CHANNEL STATE INFORMATION FEEDBACK

BACKGROUND
The present invention relates to a n apparatus and method for feeding back channel
state information.

FIELD OFTHE INVENTION
Base stations in wireless communication systems provide wireless connectivity to user
equipment within the geographic area, or cell, associated with the base station. The
wireless communication links between the base station and each of the user
equipment typically include one or more downlink (or forward) channels for transmitting
information from the base station to the user equipment and one or more uplink (or
reverse) channels for transmitting information from the user equipment to the base
station. Multiple-input-multiple-output (MIMO) techniques may be employed when the
base station and, optionally, the user equipment include multiple antennas. For
example, a base station that includes multiple antennas can transmit multiple
independent and distinct signals to multiple user equipment concurrently and on the
same frequency band.
For example, consider a cellular system with M antennas at the base station and N
antennas at the user equipment. In such communication systems, the radio channel
between the base station and the user equipment can be described in terms of NxM
links (sub-channels). Each link typically has a time-varying complex gain (i.e. amplitude
and phase). If the radio channel is wideband .e. symbol rate greater than the delay
spread of the channel), the complex gain varies across the bandwidth of the
transmitted signal. The overall state of the radio channel can therefore be described as
a series of complex weights. This channel state information is measured by the user
equipment and fed back to the base station in order to allow the base station to adapt
characteristics of the signals transmitted to the user equipment to match it in the most
appropriate way to the prevailing channel state.
Although techniques exist to provide channel state information feedback, they each
have their own shortcomings. Accordingly, it is desired to provide an improved
technique for providing channel state information.

