss it can cause to eNBl by increasing its power.
Consequently, such an increase in power would have to be done unilaterally by eNB2
which is undesirable.
DETAILED DESCRIPTION
[0031] Referring now to FIG. 1, a CoMP mobile communications system 400 comprising a
CoMP coordination zone or area or CoMP cooperating set 402 in which the embodiments
may be implemented is illustrated. One or more user equipments 410 are served by one or
more TPs or cells 404 to 408. TPs 404 to 408 can be base stations or eNBs. Each of the
user equipments includes e.g. a transmitter and a receiver, and each of the base stations or
eNBs 104 includes e.g. a transmitter and a receiver.
[0032] Embodiment A
[0033] We have captured the details of the scheduling framework in the appendix. We
assume that for each user a measurement set containing up-to three TPs among those in the
coordination zone is defined and held fixed for a time scale even coarser than the one at
which the centralized decisions (precoder tuple or muting pattern assignment and possibly
user association) are made.
[0034] From the description given in the appendix, we see that to determine the centralized
decisions (such as the precoder tuple assignment and the user associations) under the full
buffer traffic model , the designated central node (referred to here as the master TP (MTP))
should be able to obtain, R" (W^' , which we recall denotes an estimate of the average rate
that user u can obtain (over the available time-frequency resource normalized to have size
unity) when it is served data by TP b , given that the precoder tuple assigned to the
TPs in the zone and that no other user is associated with TP b . Recall also that the
precoder tuple can also correspond to a muting pattern deciding which TPs should be
active and which should be turned off in the time-frequency unit. For the joint semi-static
point muting (SSPM) and semi-static point switching (SSPS) scheme (cf. (PI) in the
appendix), this average estimate must be obtained for each user u , each TP b
its measurement set and for all precoder tuple assignments. Note that for any precoder tuple,
R ) can e considered to be negligible if the TP b n o t the measurement set of
user u . Notice also that ' can be assumed to be equal to ' for any two
precoder tuple assignments and ' which differ only in precoders assigned to TPs
not in the measurement set of user u . For the SSPM problem (cf. (P2) in the appendix)
with pre-determined user associations, the average estimate R (W^' must be obtained for
each user u only for its pre-determined serving TP b the set of users associated to
that TP must also be obtained. Thus, the following types of backhaul signaling are needed
to facilitate a centralized implementation.
[0035] Al. Backhaul signaling to enable determination of centralized actions (such
as precoder tuple/muting pattern assignments and the user associations)
R b ( )
[0036] We will now consider computation of the average rate estimates at the
MTP for some user , under a precoder tuple assignment . These rates depend on the
channels that the user sees from TPs in its measurement set. Using up-to three CSl
processes (recall that the maximum measurement set size is three) which include a common
interference measurement resource (IMR), the UE can report short-term CSl for each TP
in its measurement set, where this short-term CSl is computed based on the non-zero
CSI-reference signal (RS) transmitted by TP and the interference observed on the IMR,
which in turn includes only the interference from TPs not in the measurement set of user u .
The UE currently reports such CSl only to its designated anchor TP.
[0037] However, to fully exploit point switching gains we need to allow for the possibility
of associating a user to a non-anchor TP and then allowing that user to report instantaneous
(short-term) CSl to the non-anchor TP it has been associated to. Further, the CSl processes
should be defined in a coordinated manner so that the users measure the appropriate
interference on the constituent IMRs. Such coordinated configuration of IMRs also
provides the ability to inject the desired interference (such as isotropically distributed
interference) onto resource elements in those IMRs.
[0038] These short-term CSl can be sent to the MTP over the backhaul, which can then
filter (i.e. perform a weighted average of) the received CSl sequence to obtain an averaged
H b
channel estimate " for each TP " in the measurement set of user u . Alternatively, the
averaging (or subsampling) of the short-term CSl can be done by the TP receiving the
short-term CSl but where the averaging window (and possibly the weighting factors or
subsampling factors) can be configured for that UE on a per CSI-process basis.
[0039] In either case, these averaged or subsampled channel estimates for all TPs in that
UE's measurement set can be used by the MTP to compute ' for each precoder
tuple hypothesis and if needed each TP in its measurement set, under the
assumption that the signal transmitted by each TP (along its assigned precoder under that
hypothesis) is isotropically distributed. Another option is for the MTP to directly compute
an estimate of the rate using each received short-term CSI and then average these computed
rates to obtain an estimate of the average rate. We note that in case each precoder tuple
hypothesis is a muting pattern, the average rate estimates can be computed using only the
average received powers observed by each user from each TP in its measurement set. In
such a case only reference signal received powers (RSRPs) need to be exchanged for a
configurable set of users over the backhaul.
[0040] Moreover, the signaling of CSI (which can be RSRP) over the backhaul should
enable the identification of the users whose CSI are being signaled as well as the attributes
(such as zero-power CSI-RS or non-zero-power CSI-RS) of the corresponding CSI
processes. Recall also that in the scenario with pre-determined users associations, the set of
users associated to each TP in the zone needs to be exchanged or conveyed to the MTP.
[0041] These views are summarized in the following proposal.
[0042] Proposal: Signaling of averaged or subsampled CSI obtained over CSI processes
corresponding to a configurable set of users should be considered. Coordination in
configuring these CSI processes should be allowed.
[0043] Proposal: Possibility of configuring a user to report short-term CSI to more than
one TP or a chosen TP in a configurable set of TPs should be considered.
[0044] Next, in the more general finite buffer model estimates of the queue sizes are
needed to determine each coarse (centralized) action, where each such user queue size
represents the amount of traffic that would available for transmission to serve that user until
the next coarse action. Determining estimates of these queue sizes requires the TPs to
report their most-recently updated associated user queue sizes before the next coarse action
to the MTP.
