Non-linear precoder with separate tracking

Abstract

A network unit includes a non-linear precoder for jointly pre-processing transmit samples to be transmitted over respective communication channels for crosstalk mitigation. The non-linear precoder includes a first non-linear precoding stage configured to operate according to a first triangular precoding matrix and including a modulo function, followed by a second linear precoding stage configured to operate according to a second precoding matrix. The network unit further includes a first pilot signal generator configured to generate first pilot signals for pre-processing by the second precoding stage only to yield partially-precoded pilot signals for further transmission over the respective communication channels, and a controller configured to update the second precoding matrix based on first error measurements performed during the transmission of the partially-precoded pilot signals over the respective communication channels while keeping the first precoding matrix unaltered.

Claims

1. A network unit, comprising: a non-linear precoder for jointly pre-processing transmit samples to be transmitted over respective communication channels for crosstalk mitigation, the non-linear precoder including a first non-linear precoding stage configured to operate according to a first triangular precoding matrix and including a modulo function, followed by a second linear precoding stage configured to operate according to a second precoding matrix; a first pilot signal generator configured to generate first pilot signals for pre-processing by the second precoding stage only to yield partially-precoded pilot signals for further transmission over the respective communication channels; and a controller configured to update the second precoding matrix based on first error measurements performed during the transmission of the partially-precoded pilot signals over the respective communication channels while keeping the first precoding matrix unaltered.

2. The network unit according to claim 1, further comprising: a second pilot signal generator configured to generate at least one of second pilot signals not pre-processed by any of the first and second precoding stages to yield unprecoded pilot signals for further transmission over the respective communication channels, or third pilot signals for pre-processing by the first and second precoding stages to yield fully-precoded pilot signals for further transmission over the respective communication channels; and the controller being further configured to initialize both the first and second precoding matrices based on second error measurements performed during the transmission of the unprecoded or fully-precoded pilot signals over the respective communication channels, and by means of a matrix decomposition into a triangular matrix and a unitary matrix.

3. The network unit according to claim 1, wherein the controller is further configured to determine that an intended update of the second precoding matrix based on the first error measurements causes a transmit power mask violation over at least one identified communication channel of the communication channels, and then to lower at least one direct channel gain for respective at least one selected communication channels of the communication channels to preclude the transmit power mask violation over the at least one identified communication channel.

4. The network unit according to claim 3, wherein the controller is further configured to determine that a direct channel gain of the at least one direct channel gain is to be lowered beyond a given threshold, then to re-initialize both the first and second precoding matrices based on new second error measurements performed during the transmission of unprecoded or fully-precoded pilot signals over the respective communication channels, and by means of a new matrix decomposition into a new triangular matrix and a new unitary matrix, and next to restore initial direct channel gains for the respective communication channels.

5. The network unit according to claim 1, wherein the controller is further configured to lower initial direct channel gains for the respective communication channels by a given margin with respect to respective transmit power masks.

6. The network unit according to claim 5, wherein the controller is further configured to determine that an intended update of the second precoding matrix based on the first error measurements causes a transmit power mask violation over at least one identified communication channel of the communication channels, then to re-initialize the first and second precoding matrices based on new second error measurements performed during the transmission of unprecoded or fully-precoded pilot signals over the respective communication channels, and by means of a new matrix decomposition into a new triangular matrix and a new unitary matrix.

7. The network unit according to claim 1, wherein the first pilot signals are frequency-multiplexed together with second or third pilot signals over common Discrete Multi Tone DMT symbols.

8. The network unit according to claim 1, wherein the first pilot signals are time-multiplexed together with second or third pilot signals over distinct Discrete Multi Tone DMT symbols.

9. The network unit according to claim 1, wherein the network unit is an access node that supports wired communication with subscriber equipment over an access plant.

10. A pilot generator for use with a non-linear precoder for jointly pre-processing transmit samples to be transmitted over respective communication channels for crosstalk mittigation, the non-linear precoder including a first non-linear precoding stage configured to operate according to a first triangular precoding matrix and including a modulo function, followedby a second linear precoding stage configured to operate according to a second precoding matrix, the pilot generator configured to: generate first pilot signals for pre-processing by the second precoding stage only to yeild partially-precoded pilot signals for further transmission over the respective communication channels, and further for updating the second precoding matrix only based on first error measurements performed during the transmission of the partially-precoded pilot signals over the respective communication channels while keeping the first precodeing martrix unaltered.

11. A method for controlling a non-linear precoder for jointly pre-processing transmit samples to be transmitted over respective communication channels for crosstalk mitigation, the non-linear precoder including a first non-linear precoding stage configured to operate according to a first triangular precoding matrix and including a modulo function, followed by a second linear precoding stage configured to operate according to a second precoding matrix, the method comprises comprising: generating first pilot signals for pre-processing by the second precoding stage only to yield partially-precoded pilot signals for further transmission over the respective communication channels; and updating the second precoding matrix based on first error measurements performed during the transmission of the partially-precoded pilot signals over the respective communication channels while keeping the first precoding matrix unaltered.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The above and other objects and features of the invention will become more apparent and the invention itself will be best understood by referring to the following description of an embodiment taken in conjunction with the accompanying drawings wherein:

(2) FIG. 1 represents an overview of an access plant;

(3) FIG. 2 represents a reference model for a non-linear precoder as per the present invention;

(4) FIG. 3 represents further details for an access node as per the present invention; and

(5) FIG. 4 represents a plot of the evolution over time of the signal to Noise Ratio (SNR) of 20 vectored lines when the precoder is updated as per the present invention.