SUMMARY
According to a first aspect, there is provided a method of providing channel state
information for a wireless communications channel provided between a first network
node having at least one transmission antenna and a second network node having at
least one reception antenna, the method comprising the steps of: determining
characteristics of each tap resolvable in the time domain of at least one sub-channel
within the channel from signals received by the at least one reception antenna over
the channel from the at least one of transmission antenna; arranging the characteristics
into at least one vector; quantising the at least one vector by selecting one of a
plurality of codebook vectors at a first level of a hierarchical codebook of vectors; and
providing, to the first network node, a n indication of an index to the one of a plurality of
codebook vectors.
The first aspect recognises that increasing the accuracy of the channel state
information fed back increases the signalling overhead and there is therefore a t rade
off between accuracy of the channel state information fed back and the feed back
overhead. The first aspect also recognises that by determining the time domain
response characteristics of at least one sub channel, it is possible to provide information
relating to the dominant aspects of those time domain characteristics by identifying the
dominant taps Q.e. the primary and any secondary or reflected signals) and transmit
information relating to those taps (such as their timing, amplitude and phase) which
would provide the first network node with sufficient information to adapt its transmissions
without needing to send less important information relating to the time domain
response characteristic. With that information it is possible to adapt the transmissions
made by the first network node.
Accordingly, the characteristics of each tap resolvable in the time domain of at least
one sub channel within the channel are determined from signals received by the
reception antennas. Hence, not all characteristics of the received signal need be
determined. These taps may for example be identified in that they exceed some
predetermined signal threshold, as well as being separated sufficiently in time to enable
the taps to be distinguishable). The characteristics of the resolvable taps are then
arranged into at least one vector. Arranging the characteristics into this vector
provides for an efficient grouping without which the benefits of the subsequent
quantisation steps would be difficult to achieve. The subsequent vector or vectors are
then quantised using a hierarchical codebook of vectors. It will be appreciated that
such a hierarchical codebook of vectors provides a number of codebook vectors at
each level of the hierarchical codebook, each of which is selectable based on
predetermined criteria such as, for example, that codebook vector being, for example,
the closest match to the vector to be quantised, the best match to that vector or
offering the minimum error, although it will also be appreciated that other selection
criteria may be applied since using a hierarchical codebook enables subsequent
refinement to further codebook vectors which may better represent the vector being
quantised. An index to the selected codebook vector may then be provided to the
first network node. Hence, rather than transmitting the selected vector itself, only the
index to that vector need be provided, which it will be appreciated will be typically
representable with a smaller number of bits. The first network node, which also has a
copy of the codebook of vectors, can then identify the selected vector and utilise that
vector to determine approximate characteristics of each tap resolvable in the time
domain of each sub-channel from that codebook vector and adapt its transmissions
accordingly. Hence, it can be seen that the amount of channel state feedback
information provided can be drastically reduced through the combination of only
identifying characteristics of each resolvable tap and by utilising a hierarchical
codebook to quantise those characteristics, which is possible because of the
arrangement of those characteristics into a vector. Such an approach enables a
sufficiently accurate indication of the channel state information to be provided to the
first network node whilst minimising the amount of information to be fed back.
In one embodiment, the step of quantising comprises: quantising the at least one
vector by selecting, for each vector, one of a plurality of codebook vectors at a first
level from one of a plurality of hierarchical codebooks of vectors. Accordingly, a
different codebook may be provided for each vector being quantised. It will be
appreciated that this enables a suitable codebook to be selected for each vector.
In one embodiment, the method comprises the steps of: requantising the at least one
vector by selecting one of a plurality of hierarchically-related codebook vectors from
hierarchically-related levels of the hierarchical codebook of vectors; and providing, to
the first network node, an indication of an index to the one of a plurality of
hierarchically-related codebook vectors. Accordingly, the quantisation of the vector
may be successively refined by selecting codebook vectors which are related to the
vector whose index has been previously advised to the first base station. This successive
refinement enables an improved indication of the channel state information to be
provided back to the first network node over time.
In one embodiment, the one of a plurality of hierarchically-related codebook vectors
from hierarchically-related levels of the hierarchical codebook of vectors comprises
one of a plurality of child codebook vectors from child levels of the hierarchical
codebook. Accordingly, for vectors which change slowly over time, subsequent
refinements in their quantisation may be possible by selecting child codebook vectors
of the codebook vector previously indicated to the first network node. Such child
codebook vectors are typically closer refinements of the previously advised parent
codebook vector.
In one embodiment, the one of a plurality of hierarchically-related codebook vectors
from hierarchically-related levels of the hierarchical codebook of vectors comprises
one of a plurality of parent codebook vectors from parent levels of the hierarchical
codebook. Hence, for more rapidly time varying characteristics, it may be necessary to
traverse to parent levels of the hierarchical codebook to select a more appropriate
codebook vector which quantises the vector.
In one embodiment, the method comprises the step of: pre-processing the at least one
vector prior to the step of quantising.
In one embodiment, the method comprises the step of: when the pre-processing
comprises normalising, quantising at least one scalar value produced by the step of
normalising and providing, to the first network node, an indication of the at least one
scalar. Accordingly, the efficiency of the quantisation process can be further improved
by normalising the vectors prior to performing the quantisation. Also, as well as
quantising the normalised vectors, a scalar produced during the normalising process
will also need to be quantised, which may also be achieved through a hierarchical
codebook or by any other appropriate process.
In one embodiment, the step of determining comprises: determining characteristics of
each tap resolvable in the time domain of at least one sub-channel within the channel
from signals received by the at least one reception antenna over the channel from the
at least one transmission antennas by comparing signals received over time on at least
one sub-channel with a predetermined threshold. It will be appreciated that other
techniques may be utilised to determine the time domain characteristics of each
resolvable tap.
In one embodiment, the first network node comprises at least M transmission antennas,
the second network node comprises at least N reception antennas and the step of
determining comprises: determining characteristics of Ltaps resolvable in the time
domain of each sub-channel within the channel from signals received by the at least N
reception antennas over the channel from the at least M transmission antennas, where
Lis a positive integer.
In one embodiment the step of arranging comprises: arranging the characteristics of
the Ltaps resolvable in the time domain of each sub-channel within the channel into
vectors of characteristics for each of the at least N reception antennas at each tap,
each vector having a dimension of lxM.
In one embodiment, the step of arranging comprises: arranging the characteristics of
the Ltaps resolvable in the time domain of each sub-channel within the channel into
vectors of characteristics for each of the at least M transmission antennas at each tap,
each vector having a dimension of lxN.
In one embodiment, the step of arranging comprises: arranging the characteristics of
the Ltaps resolvable in the time domain of each sub-channel within the channel into a
concatenation of Nvectors of characteristics for each of the at least N reception
antennas at each tap, each of the Nvectors having a dimension of lxM.
In one embodiment, the step of arranging comprises: arranging the characteristics of
the Ltaps resolvable in the time domain of each sub-channel within the channel into a
concatenation of M vectors of characteristics for each of the at least M transmission
antennas at each tap, each of the M vectors having a dimension of lxN.
In one embodiment, the step of arranging comprises: arranging the characteristics of
the Ltaps resolvable in the time domain of each sub-channel within the channel into
vectors of characteristics for each of the at least M transmission antennas and each of
the at least N reception antennas, each vector having a dimension of lxL
In one embodiment, the network node comprises more than M transmission antennas
and the M transmission antennas are selected from the more than M transmission
antennas, either by the first network node which signals the result of the selection to the
second network node, or by the second network node which signals the result of the
selection to the first network node.
In one embodiment, the second network node comprises more than N reception
antennas and the N reception antennas are selected from the more than N reception
antennas, either by the first network node which signals the result of the selection to the
second network node, or by the second network node which signals the result of the
selection to the first network node.