[0045] Proposal: Signaling of associated user queue sizes by a TP to another TP should be
considered, possibly by enhancing the status report.
[0046] A2. Backhaul Signaling from MTP to TPs
[0047] Each TP in the coordination zone is informed (semi-statically) about the precoder it
should use and possibly the users it should serve on a time-frequency resource. The
decision made by the MTP can be represented using a CoMP hypothesis. This can be
achieved for instance, by assigning an identifier to each TP in the coordination zone and
then including pairs representing (TP identifier, corresponding part of decision) in the
CoMP hypothesis. Each TP then implements its own per-subframe scheduling based on the
instantaneous CSI it receives from the users associated to it. Some comments on the set^
which contains the set of precoders that can be assigned to each TP, are on order. We recall
that this set includes codeword 0 to subsume muting as a special case. It can also include
codewords of the form I where a denotes a positive power level. In addition, it can
include sector beams as its codewords. Notice that so far we have implicitly assumed that
each TP will accept the decision made by the MTP. This assumption need not always hold,
in which case it is beneficial (even necessary) to have an acknowledgement from the
receiving TP conveying whether or not it accepts to implement its part of the decision in the
CoMP hypothesis.
[0048] Note that since the decision represented by the CoMP Hypothesis should be valid
for a period longer than the (maximum) backhaul delay. Henceforth we will refer to the
time period over which a CoMP hypothesis is supposed to be valid (or supposed to apply)
as a frame. Thus, the CoMP hypothesis should be signaled at a time granularity (i.e., the
time interval between successive CoMP hypotheses) that is a multiple of the largest
backhaul delay. Note that it in some scenarios it may be preferable for the MTP to receive
the acknowledgement, in which case the multiple should be at-least 2. A small value of this
multiple would help the system adapt faster, so we suggest a value for this multiple that is
less than or equal to 3.
[0049] Proposal: Signaling of decisions made by one TP (such as precoder set or muting
pattern assignment) to all other TPs over the backhaul should be considered. Such a
decision can be represented by a CoMP Hypothesis. Signaling of an acknowledgement
conveying a yes/no response to a received CoMP hypothesis should be considered.
[0050] A3. Distributed Implementation
[0051] In order to enable a de-centralized or distributed operation, a benefit metric
corresponding to each CoMP hypothesis can be defined. In [1] a distributed implementation
of power control is provided. An example distributed operation considering binary power
control is described next and we note that extension to multiple power levels can be
developed following the same approach. Each TP b in the coordination set can determine
its set of interfering TPs, where a TP is labelled interfering for TP b if it is in the
measurement set of at-least one user associated to TP b . Note that TP b can determine its
set of interfering TPs. Further, let us refer to all TPs in whose interfering sets TP b is
present as the out neighbor set of TP b . Each CoMP hypothesis can be defined such that the
sending TP, say TP b , suggests a muting (or in general a power level) pattern for a set of
time-frequency resources to a receiving TP, say TP a, in its interfering set of TPs. The
benefit metric for that hypothesis comprises of a set gain (or loss, i.e., the gain can be
negative) values (one for each time-frequency resource), where each gain represents the
incremental average throughput or utility that would be achieved for the sending node (TP
b) if the receiving node (TP a) accepts the suggested muting or power level (henceforth
termed suggested action) on that time-frequency resource, while the other TPs in the
interfering set of TP b as well as TP b do not alter their current status (current power level).
TP a can then consider each time-frequency resource and add up all the gain values it has
received for each suggested action on that resource. To this sum it can then add the gain (or
loss) that it would obtain upon following the suggested action, assuming that all TPs in its
interfering set do not alter their current status. This sum gain for each action can then
represent the system utility gain that can be achieved by a one-step change, i.e., the
incremental throughput or utility gain for the coordination set achieved when TP a accepts
that suggested action on that resource and all the other TPs in the coordination set keep
their current respective status. TP a can then independently choose its action on each
time-frequency resource using a probabilistic rule [1], and this distributed operation can be
shown to converge. Further, the TP a can signal its choice of actions using an enhanced
RNTP. Note here that as an alternative the CoMP hypothesis can consider only one
time-frequency resource and suggest multiple actions, one for each TP in its interfering set
and the corresponding benefit metric can include a gain (or a loss) for each suggested
action. In general, the CoMP hypothesis can include multiple tuples, where each tuple
contains a TP identifier and a suggested action identifier, and one time-frequency resource
identifier that is common for all tuples in that hypothesis. Alternatively, each tuple can
include a time-frequency resource identifier and a suggested action identifier while the
hypothesis includes a TP identifier that is common across all its constituent tuples.
Combinations of these two general alternatives can also be used to define a CoMP
hypothesis. In each case the benefit metric includes a gain (or loss) for each suggested
action and a TP receiving the benefit metric must be able to determine which gain
corresponds to which suggested action.
[0052] We next discuss efficient signaling mechanisms. First note that in order to reduce
the signaling overhead, the network can configure to allow only a subset of TPs in the
coordination set to make a change. This can be done in a de-centralized manner using a
pre-determined function (known to all TPs in the coordination set), where this function
returns the indices (or identifiers) of all TPs that are permitted to make a change, given the
frame or sub-frame index as input. Alternatively, a designated TP can convey the set of TPs
that are permitted to make a change, to all the other TPs in the coordination set, at the start
of each frame. In either case, a TP b will send one or more CoMP hypothesis for TP a and
corresponding benefit metrics, only if TP a is in its interfering set and TP a is in the set of
TPs that are permitted to make a change on that frame. Further, the cardinality of the
aforementioned set of TPs can be used to control the backhaul signaling overhead, as well
as the size of the enhanced relative narrowband TX power (RNTP) which is used by each
TP in that set to convey its actions to the other TPs. Note that each TP which changes its
action on a time-frequency resource must report its changed action only to TPs in its out
neighbor set.