DETAILED DESCRIPTION OF THE INVENTION

(6) There is seen in FIG. 1 an access plant 1 comprising a network unit 10 at a CO, a DPU 20 coupled via one or more optical fibers to the network unit 10, and further coupled via a copper loop plant to Customer Premises Equipment (CPE) 30 at various subscriber premises.

(7) The copper loop plant comprises a common access segment 40, wherein the subscriber lines are in close vicinity with each other and thus induce crosstalk into each other, and dedicated loop segments 50 for final connection to the subscriber premises. The transmission media is typically composed of copper Unshielded Twisted Pairs (UTP).

(8) The DPU 20 comprises a vectoring processing unit for jointly processing the data symbols that are being transmitted over, or received from, the loop plant in order to mitigate the crosstalk induced within the common access segment and to increase the communication data rates achievable over the respective subscriber lines.

(9) There is seen in FIG. 2 a conceptual representation of a data communication system 300 over a loop plant using a non-linear precoder for mitigating the crosstalk incurred within the loop plant.

(10) The mathematical notations that were introduced for the discussion of the prior art are still adhered to.

(11) A transmit vector U, which comprises the respective transmit frequency samples to be transmitted over the respective communication channels, being user or control data, is input to a non-linear precoder comprising a first non-linear precoding stage 301 and a second linear precoding stage 302. The first non-linear precoding stage 301 uses a first lower-triangular and unit-diagonal precoding matrix L and a modulo function , and outputs a partially-precoded vector X to the second linear precoding stage 302. The second linear precoding stage 302 uses a second unitary or almost-unitary matrix Q, and outputs a fully-precoded vector X for further transmission over the MIMO channel 303, which is represented by a channel matrix H=D(I+C)=DR*Q*.

(12) The noisy received vector Y, which comprises the frequency samples received through the respective communication channels, and corrupted by some additive white Gaussian noise source Z, goes through the equalization stages 304 and 305. The first equalization stage 304, which corresponds to the typical frequency equalization step (FEQ), uses a diagonal matrix D.sup.1, and outputs a partially-equalized received vector Y. The second equalization stage 305 uses another diagonal matrix S.sup.1, whose element are the diagonal elements of R*. The first and second equalization stages 304 and 305 can be merged into one single equalization stage D.sup.1S.sup.1 or S.sup.1D.sup.1. The properly equalized frequency samples go then through the modulo function to yield a vector for processing by a decision stage 307 (or demodulator) to yield an estimate .Math. of the transmit vector U.

(13) The present invention proposes to insert a first pilot signal X1 between the first precoding stage 301 and the second precoding stage 302 for update of the second precoding matrix Q in a low-complex manner. The first pilot signal X1 is generated by a first pilot generator 311, and is combined together with the partially-precoded data signal X by means of a multiplexer 321. The pilot signal X1 is processed by the second precoding stage 302 only to yield a partially-precoded pilot signal.

(14) In order to get an estimate of the channel matrix H, a second pilot generator 312 generates a second pilot signal X2 that is transmitted over the channel H without going through the first precoding stage 301 and the second precoding stage 302 to yield an unprecoded pilot signal. The second pilot signal X2 is combined with the fully-precoded signal X by means of a multiplexer 322 before transmission over the channel H. The second pilot signal X2 is used for properly initializing the precoding matrices L and Q through QR matrix decomposition.

(15) The multiplexers 321 and 322 multiplex the respective pilot signals together with the respective data signals over particular tones and/or particular DMT symbols.

(16) Further details are now given as per the update procedure of the second precoding matrix Q by means of the first pilot signal X1.

(17) The precoder comprises a non-linear precoder ({circumflex over (L)}) followed by linear precoder {circumflex over (Q)}:
P={circumflex over (Q)}({circumflex over (L)})(11)
Wherein H=DSL.sup.1Q*(12),
and wherein {circumflex over (L)} and {circumflex over (Q)} denote estimates of the actual channel matrices L and Q.

(18) The receivers compensate for the respective direct channel gains D and precoding gains S, and apply modulo operation on the received symbols. The overall system model is:
{circumflex over (Y)}=(S.sup.1D.sup.1(H{circumflex over (Q)}({circumflex over (L)})U+Z))=(S.sup.1D.sup.1(DSL.sup.1Q*{circumflex over (Q)}({circumflex over (L)})U+Z))(13).

(19) In order to decouple modulo decision among lines, the precoding matrix L needs to remain essentially triangular at all times. Also, an update of L requires an update of the gain matrix S used at the receiver.

(20) The precoding matrix Q needs to remain essentially unitary. Q is allowed to deviate from a unitary matrix as long as any subsequent PSD increase remains within bounds.