According to a second aspect, there is provided a computer program product
operable, when executed on a computer, to perform the method steps of the first
aspect.

According to a third aspect, there is provided a network node having at least one
reception antenna and operable to provide channel state information for a wireless
communications channel provided between a another network node having at least
one transmission antenna and the network node, the network node comprising:
determination logic operable to determine characteristics of each tap resolvable in the
time domain of at least one sub-channel within the channel from signals received by
the at least one reception antenna over the channel from the at least one transmission
antenna; arrangement logic operable to arrange the characteristics into at least one
vector; quantisation logic operable to quantise the at least one vector by selecting one
of a plurality of codebook vectors at a first level of a hierarchical codebook of vectors;
and provision logic operable to provide, to the another network node, an indication of
an index to the one of a plurality of codebook vectors.
In one embodiment, the quantising logic is operable to quantise the at least one vector
by selecting, for each vector, one of a plurality of codebook vectors at a first level from
one of a plurality of hierarchical codebooks of vectors.
In one embodiment, the quantising logic is operable to requantise the at least one
vector by selecting one of a plurality of hierarchically-related codebook vectors from
hierarchically-related levels of the hierarchical codebook of vectors; and the providing
logic is operable to provide, to the first network node, an indication of an index to the
one of a plurality of hierarchically-related codebook vectors.
In one embodiment, the one of a plurality of hierarchically-related codebook vectors
from hierarchically-related levels of the hierarchical codebook of vectors comprises
one of a plurality of child codebook vectors from child levels of the hierarchical
codebook.
In one embodiment, the one of a plurality of hierarchically-related codebook vectors
from hierarchically-related levels of the hierarchical codebook of vectors comprises
one of a plurality of parent codebook vectors from parent levels of the hierarchical
codebook.
In one embodiment, the quantisation logic is operable to pre-process the at least one
vector prior to quantising.
In one embodiment, the quantisation logic is operable, when the pre-process
comprises normalising, to quantise at least one scalar value produced by the step of
normalising and the provision logic is operable to provide, to the first network node, an
indication of the at least one scalar.
In one embodiment, the determination logic is operable to determine characteristics of
each tap resolvable in the time domain of at least one sub-channel within the channel
from signals received by the at least one reception antenna over the channel from the
at least one transmission antennas by comparing signals received over time on at least
one sub-channel with a predetermined threshold.
In one embodiment, the first network node comprises at least M transmission antennas,
the second network node comprises at least N reception antennas and the
determination logic is operable to determine time of Ltaps resolvable in the time
domain of each sub-channel within the channel from signals received by the at least N
reception antennas over the channel from the at least transmission antennas, where
Lis a positive integer.
In one embodiment, the arrangement logic is operable to arrange the characteristics
of the Ltaps resolvable in the time domain of each sub-channel within the channel into
vectors of characteristics for each of the at least N reception antennas at each tap,
each vector having a dimension of lxM.
In one embodiment, the arrangement logic is operable to arrange the characteristics
of the Ltaps resolvable in the time domain of each sub-channel within the channel into
vectors of characteristics for each of the at least M transmission antennas at each tap,
each vector having a dimension of lxN.
In one embodiment, the arrangement logic is operable to arrange the characteristics
of the Ltaps resolvable in the time domain of each sub-channel within the channel into
a concatenation of Nvectors of characteristics for each of the at least N reception
antennas at each tap, each of the N vectors having a dimension of lxM.
In one embodiment, the arrangement logic is operable to arrange the characteristics
of the Ltaps resolvable in the time domain of each sub-channel within the channel into
a concatenation of M vectors of characteristics for each of the at least M transmission
antennas at each tap, each of the M vectors having a dimension of lxN.
In one embodiment, the arrangement logic is operable to arrange the characteristics
of the Ltaps resolvable in the time domain of each sub-channel within the channel into
vectors of characteristics for each of the at least M transmission antennas and each of
the at least N reception antennas, each vector having a dimension of lxL.
In one embodiment the first network node comprises more than M transmission
antennas and the transmission antennas are selected from the more than M
transmission antennas, either by the first network node which signals the result of the
selection to the second network node, or by the second network node which signals
the result of the selection to the first network node.
In one embodiment, the second network node comprises more than N reception
antennas and the N reception antennas are selected from the more than N reception
antennas, either by the first network node which signals the result of the selection to the
second network node, or by the second network node which signals the result of the
selection to the first network node.
Through this approach it can be seen that the accuracy of the channel state
information feedback can to be improved for a given feedback overhead, or the
feedback overhead can be reduced for a given accuracy.
Further particular and preferred aspects are set out in the accompanying independent
and dependent claims. Features of the dependent claims may be combined with
features of the independent claims as appropriate, and in combinations other than
those explicitly set out in the claims.
BRIEF DESCRIPTION OF THE DRAWINGS
Embodiments of the present invention will now be described further, with reference to
the accompanying drawings, in which:
Figure 1 illustrates the main components of a wireless telecommunications network
according to one embodiment;
Figures 2A to 2C illustrate a technique for determining tap characteristics;
Figure 3 illustrates a n example hierarchical codebook structure; and
Figures 4 and 5 illustrate example utilisations of a hierarchical codebook structure.