[0053] Note that the distributed procedure described above can be implemented
independently on each time-frequency resource. Then, the set of time frequency resources
on which TPs can change their actions in a frame can also be controlled to reduce the
signaling overhead. This can be accomplished as before, for instance by defining a rule
using the frame index (known to all TPs in the coordination set) to decide the set of
time-frequency resources at the start of each frame. A combination is also possible where in
each frame a set of TPs which are permitted to change their actions and a set of
time-frequency resources on which those TPs can change their actions is identified for each
frame.
[0054] The configuration (or identification) of these sets can instead be done at a time-scale
coarser than the frame duration, i.e. once in every n frames, where n is configurable. We
have assumed that the set of TPs permitted to change their actions is the same across all
time-frequency resources in the set of such resources. A more general approach would be to
configure a separate set of TPs for each time-frequency resource. Here a designated node
can optionally be used convey the configured sets to all other TPs.
[0055] However, a potential drawback with the distributed approach described above is if
the benefit metrics do not allow a TP to infer (a good approximation of) the system utility
gain (or loss) accrued by a suggested action on a time-frequency resource, in which case
oscillatory behavior or convergence to a highly sub-optimal operating point can result. We
summarize our views in the following proposal.
[0056] Proposal: The benefit metrics received by a TP should enable it to compute a
system utility change for each action suggested for that TP in each of its received CoMP
hypothesis.
[0057] Thus, we provided our views on backhaul signaling needed for CoMP-NIB
comprising of the following proposals:
[0058] Proposal: Signaling of averaged or subsampled CSI obtained over CSI processes
corresponding to a configurable set of users should be considered. Coordination in
configuring these CSI processes should be allowed.
[0059] Proposal: Possibility of configuring a user to report short-term CSI to more than
one TP or a chosen TP in a configurable set of TPs should be considered.
[0060] Proposal: Signaling of associated user queue sizes by a TP to another TP should be
considered, possibly by enhancing the status report.
[0061] Proposal: Signaling of decisions made by one TP (such as precoder set or muting
pattern assignment) to all other TPs over the backhaul should be considered. Such a
decision can be represented by a CoMP Hypothesis. Signaling of an acknowledgement
conveying a yes/no response to a received CoMP hypothesis should be considered.
[0062] Proposal: The benefit metrics received by a TP should enable it to compute a
system utility change for each action suggested for that TP in each of its received CoMP
hypothesis.
[0063] Embodiment B
[0064] We present our views on the signalling that is appropriate to extract network
assisted interference cancellation and suppression (NAICS) gain.
[0065] We assume that a candidate list of potentially interfering cells is configured by the
network for the user of interest. For each cell in this list (identified by an index, a natural
choice of which is the corresponding cell ID) the network can specify a set of parameters.
Such a candidate list (along with its constituent parameters) should be semi-statically
configured by the network for the user in order to simplify and assist the user's blind
detection.
[0066] Bl. Signaling parameters pertaining to reference signal (RS)
[0067] Bl.l Signaling parameters associated with the cell-specific reference signal
(CRS)
[0068] We first consider the signalling needed to convey parameters associated with the
CRS transmitted by each cell in the candidate list. In our view, the number of CRS ports for
each cell in the list (and optionally its corresponding frequency shift or the multimedia
broadcast multicast service (MBMS) or single frequency network (MBSFN) sub-frame
configuration) is quite beneficial in reducing the blind detection complexity at the user of
interest. In this context, we note that the possibility of CRS not being transmitted at-all by
the interferer might also need to be considered by the user over any sub-frame in order to
incorporate dynamic cell ON-OFF. Another useful parameter is the (expected) physical
downlink shared channel (PDSCH) start symbol. The signalling of this parameter conveys
the actual (or likely) starting symbol of the interfering PDSCH and is needed to fully
exploit NAICS gain (over all transmitted interfering PDSCH symbols). Moreover, blind
detection of the starting symbol by the user appears to be quite challenging.
[0069] B1.2 Signaling CSI-RS related parameters
[0070] Next, we consider configuration parameters associated with the CSI-RS (including
both zero-power and non-zero power CSI-RS). In this case, the user upon knowing one or
more CSI-RS configurations that can be employed by each potential interferer in its list,
knows the PDSCH resource element (RE) mappings possible under each such interferer
hypothesis, which clearly will improve interference cancellation/suppression gains (for a
given feasible level of complexity).
[0071] On the other hand, signalling for quasi co-location (QCL) indication needs further
evaluation since purely demodulation reference signal (DMRS) based channel estimation
was sufficient for desired signal demodulation in several evaluated instances during 3GPP
Release 11 and it is unclear if enhanced estimation of the channel seen from an interferer is
really needed for cancellation/suppression gains.
[0072] In summary, we have the following proposal for the parameters pertaining to the
RS.
[0073] Proposal: Convey via semi-static signaling about each cell in a candidate list:
[0074] (1) Number of CRS ports and PDSCH start symbol
[0075] (2) CSI-RS configuration(s)
[0076] B2. Signalling to aid blind detection of other dynamic parameters
[0077] B2.1 Modulation Classification
[0078] We note that the joint blind detection of modulation, PMI, RI and presence of one
dominant interferer using a CRS based TM (transmission mode) has been deemed feasible
for 2 CRS ports, at-least under the simulated scenarios and provided that the other required
parameters are perfectly known. Similarly, in the case of DMRS based TM, joint blind
detection of modulation, nSCID and presence of one dominant interferer using up-to two
DMRS ports (ports 7 and 8) has been deemed feasible, again under the simulated scenarios
and provided that the other required parameters are perfectly known.