DESCRIPTION OFTHE EMBODIMENTS OVERVIEW
Figure 1 illustrates an arrangement of a base station 20 and user equipment 30 of a
wireless telecommunications network, generally 10, according to one embodiment.
The base station 20 and the user equipment 30 are examples of the first and second
network nodes although it will be appreciated that other examples of network nodes to
which the present technique may be applied exist and indeed the functionality of the
base station and user equipment may be reversed. Although just one base station and
user equipment is shown for sake of clarity, it will be appreciated that many such base
stations and user equipment may be deployed in such a wireless telecommunications
network. Each base station 20 is provided with M antennas, whilst each user equipment
30 is provided with N antennas. Typically, at least one of M and N is an integer value
greater than 1. A MIMO radio channel is established between the base station 20 and
the user equipment and sub-channels are provided between each antenna of the
base station 20 and user equipment 30. As will be explained in more detail below,
each sub-channel will have a number of taps which represent the various signal paths
between any one antenna of the base station 20 and user equipment 30. Typically, a
first tap will exist which may be the direct path between these antennas. Subsequent
taps exist which are typically the secondary or reflected paths between these
antennas. It will be appreciated that a transmitted signal may be decodable from any
one or more of these taps or a combination thereof.
Each base station 20 comprises at least one processing means adapted to receive an
indication of an index to at least one hierarchical codebook from which channel state
information can then be derived, and means to adjust its transmissions based on this
channel state information.
Each user equipment 30 comprises the functionality of a mobile terminal for transmission
and reception of signals in a network using radio transmission. Furthermore, the user
equipment 30 comprises at least one processing means adapted to determine the at
least one channel impulse response in the time domain for at least one sub-channel
between an antenna of the base station 20 and an antenna of the user equipment 30,
compare the power of said at least one channel impulse response with a predefined
threshold to identify the taps, and determine only the at least one complex coefficient
related to one or more time intervals of the at least one channel impulse response with
a power higher than said predefined threshold, i.e. the coefficients of each of the
resolvable taps. The user equipment 30 also includes means for arranging these
complex coefficients into one or more vectors, means for quantising these vectors using
at least one hierarchical codebook and means for signalling an index identifying the
vector selected from the hierarchical codebook to the base station.
The combination of identifying the coefficients of the resolvable taps, arranging those
coefficients into vectors and quantising those vectors using a hierarchical codebook to
identify a n index for transmission to the base station provides a particularly efficient
technique to providing channel state information.

TAP CHARACTERISTIC DETERMINATION
The base-band channel time domain response representation for the link between the
" base station transmission antenna of the M antennas and the user equipment
reception antenna of the N antennas of a generic user is denoted as:
= - m=\,...,M
where = ™,...,™} denotes the set of resolvable paths for this link. Such a ba se
band channel time domain response representation for example sub-channels is
derivable by the user equipment 30 and illustrated graphically in Figure 2C, a s will be
explained in more detail below.
In the following example it is assumed | |= L V n,m ; in other words it is assumed that
the number of resolvable paths is the same for the different links and that the timing of
the taps is the same for each sub-channel.
The base station 20 sends data and pilot signals in the downlink over each of the sub
channels. The user equipment 30 receives the data and pilots and performs channel
estimation preferably using the pilots, which results in a so-called channel transfer
function (CTF) per antenna-to-antenna link, i.e. for each sub-channel, and per
subcarrier.
In this example, by means of a n inverse fast Fourier transformation ( IFFT performed in
the user equipment 30, the channel transfer function (CTF) can be transferred from the
frequency domain into the time domain resulting in a channel impulse response (CIR)
for each an†enna-†o-an†enna-link. However, it will be appreciated that other
techniques may be used to derive this time domain information.
In Figure 2A the amplitudes of the channel impulse response over time are depicted for
each antenna-to-antenna link, i.e. for each sub-channel. In this example, a path
profiling procedure performed in the user equipment 30 similar to the procedure
performed in Rake receivers is used to identify the most significant taps of the sub¬
channel impulse response i.e. the taps which have the strongest power, although other
techniques may be used. The amplitudes of the channel impulse response of each
single antenna-to-antenna link are time-averaged, and the taps which have the
strongest power are identified.
Path profiling is realized by averaging the absolute values, i.e. the amplitudes, of the
channel impulse response both over time and over all antenna-to-antenna links using a
low pass filter, e.g. a first order infinite impulse response filter. The averaging over time
or antenna-to-antenna-links results in a channel impulse response that is free from fast
fading components and is called a power delay profile. Such a power delay profile, i.e.
channel impulse response, that is averaged over time and antennas is shown in Figure
2B. All taps below a certain threshold, which is e.g. defined by a combination of noise
level, interference level and a certain margin, or by the sensitivity level of the receiver in
the user equipment 30, are deleted in order to avoid wasting feedback resources for
noisy taps. In figure 2B, the threshold is indicated by a dashed line. Preferably, in the
tap assignment, a certain minimum distance between selected taps is kept which is in
the range of the sampling rate.
For the time intervals in which the amplitude of the channel impulse response that has
been averaged over time and antennas is higher than the certain threshold, i.e. for the
taps, the complex coefficients of the channel impulse response are determined for
each antenna-to-antenna-link. Preferably, the channel impulse response is timeaveraged
for each antenna-to-antenna-link before determination of the complex
coefficients in the time intervals. In other words, the complex coefficients are taken at
the delays of the assigned taps from instantaneous channel impulse responses for each
antenna-to-antenna-link.

As shown in Figure 2C, the amplitudes of the complex coefficients of the taps, i.e. of the
channel impulse response in the time intervals, are depicted for each antenna-toantenna
sub-channel over the time.

COEFFICIENT ARRANGEMENT
The channel matrix H asseen at the n" receive antenna can be defined as follows:
In one embodiment the channel matrix can be organized as follows to capture both
transmit and receive correlation:
H( ) may be defined from H ) by normalizing the columns by their 2-norm:
A vector-quantisation approach is used to feedback (i)J, while a scalar
quantisation approach may be used to feedback ||h, ( )|| , l = \,. ..,L .

It will be appreciated that vector quantisation can be applied to any suitable vector
from the channel representation. The embodiment above applies vector quantisation
to the concatenation of the link vectors at each tap. However, other embodiments
may apply vector quantisation to the individual vectors for each receive antenna at
each tap independently, or for each transmit antenna at each tap independently, or
for the vector of coefficients corresponding to the taps of one or more links.