[0079] However, the evaluation so far has assumed only the three modulation types that
can be employed up-to 3GPP Release 11, i.e., quadrature phase shift keying (QPSK), 16
quadrature amplitude modulation (QAM) and 64 QAM. It is likely (or imminent) that a
higher modulation order (256 QAM) will be agreed in 3GPP Release 12. This then raises
the question about feasibility of blind detection in scenarios where 256 QAM can be
employed by the interferer. In this context, we note that applying blind modulation
classification when multiple higher order modulation types can be employed by the
interferer is more complicated (indeed the classification errors tend to be increasing with
the modulation order). Moreover, NAICS gain (even after correctly classifying an interferer
employing a higher order modulation) over the baseline interference rejection combining
(IRC) receiver will be smaller, since the IRC receiver regards interference as a
(un-constrained) Gaussian variable, an assumption that becomes increasingly suitable for
denser QAM constellations. To summarize, support of 256 QAM with NAICS needs to be
further evaluated. Our preference is thus the following.
[0080] Proposal: Blind modulation classification is done by the user assuming that QPSK,
16 QAM and 64 QAM are the modulation types that can be employed by any interferer.
[0081] It is desirable that the assumption made by the user is indeed respected by each
interferer in its candidate list, i.e., it is desirable that the network enable NAICS
functionality only in the regime where 256 QAM is not employed in a cluster of cells. In
case, this is not true, the user can itself disable its NAICS capability and fallback to IRC
based reception, following some decision rule, when it perceives degraded performance due
to operation in a scenario where 256 QAM is often employed by one or more interferers.
[0082] B2.2 Supporting 4TX
[0083] The support for 4TX is important and NAICS gain should hold for such
deployments. Let us consider the case where the dominant 4TX interferer employs a
CRS-based TM. Here, blind detection of the assigned transmit rank of the interferer among
all the four possible transmit ranks can result in an excessive complexity expended to
chase gains that become increasingly marginal for larger ranks. It is thus meaningful to
restrict the transmit rank assigned by the interferer. The user can be informed via
semi-static signaling about an upper bound to the transmit rank that can be assigned by
each potentially interfering cell in its candidate list. Alternatively, the semi-static signaling
can indicate an expected transmit rank that is likely to assigned by that interferer, which can
be used as a more probable seed value for the blind detection implementations.
[0084] Next, we suppose a dominant interferer (from the candidate list) employing a
DMRS based TM. In this case, physical resource block (PRB)-pair has been agreed as the
minimum resolution of the time-frequency unit that can be assigned by any such interferer.
[0085] Here, it is particularly beneficial if the user has to consider only ports 7 and 8 in
order to detect the presence and absence of interferer and classify the rank on each
PRB-pair, possibly by determining the norms of the columns of the corresponding
equivalent channel estimate. Recall that joint blind detection has been deemed feasible only
with such a qualification. Consequently, semi-static signaling a transmit rank upper bound
adhered to by each potential interferer is useful here as well.
[0086] Proposal: Convey via semi-static signaling about each cell in a candidate list:
[0087] An upper bound on the transmit rank that can be assigned.
[0088] B3. Other Issues
[0089] We believe that synchronization should be assumed by the user without any explicit
signaling since this is in any case the main operating regime where NAICS gain can be
achieved in a feasible manner. While, the user can itself disable its NAICS capability and
fallback to IRC based reception, following some decision rule, when it perceives degraded
performance due to operation in an asynchronous scenario, it is desirable that the network
enable NAICS functionality only in the synchronous regime.
[0090] The user can perform blind detection (classification) after assuming a certain
minimum time-frequency unit that can be assigned by an interferer under each transmission
scheme, in other words, after assuming that the parameters that it seeks to classify remain
constant within that unit. This minimum assignable time-frequency unit can be set or
assumed, for instance, to be one PRB-pair. This is a choice that is indeed accurate at-least
for DMRS based TMs and has been found to ensure reliable blind detection. One PRB-pair
for all DMRS based TMs has been found sufficient to ensure reliable blind detection. For
CRS based TMs the minimum assumed unit can be configured (by the network for the user)
to be either a slot or a PRB pair. It is beneficial with respect to NAICS gain that this
assumption is indeed respected by each interferer in the list, i.e., it is desirable that the
network enable NAICS functionality only in the regime where the respective assumed
minimum assignable time-frequency units are followed by all the cells. Then, note that
configuring the minimum assumed unit for CRS based TMs to be a slot makes blind
detection challenging but does not preclude distributed virtual resource block (DVRB)
based allocation, while configuring the minimum assumed unit to be a PRB-pair makes
blind detection more feasible but precludes DVRB based allocation. While these assumed
minimum assignable time-frequency units can be made further configurable on a
per-interferer basis for each user, i.e., the assumed minimum assignable time-frequency
units can be altered semi-statically for each cell in that user's candidate list of interferers,
further evaluation is needed to assess if this is beneficial. This is because such semi-static
configuration in the absence of any explicit scheduling restrictions will not lead to
significant NAICS gain, while placing scheduling restrictions can be counter-productive
due to the bursty nature of the traffic. In this context, we note that a significant portion of
the traffic is expected to be bursty and formed by very small per-user data demands.
[0091] Proposal: Interference cancellation/suppression is attempted by the user assuming
synchronization and a minimum time-frequency unit that can be assigned by a dominant
interferer for each transmission scheme.
[0092] We note that in case the assumed minimum assigned unit is configured to be a slot
for the CRS based TMs, it is still possible to exploit for blind detection the fact that the
minimum unit can be more than a slot (i.e., can be a PRB-pair) even under CRS based TMs
when the resource allocation is not DVRB based.
[0093] Finally, for each cell in the candidate list of the user, a possible set of transmission
schemes that could be utilized by that cell, should be specified. This will obviously reduce
the blind detection complexity at the user end and will also enable the network to configure
the best possible scenario for NAICS (if deemed beneficial by the network), where the
users sees the same transmission scheme (such as a DMRS based scheme) being used by
both the serving cell and the interferer.