VECTOR QUANTISATION
A different quantisation codebook can be built for each column of the normalized
channel matrix H(i) . C i = \,. ..,L is the codebook associated to , (/) . In general,
each codebook can have a different size, such that a different resolution is used to
quantise the different paths. For example a larger number of bits could be used for
quantising the dominant paths. Both the base station 20 and the user equipment 30
share the same codebooks. In one embodiment, the base station 20 is operable to
transmit codebooks to the user equipment 30 for storage therein.

The number of bits used for the i" codebook is denoted as follows:
=iog |c |
and the following constraint on the total number of bits for signalling all the columns of
H ( ) is used:
I† will be appreciated that C i = l, ...,L and Z ,. , i = \,. .., L can be optimised following a
given criterion. Successive refinement of the quantisation in time using hierarchical
quantisation is used as will now be described.
The hierarchical feedback approach assumes that if the channel is changing
sufficiently slowly the feedback can be aggregated over multiple feedback intervals so
that the aggregated bits index a larger codebook.
As it can be seen in Figure 3, each quantisation codebook can be organized using a
binary tree structure, in a way that all the codewords at the j level have the same j-1
significant bits. Such a structure can be used as an enabler for hierarchical
quantisation.

The hierarchical quantisation method can be explained referring to two messages, the
basic feedback message and the refined feedback message. The basic feedback
message is obtained by sending back to the base station 20 b bits specifying the best
quantisation codeword at level of the binary tree, typically by specifying its index.
The refined feedback message is obtained as follows. Assuming that at the time
interval t- 7, in one embodiment, both the base station 20 and user equipment 30 share
a codeword h, ( -1) C belonging to the l(t - l ) > b level. It will be appreciated that
the base station 20 and user equipment 30 need to save the sequence of selected
codewords, starting from the level b , up to level ( - 1) ) . The case / ( - 1) = b, can be
obtained as a particular case, where an 'UP' move corresponds to a new codeword in
the b level.

If c= h, ( - l ) , then a new candidate is chosen in the /(i - l )+6, - 1level, b, - l bits are
sent back to the base station 20 indicating the position in the subtree starting from
h ( - l ) , while one bit is used to signal a 'DOWN' move within the tree. If c h (/ - l ) ,
then a new candidate is chosen in the / ( - ΐ ) - ( ,. - ΐ ) level, b, - \ bits are sent back to
the base station 20 indicating the position in the subtree starting from h ( - l ) , while
one bit is used to signal a 'UP' path within the tree.

Consider the following two examples shown in Figures 4 and 5.
In Figure 4, a †the h time interval, a refinement ('DOWN') of the quantisation codeword
is done starting from / ( -l) = 3 . A new vector in level 5 is sent back to the base station
20. At the + l time interval, a second refinement ( OWN') is done starting from
/(/) = 5 . A new vector in level 7 is sent back to the base station.

In Figure 5, at the 1 time interval, a refinement ('DOWN') of the quantisation codeword
is done starting from /(/-l) = 3 . A new vector in level 5 is sent back to the base station
20. At the t+ 7h time interval, due to a channel variation, a n 'UP' move is signalled to the
base station 20, in the same subtree h,. ( ) belongs to.

The previously described embodiment using bi-1 bits for signalling the codeword in a
given tree level, and 1 bit for signalling a n "UP" or "DOWN" move in the tree, can be
generalized to the case where xi bits are used for signalling a given codeword in a
given tree level, whereas yi bits are used to signal a move in the tree. And these xi,yi
can be a function of the level in the tree, user,...

Feedback of the channel norms ||h, (f)|| , = i,...,L
The scalar ||h,. ( ) , i = \,...,L can be quantised using a traditional approach for scalar
quantisation. It will be appreciated that successive quantisation using hierarchical
codebooks can be also applied.
Hence, two approaches for efficient channel state feedback are combined. The first
one, is the so-called hierarchical feedback approach, in which the feedback is refined
over multiple feedback intervals so that the aggregated bits index a larger codebook.
The second one, is the so called time-domain compressed feedback approach. This
combined a technique enables successive refinement in different feedback intervals of
a compressed time-domain feedback information.
Embodiments provide a method for successive refinement of feedback (from a
secondary station to a primary station) of a time-domain representation of the radio
channel between a primary station and a secondary station, applicable to one or
more antennas at each of the primary and secondary stations.
The embodiments recognise that the accuracy of the channel state information
feedback can be improved over time by feeding back hierarchical refinements of the
previously fed-back time-domain representation of the channel state information.
Embodiments therefore provide a method by which a time-domain representation of
the channel state information can be successively refined by means of a hierarchical
quantisation of the time-domain representation.