[0094] B4. Benefit metric in Coordinated Multi-point Transmission and Reception
with Non-Ideal backhaul (CoMP-NIB)
[0095] With reference to FIG. 2, in order to allow CoMP-NIB implementation, CoMP
coordination request including (but not limited to) the followings can be sent from one eNB
to another:
- One or more CoMP hypotheses, each comprising a hypothetical resource allocation
associated with a cell ID, where the cell identified by the cell ID is not necessarily
controlled by the receiving eNB,
- A benefit metric associated with one or more CoMP hypothesis/es, quantifying the benefit
that a cell of the sender node expects in its scheduling when the associated CoMP
hypothesis/es is assumed, and
- Necessary time/frequency granularity and signaling period: Same as the associated
CoMP hypothesis/es.
[0096] Consider the benefit metric associated with one CoMP hypothesis and suppose that
the cell ID in that hypothesis identifies a cell controlled by the receiving eNB. The intention
of benefit metric is to help the receiving eNB gauge the benefit that will be accrued by the
sending eNB, if it follows the suggestion in the associated CoMP hypothesis. The receiving
eNB can weigh this benefit against the loss it might accrue upon following that suggestion,
and then decide its response. However, implicit in the derivation of this cell-specific benefit
metric is the use of a reference state that the sending eNB assumes for the receiving eNB
(or equivalently for the cell identified by the ID) over the time-frequency resource indicated
in the CoMP hypothesis. For instance, if the CoMP hypothesis suggests muting" (or zero
power-level) over a time-frequency resource, the sending eNB could have computed the
benefit metric after assuming a reference state of non-muting (i.e., a certain non-zero power
level) for the receiving eNB over the same indicated time-frequency resource. In the
multi-vendor scenario and particularly in the case when multiple power levels (not just
binary) can be indicated via a CoMP hypothesis, it is desirable that the reference state used
to by each sending eNB in deriving its benefit metric be known to the receiving eNB, so
that the latter can properly decide its response. This can be done without explicit signaling
if it is agreed that the benefit metric is computed by each sending eNB using a pre-defined
reference state. This pre-defined reference state can for instance be the highest power level
that can be used over a time-frequency resource or it can be the current power level being
used by the receiving eNB over the time-frequency resource.
[0097] Next, let us consider a common benefit metric associated with multiple CoMP
hypotheses.
[0098] Here, again the aforementioned reference state can be assumed for all cells
indicated via their IDs in the multiple hypotheses. The use of benefit metric is better
justified when it is associated to one hypothesis rather than multiple hypotheses, since in
the latter case it is not possible to determine which individual hypothesis contributes what
fraction of that overall common benefit metric. Consequently, for a given number of bits
available to convey the benefit metric, the range of the benefit metric must be optimized for
the case when it is used for an individual hypothesis rather than multiple hypotheses.
Further, as an alternative, a scaling factor for the benefit metric should be separately
configurable (on a per-eNB basis if needed). Then, the receiving eNB can scale the
received benefit metric by the scaling factor associated with the sending eNB (which could
be common for all eNBs or as an option could be configured separately for each sending
eNB) to decide its response. Another alternative would be for each eNB to obtain a time
average of the benefit metrics sent by a sending eNB and then determine the scaling factor
for that sending eNB using that average.
[0099] Embodiment C
[00100] In 3GPP RAN3 Meeting #84, the following agreements on X2 messages to
support the inter-eNB CoMP were reached [3]:
"The task of inter-eNB CoMP is to coordinate multiple eNBs in order that the coverage of
high data rates and the cell-edge throughput are improved, and also the system throughput
is increased. The coordination of multiple eNBs is achieved by signalling between eNBs of
hypothetical resource allocation information, CoMP hypotheses, associated with benefit
metrics. Each of the signalled CoMP hypotheses is concerned with a cell belonging to
either the receiving eNB, the sending eNB or their neighbour. The benefit metric associated
with the CoMP hypotheses quantifies the benefit assuming that the CoMP hypotheses are
applied. The receiving eNB of the CoMP hypotheses and the benefit metrics may take them
into accountfor RRM and may triggerfurther signalling FFS.
RSRP measurement reports can also be exploitedfor inter-eNB CoMP. For example, the
RSRP measurement reports can be used to determine and/or validate CoMP hypotheses
and benefit metrics. [Further explanation on the RSRP measurement reports ofUEs: FFS]
Inter-eNB CoMP is located in the eNB."
[00 101] In the following, we provide our views along with the required message
structure.
[00102] Cl.l CoMP Hypothesis for inter-eNB CoMP
[00103] Each CoMP hypothesis (CH) contains a hypothetical resource allocation for
a cell that is not necessarily controlled by the receiving eNB. The design of signaling
associated with such CoMP hypotheses must facilitate both centralized and distributed
radio resource management (RRM). In centralized RRM a potential use of CH would be a
mandatory resource allocation that the cell indicated in that CH will (or must) follow,
whereas in a distributed RRM scenario the CH would be a request which the indicated cell
may or may not follow. As a result, including an element in the CH to indicate whether the
constituent resource allocation is mandatory or not, is desirable. This element is also useful
when the CH is sent to the eNB not controlling the indicated cell, since then the latter eNB
can have more information about the possible resource allocation of neighboring cells, to
make its own resource allocation decision. We note that when the CH is used to convey a
mandatory resource allocation (or a final decision of centralized RRM) there is limited use
of the associated benefit metric. Thus, one approach of realizing the element would be via a
special value of the benefit metric. In particular, when the associated benefit metric is null
or set to that special value then the resource allocation in the CH is mandatory, otherwise,
the resource allocation is not mandatory. An example of centralized coordination is given in
FIGs. 3(a) and 3(b), and that of a distributed coordination is given in FIG. 4. Note that in
the distributed case, eRNTP can be used to convey the resource allocation decisions.
[00104] Proposal CI : Include an element in CoMP hypothesis message to indicate
whether the included resource allocation for the indicated cell is mandatory or not.
[00105] Another relevant point here is that a cell needs to be indicated in the CH
using an ID. This ID should be unique for each cell. This requirement rules out using the
physical cell ID, since in certain deployments multiple neighboring cells (or transmission
points) can share the same physical cell ID. It is nevertheless important to be able to specify
or signal a CH for a particular cell among a set of cells sharing the same physical cell ID.
[00106] C1.2 Benefit Metric
[00107] We first consider the role of benefit metric in a distributed setup. In such a
case the cell indicated in the associated CoMP hypothesis will typically be controlled by the
receiving eNB. Then, the intention of benefit metric (as stated in RANI proposals such as
[4]) is to help the receiving eNB gauge the benefit that will be accrued by the sending eNB,
if it follows the suggested resource allocation in the associated CoMP hypothesis. The
receiving eNB can then add up all the metrics it receives for a particular cell controlled by
it and a particular resource allocation, and compare the sum against the gain or loss it might
incur, in order to decide the resource allocation for its cell. For the receiving eNB to make a
decision that will lead toward a social optima, it should have information about the loss it
can cause to other eNBs by certain allocation (such as power boosting on some PRB that
was muted previously in response to a request). This point is illustrated in FIG. 5.
Moreover, in the case the cell identified by the sending eNB is controlled by the sender, a
negative value can be used to convey the loss the sending eNB can incur by muting a
certain resource. For instance, we note that the sign of the benefit metric value can be
separately conveyed via a separate binary valued element in the benefit metric field, which
is one if the metric is positive and is zero otherwise, or vice versa.
[00 108] Proposal C2 : Allow negative values in the benefit metric.
[00109] The guiding principle behind benefit metric was that it could be used to
convey the change in a utility function in a succinct manner. The utility function usually
depends on several factors such as queue sizes, channel states, priorities (or quality of
service (QoS) classes) of the users being served by that eNB or cell. The benefit metric has
the potential to convey the change resulting from a hypothetical resource allocation,
without the need of signaling all the constituent terms of the utility function. However, this
potential can be realized only if the benefit metric field is large enough. Moreover, a
potentially serious drawback of not having a benefit metric field that allows for a fine
quantization of the utility change is that it can lead to oscillatory behavior in distributed
coordination. An additional use of a larger benefit metric field is that it provides the
operator the flexibility to simultaneously convey different utility changes for the same
hypothetical resource allocation (or set of resource allocations in the CoMP hypothesis set
associated with that benefit metric), where each such change can be computed by
emphasizing different terms of the utility function.
[001 10] Proposal C3: The benefit metric field should be sufficiently large, e.g., 3
bytes or 2 bytes.
[001 11] It has been agreed that a single benefit metric can be associated with
multiple CoMP hypotheses, i.e., a CoMP hypothesis set. Consider such a scenario where
one benefit metric is associated with L hypotheses in a CoMP hypothesis set. In such a
case, where L>1, it will be helpful if the benefit metric field represents a string of L+l
numbers. This will enable differential encoding of benefit metric. For instance, the first
number could be the base value (quantized by a certain number of bits, where that number
is less than the benefit metric field size which is for instance 3 bytes or 24 bits) which
represents the utility change when all the resource allocations are together applied. On the
other hand, each of the other L numbers can be offsets (represented by D bits each)
computed with respect to the base value, such that the sum of the base value and the offset
captures the utility change when only the corresponding individual resource allocation is
applied. It is well established that differential encoding allows for finer quantization for a
given payload size. Note that L and D can be separately conveyed and are configurable, for
instance L can be conveyed in the range of the CoMP hypothesis set. So L=l or D=0 would
mean that the benefit metric reduces to a single number that is common for all the
associated hypothesis or hypotheses. An alternative benefit of this differential encoding
feature is that it provides the operator the flexibility to convey different utility changes for
the same hypothetical resource allocation, where each such change can be computed by
emphasizing different terms of the utility function. Note that the value of L can vary
between 1 and a maximum, denoted by maxnoofCoMPCells. Example values for
maxnoofCoMPCells are 4, 8, 16, or 256. We note here that a larger value of
maxnoofCoMPCells can help to reduce overhead (since a single benefit metric field is
associated with all the hypotheses in the set) and is useful if the CoMP hypothesis set is
being used to convey the final decision in a centralized RRM, since in that case the
associated single benefit metric value can be set to a special value (or null) to indicate that
the hypothesis set is mandatory.
[001 12] Proposal C4: Differential encoding of the benefit metric field should be
supported.
[001 13] We discussed the necessary X2 message to support the inter-eNB CoMP.
[001 14] C2. Text proposal
[001 15] 9.2.XX CoMP Information
[001 16] This Information element (IE) provides the list of CoMP hypothesis sets,
where each CoMP hypothesis set is the collection of CoMP hypothesis(es) of one or
multiple cells and each CoMP hypothesis set is associated with a benefit metric.
[001 17] Example-la
Index: FFS]
[001 18] Example-lb
Index: FFS]
[001 19] Example-2a
Index: FFS]
[00120] Example-2b
Index: FFS]
[00121] Example sizes for maxnoofCoMPInformation are 4, 8, 16, or 256.
[00122] Embodiment D
[00123] In the following we provide our views on X2 messages to support the
inter-eNB CoMP along with the required message structure.
[00124] Dl. CoMP Hypothesis for inter-eNB CoMP
[00125] Each CoMP hypothesis (CH) contains a hypothetical resource allocation for
a cell that is not necessarily controlled by the receiving eNB. The design of signaling
associated with such CoMP hypotheses and associated benefit metrics must facilitate both
centralized and distributed RRM. The use cases in both centralized and distributed RRM is
described in the appendix. Our preference for computing the benefit metric on a linear scale
is justified there.