In one embodiment, the steps of the method comprise, at the secondary station:
1. constructing a time-domain representation of the radio channel from the M
primary station antennas to the N secondary station antennas (where M>1 and N>1).
2 . collecting the elements of the time-domain representation into one or more
vectors.
3. quantising the or each vector according to a first predetermined quantisation
codebook comprising a first level of a hierarchical codebook.
4. signalling the quantised vec†or(s) to the primary station.
5. refining the quantisation of at least one vector according to a second
predetermined quantisation codebook comprising a second level of the hierarchical
codebook.
6. signalling the refined quantisation of the at least one vector to the primary
station.
The step of collecting the elements of the time-domain representation into one or more
vectors may be performed in a variety of ways in different embodiments. In one
embodiment, each vector to be quantised is the I M vector of link coefficients for the
n * receive antenna at the l channel tap. In another embodiment, each vector to be
quantised is the l N vector of link coefficients for the m transmit antenna at the I*
channel tap. In another embodiment, each vector to be quantised is the
concatenation of N, IxM vectors of link coefficients for the n receive antenna at the Ih
channel tap. In another embodiment, each vector to be quantised is the
concatenation of M, lxN vectors of link coefficients for the transmit antenna at the
I* channel tap. In another embodiment, each vector to be quantised is the lxL vector
of link coefficients for the n* receive antenna and m transmit antenna. In another
embodiment, each vector to be quantised is another concatenation or partial
concatenation of one or more of the above.
A person of skill in the art would readily recognize that steps of various above-described
methods can be performed by programmed computers. Herein, some embodiments
are also intended to cover program storage devices, e.g., digital data storage media,
which are machine or computer readable and encode machine-executable or
computer-executable programs of instructions, wherein said instructions perform some
or all of the steps of said above-described methods. The program storage devices may
be, e.g., digital memories, magnetic storage media such as a magnetic disks and
magnetic tapes, hard drives, or optically readable digital data storage media. The
embodiments are also intended to cover computers programmed to perform said
steps of the above-described methods.
The functions of the various elements shown in the Figures, including any functional
blocks labelled as "processors" or "logic", may be provided through the use of
dedicated hardware aswell as hardware capable of executing software in association
with appropriate software. When provided by a processor, the functions may be
provided by a single dedicated processor, by a single shared processor, or by a
plurality of individual processors, some of which may be shared. Moreover, explicit use
of the term "processor" or "controller" or "logic" should not be construed to refer
exclusively to hardware capable of executing software, and may implicitly include,
without limitation, digital signal processor (DSP) hardware, network processor,
application specific integrated circuit (ASIC), field programmable gate array (FPGA),
read only memory (ROM) for storing software, random access memory (RAM), and non
volatile storage. Other hardware, conventional and/or custom, may also be included.
Similarly, any switches shown in the Figures are conceptual only. Their function may be
carried out through the operation of program logic, through dedicated logic, through
the interaction of program control and dedicated logic, or even manually, the
particular technique being selectable by the implementer as more specifically
understood from the context.
It should be appreciated by those skilled in the art that any block diagrams herein
represent conceptual views of illustrative circuitry embodying the principles of the
invention. Similarly, it will be appreciated that any flow charts, flow diagrams, state
transition diagrams, pseudo code, and the like represent various processes which may
be substantially represented in computer readable medium and so executed by a
computer or processor, whether or not such computer or processor is explicitly shown.
The description and drawings merely illustrate the principles of the invention. It will thus
be appreciated that those skilled in the art will be able to devise various arrangements
that, although not explicitly described or shown herein, embody the principles of the
invention and are included within its spirit and scope. Furthermore, all examples recited
herein are principally intended expressly to be only for pedagogical purposes to aid the
reader in understanding the principles of the invention and the concepts contributed
by the inven†or(s) to furthering the art, and are to be construed as being without
limitation to such specifically recited examples and conditions. Moreover, all statements
herein reciting principles, aspects, and embodiments of the invention, as well as
specific examples thereof, are intended to encompass equivalents thereof.