[00126] We next present our view on the coding structure of the CoMP hypothesis.
[00127] From the agreements made so far ([2] and [3]), it is clear that a benefit
metric is associated with multiple CoMP hypotheses, where each CoMP hypothesis
indicates a resource allocation in the frequency domain (on a per-RB basis) as well as the
time domain (across multiple sub-frames). The guiding principle behind benefit metric was
that it could be used to convey the change in a utility function in a succinct manner. The
utility function usually depends on several factors such as queue sizes, channel states,
priorities (or QoS classes) of the users being served by that eNB or cell. The benefit metric
has the potential to convey the change resulting from a hypothetical resource allocation,
without the need of signaling all the constituent terms of the utility function. However, this
potential can be realized only if the benefit metric value represents a fine enough
quantization. Moreover, a potentially serious drawback of not having a benefit metric field
that allows for a fine quantization of the utility change is that it can lead to oscillatory
behavior in distributed coordination.
[00128] It is apparent that the amount of information we can convey using a single
benefit metric value (effective quantization level) becomes increasingly diminished as we
include more hypotheses in the CoMP hypothesis set, as well as when we increase the
choices (possibilities) of the resource allocation that can be conveyed by each hypothesis.
Thus, the predominant use case would be to have a limited CoMP hypothesis set size
(which is controllable with the maximum being 32) and have limited choices of resource
allocation possibilities conveyed by each hypothesis.
[00129] This can be achieved by conveying resource allocation associated with each
hypothesis across frequency (on a per-RB basis) and over one (or a few) sub-frames in the
time domain (via a list). The pattern represented by the list is understood to be repeated
continuously. Furthermore, it is sensible to restrict all patterns (corresponding to different
hypotheses in the set) to have the same size in terms of the number of sub-frames spanned
by them. Such a design permits all the flexibility needed by the typical use-cases and also
achieves overhead reduction. We further note that patterns of unequal sizes also complicate
the benefit metric computation. This design is described in our text proposal.
[00130] We discussed the necessary X2 message to support the inter-eNB CoMP and
presented corresponding text proposals.
[00131] D2. Text proposal
[00132] 9.2.XX CoMP Information
[00133] This IE provides the list of CoMP hypothesis sets, where each CoMP
hypothesis set is the collection of CoMP hypothesis(ses) of one or multiple cells and each
CoMP hypothesis set is associated with a benefit metric.
IE/Group Name Presen Range IE type and Semantics
ce reference description
CoMP Information 1 ..
Item
>CoMP Hypothesis M 9.2.xy
Set
>CoMP Hypothesis M l.. (cardinality) of each
CoMP Hypothesis
list in the CoMP
Hypothesis set.
>Benefit Metric M INTEGER Value - 100 indicates
(-1 0 1..100, the maximum cost,
) and 100 indicates
the maximum
benefit.
Value - 10 1 indicates
unknown benefit.
The value is
computed on a
linear scale.
CoMP Information 0.. 1
Start Time
>Start SFN M INTEGER SFN of the radio
(0..1023) frame containing the
first subframe when
the CoMP
Information IE is
valid.
>Start Subframe M INTEGER Subframe number,
Number (0..9) within the radio
frame indicated by
the Start SFN IE, of
the first subframe
when the CoMP
Information IE is
valid.
Range bound Explanation
maxnoofCoMPInformation Maximum number of CoMP Hypothesis sets. The
value is 256.
maxnoofSubframes Maximum number of Subframes. The value is 40.
[00134] maxnoofSubframes can alternatively be 20 or 80.
[00135] 9.2.xy CoMP Hypothesis Set
[00136] This IE provides a set of CoMP hypotheses. A CoMP hypothesis is
hypothetical PRB-specific resource allocation information for a cell.
[00137] D3. Use of special value
[00138] In centralized RRM a typical use of CoMP hypothesis (CH) set would be a
mandatory resource allocation that each cell indicated in the respective CH will (or must)
follow, whereas in a distributed RRM scenario the CH would be a request which the
indicated cell may or may not follow. As a result, using a special value of the associated
benefit metric to indicate whether the constituent resource allocations are mandatory or not,
is desirable. This is also useful when the CH is sent to the eNB not controlling the indicated
cell, since then the latter eNB can have more information about the possible resource
allocation of neighboring cells, to make its own resource allocation decision. An example
of centralized coordination is given in FIG.3(a), and that of a distributed coordination is
given in FIG. 4. Note that in the distributed case, eRNTP can be used to convey the
resource allocation decisions.
[00139] D4. Use of Benefit Metric
[00 140] In the context of Section C1.2, we note that comparing different benefit
metric values for a given (hypothetical) resource allocation is simplified if these values are
computed using a linear scale. In that case we can simply add the values together (after
scaling or shifting) to assess the net benefit (or cost). The scaling or shifting parameters (if
needed) can be determined by each eNB based on previously received reports. The other
option is for an entity (operator) to provide each eNB with a loop-up-table corresponding to
each of its neighbors, which that eNB can use to first map each received benefit value to an
estimated value using the appropriate look-up-table and then compare the estimated values.
We slightly prefer the first option since the second one is more complex.
[00141] Appendix Optimizing Proportional Fairness Utility Metric
[00142] Suppose that there are K users and B transmission nodes or transmission
points (TPs) in the CoMP cluster, i.e., coordination set or of interest, where these TPs can
include multiple eNBs. For convenience in exposition, here we assume a full buffer traffic
model and let W denote the set of K users. We consider hybrid schemes where the
assignment of precoding matrices (beamforming vectors or sectored beams) to the B TPs
and the association of users with those TPs (i.e., point switching) are done in a semi-static
centralized manner based on average estimates of SINRs, rates etc. On the other hand,
given its assigned precoder (or beam) and the users associated with it, each TP does per
sub-frame scheduling independently based on the instantaneous short-term CSI.
[00143] Let W = ( 5 ) denote an assignment of a precoder tuple, where W
is the precoder assigned to the b TP. Here each precoder W can be chosen from a
pre-determined finite set which includes a codeword 0 and Wfc = 0 means that the b
TP is muted. Thus, SSPM is subsumed as a special case.
[00144] Then, let R * W ) denote an estimate of the average rate that user u can
obtain (over the available time-frequency resource normalized to have size unity) when it is
served data by TPb , given that the precoder tuple W is assigned to the B TPs and that no
other user is associated with TP b . This time-frequency unit could for example be a set of
resource blocks. Next, suppose that m total users are associated with TR . Following the
conventional approach, the average rate that user u can then obtain under proportional fair
per-subframe scheduling can be approximated asm
[00145] With these definitions in hand, we can jointly determine the assignment of a
precoding tuple and the user association (i.e., jointly consider semi-static coordinated
beamforming (SSCB) and semi-static coordinated point-switching (SSPS) problems) by
solving the following optimization problem:
[00146]
[00147] Note that in (PI), each x u is an indicator variable which is equal to one if
user u is associated with TP b and zero otherwise. Therefore the constraint in (PI)
enforces that each user must be associated with only one TP. It can be shown that (PI)
cannot be solved optimally in an efficient manner, which necessitates the design of
low-complexity algorithms that can approximately solve (PI). For any given precoder tuple
W he SSPS sub-problem can be optimally solved. Alternatively, a greedy approach can be
adopted to achieve further complexity reduction.
[00148] These solutions to the SSPS problem can be leveraged to obtain an
algorithm to sub-optimally solve the joint SSCB and SSPS problem (PI).
[00149] We next consider the SSPM-only problem where user associations are
pre-determined.
1 5 s.t. W= ,...,W , , i
[00 151] Here Sb denotes the pre-determined set of users associated to TP b and
ISb I denotes its cardinality.
[00152] (P2) is also in general a hard problem which cannot be solved optimally in
an efficient manner. Good heuristics can nevertheless be developed to solve (P2).
[00153] The foregoing is to be understood as being in every respect illustrative and
exemplary, but not restrictive, and the scope of the invention disclosed herein is not to be
determined from the Detailed Description, but rather from the claims as interpreted
according to the full breadth permitted by the patent laws. It is to be understood that the
embodiments shown and described herein are only illustrative of the principles of the
present invention and that those skilled in the art may implement various modifications
without departing from the scope and spirit of the invention. Those skilled in the art could
implement various other feature combinations without departing from the scope and spirit
of the invention.

What is claimed is:
1. A wireless communications method implemented in a transmission point (TP) used
in a wireless communications system, the wireless communications method comprising:
receiving, from another TP, channel state information (CSI) for a user equipment
(UE); and
receiving, from said another TP, user identification for the user equipment,
wherein the signaling of the CSI for the user equipment enables user identification
of the user equipment.
2. The wireless communications method as in claim 1, wherein the CSI comprises
reference signal received power (RSRP).
3. The wireless communications method as in claim 1, wherein the TP comprises a
master transmission point (MTP) and said another TP comprises an anchor transmission
point (anchor TP).
4. The wireless communications method as in claim 1, wherein the CSI is received
from said another TP over backhaul.
5. The wireless communications method as in claim 1, further comprising:
receiving, from said another TP, an attribute of a corresponding CSI process.
6. The wireless communications method as in claim 5, wherein the attribute comprises
of a zero-power CSI reference signal (CSI-RS) or non-zero-power CSI-RS.
7. The wireless communications method as in claim 1, further comprising:
processing the CSI.
8. The wireless communications method as in claim 7, wherein the processing
comprises filtering the CSI.
9. The wireless communications method as in claim 7, wherein the processing
comprises performing an average of the CSI.
10. The wireless communications method as in claim 9, wherein the average comprises
a weighted average.
11. The wireless communications method as in claim 7, wherein the processing
comprises performing subsampling of the CSI.
12. The wireless communications method as in claim 1, wherein the CSI comprises
short-term channel state information (short-term CSI).
13. The wireless communications method as in claim 12, wherein an estimate of an
average rate is computed using the processed short-term CSI.
14. A wireless communications method implemented in a transmission point (TP) used
in a wireless communications system, the wireless communications method comprising:
transmitting, to another TP, channel state information (CSI) for a user equipment
(UE); and
transmitting, to said another TP, user identification for the user equipment,
wherein the signaling of the CSI for the user equipment enables user identification
of the user equipment.
15. A transmission point (TP) used in a wireless communications system, the TP
comprising:
a receiver to receive, from another TP, channel state information (CSI) for a user
equipment (UE) and user identification for the user equipment,
wherein the signaling of the CSI for the user equipment enables user identification
of the user equipment.
16. A transmission point (TP) used in a wireless communications system, the TP
comprising:
a transmitter to transmit, to another TP, channel state information (CSI) for a user
equipment (UE) and user identification for the user equipment,
wherein the signaling of the CSI for the user equipment enables user identification
of the user equipment.
17. A wireless communications method implemented in a wireless communications
system, the wireless communications method comprising:
transmitting, from a transmission point (TP) to another TP, channel state
information (CSI) for a user equipment (UE); and
transmitting, from the transmission point (TP) to said another TP, user identification
for the user equipment,
wherein the signaling of the CSI for the user equipment enables user identification
of the user equipment.
18. A wireless communications system comprising:
a first transmission point (TP); and
a second transmission point (TP) to transmit to the first TP, channel state
information (CSI) for a user equipment (UE) and user identification for the user equipment,
wherein the signaling of the CSI for the user equipment enables user identification
of the user equipment.