CLAIMS
1. A method of providing channel state information for a wireless communications
channel provided between a first network node having at least one transmission
antenna and a second network node having at least one reception antenna, said
method comprising the steps of:
determining characteristics of each tap resolvable in the time domain of at
least one sub-channel within said channel from signals received by said at least one
reception antenna over said channel from said at least one transmission antenna;
arranging said characteristics into at least one vector;
quantising said at least one vector by selecting one of a plurality of codebook
vectors at a first level of a hierarchical codebook of vectors; and
providing, to said first network node, an indication of an index to said one of a
plurality of codebook vectors.
2. The method of claim 1, wherein said step of quantising comprises:
quantising said at least one vector by selecting, for each vector, one of a
plurality of codebook vectors at a first level from one of a plurality of hierarchical
codebooks of vectors.
3 . The method of claim 1 or 2, comprising the steps of:
requantising said at least one vector by selecting one of a plurality of
hierarchically-related codebook vectors from hierarchically-related levels of said
hierarchical codebook of vectors; and
providing, to said first network node, an indication of an index to said one of a
plurality of hierarchically-related codebook vectors.
4 . The method of claim 3, wherein said one of a plurality of hierarchically-related
codebook vectors from hierarchically-related levels of said hierarchical codebook of
vectors comprises one of a plurality of child codebook vectors from child levels of said
hierarchical codebook.
5. The method of claim 3, wherein said one of a plurality of hierarchically-related
codebook vectors from hierarchically-related levels of said hierarchical codebook of
vectors comprises one of a plurality of parent codebook vectors from parent levels of
said hierarchical codebook.
6. The method of any preceding claim, comprising the step of:
pre-processing said a † least one vector prior to said step of quantising.
7. The method of claim 6, comprising the step of:
when said pre-processing comprises normalising, quantising at least one scalar
value produced by said step of normalising and providing, to said first network node, an
indication of said at least one scalar.
8 . The method of any preceding claim, wherein said first network node comprises
at least M transmission antennas, said second network node comprises at least N
reception antennas and said step of determining comprises:
determining characteristics of Ltaps resolvable in the time domain of each sub
channel within said channel from signals received by said at least N reception
antennas over said channel from said at least M transmission antennas, where Lis a
positive integer.
9. The method of claim 8, wherein said step of arranging comprises:
arranging said characteristics of said Ltaps resolvable in the time domain of
each sub-channel within said channel into vectors of characteristics for each of the at
least N reception antennas at each tap.
10. The method of claim 8, wherein said step of arranging comprises:
arranging said characteristics of said Ltaps resolvable in the time domain of
each sub-channel within said channel into vectors of characteristics for each of the at
least M transmission antennas at each tap.
11. The method of claim 8, wherein said step of arranging comprises:
arranging said characteristics of said Ltaps resolvable in the time domain of
each sub-channel within said channel into a concatenation of vectors of
characteristics for each reception antenna at each tap.
12. The method of claim 8, wherein said step of arranging comprises:
arranging said characteristics of said Ltaps resolvable in the time domain of
each sub-channel within said channel into a concatenation of vectors of
characteristics for each transmission antenna at each tap.
13. The method of claim 8, wherein said step of arranging comprises:
arranging said characteristics of said Ltaps resolvable in the time domain of
each sub-channel within said channel into vectors of characteristics for each of the at
least M transmission antennas and each of the at least N reception antennas, each
vector having a dimension of lxL.
14. Computer program product operable, when executed on a computer, to
perform the method steps of any one of claims 1to 13.
15. A network node having at least one reception antenna and operable to
provide channel state information for a wireless communications channel provided
between a another network node having at least one transmission antenna and said
network node, said network node comprising:
determination logic operable to determine characteristics of each tap
resolvable in the time domain of at least one sub-channel within said channel from
signals received by said at least one reception antenna over said channel from said at
least one transmission antenna;
arrangement logic operable to arrange said characteristics into at least one
vector;
quantisation logic operable to quantise said at least one vector by selecting
one of a plurality of codebook vectors at a first level of a hierarchical codebook of
vectors; and
provision logic operable to provide, to said another network node, an indication
of an index to said one of a plurality of codebook vectors.