(21) Therefore, the proposed procedure is to update Q on a more regular basis, and only once in a while (e.g., triggered by a too high PSD increase caused by an update of Q) to derive a new {Q, L} pair by QR matrix decomposition. At this stage, new scaling factors S need also to be sent to the remote transceivers.

(22) In this method, pilots are inserted in two locations to obtain estimates of different concatenated channels.

(23) At tracking cycle i, an estimate R.sub.H.sup.i of the channel matrix S.sup.1D.sup.1H=L.sup.1Q* is obtained by inserting a pilot signal after the precoder, namely the unprecoded pilot signal X2, and by correlating the error samples gathered over the respective subscriber lines with the respective pilot sequences that modulate the pilot signal X2.

(24) Similarly, an estimate R.sub.HQ.sup.i of the channel matrix S.sup.1D.sup.1HQ.sup.i=L.sup.1Q*Q.sup.i is obtained by inserting a pilot signal between L and Q, namely the partially-precoded pilot signal X1, and by correlating the error samples gathered over the respective subscriber lines with the respective pilot sequences that modulate the pilot signal X1. Q.sup.i denotes the precoding matrix that is in use within the second precoding stage 302 at tracking cycle i. As Q.sup.i is kept close to unitary, the power increase due to Q is minimal. No modulo operation is required. The orthogonality between the pilot sequences is preserved.

(25) Since the aim is to use Q for tracking variations or residues in the overall channel L.sup.1Q*, we are interested in the overall residual error:
E.sup.i=R.sub.HQ.sup.iL.sup.iI(14).

(26) We now want to update Q.sup.i+1Q.sup.i in a manner that drives the residual error E.sup.i down:
Q.sup.i+1=Q.sup.i+A.sup.i(15).

(27) In order to guarantee convergence, the linear update can be preconditioned. Preconditioning requires knowledge on the unprecoded channel R.sub.H.sup.i estimated through unprecoded pilot signals. Through multiplication with a transformation matrix an update is preconditioned to have the desired properties leading to stable convergence, such as positive definiteness.

(28) Also, the effect of L on the convergence is removed by multiplying A.sup.i with (L.sup.i).sup.1. Note that it is low-complex to invert a unit-diagonal triangular matrix, and that L.sup.i is kept static for multiple tracking cycles.

(29) With preconditioned linear update, we have:
A.sup.i=R.sub.H.sup.i*E.sup.i(L.sup.i).sup.1=R.sub.H.sup.i*(R.sub.HQ.sup.i(L.sup.i).sup.1)(16),
and the residual error at the next tracking cycle is:
E.sup.i+1=R.sub.HQ.sup.i+1L.sup.i+1I=R.sub.HQ.sup.iL.sup.iL.sup.1Q*A.sup.iL.sup.i, or
E.sup.i+1=(IL.sup.1Q*R.sub.H.sup.i*)E.sup.i=(IL.sup.1L.sup.1*)E.sup.i=(IL.sup.1L.sup.1*).sup.i+1E.sup.0(17).

(30) Since the eigenvalues of L.sup.1L.sup.1* are positive, an exists for which the eigenvalues of (IL.sup.1L.sup.1*) are between 0 and 1, leading to convergence.

(31) There is seen in FIG. 3 further details about a particular implementation of a nonlinear precoder and a pilot generator as per the present invention.

(32) A DPU 100 comprises: transceivers 110; a Vectoring Processing Unit (VPU) 120; and a Vectoring Control Unit (VCU) 130 for controlling the operation of the VPU 120.

(33) The DPU 100 may also comprises a postcoder for canceling the crosstalk from upstream receive signals. The corresponding blocks have been purposely omitted in FIG. 3 as they are irrelevant for the present invention.

(34) The transceivers 110 are individually coupled to the VPU 120 and to the VCU 130. The VCU 130 is further coupled to the VPU 120.

(35) The transceivers 110 individually comprise: a Digital Signal Processor (DSP) 111; and an Analog Front End (AFE) 112.

(36) The DPU 100 is coupled to CPEs 200 through respective subscriber lines L, which are assumed to form part of the same vectoring group.

(37) The CPE 200 comprises respective transceivers 210 individually comprising: a Digital Signal Processor (DSP) 211; and an Analog Front End (AFE) 212.

(38) The AFEs 112 and 212 individually comprise a Digital-to-Analog Converter (DAC) and an Analog-to-Digital Converter (ADC), a transmit filter and a receive filter for confining the signal energy within the appropriate communication frequency bands while rejecting out-of-band interference, a line driver for amplifying the transmit signal and for driving the transmission line, and a Low Noise Amplifier (LNA) for amplifying the receive signal with as little noise as possible.

(39) In case of Frequency Division Duplexing (FDD) operation where downstream and upstream communications operate simultaneously over the same transmission medium in distinct and non-overlapping frequency bands, the AFEs 112 and 212 further comprise a hybrid for coupling the transmitter output to the transmission medium and the transmission medium to the receiver input while achieving low transmitter-receiver coupling ratio. The AFE may further accommodate echo cancellation filters to reduce the coupling ratio at a further extent.

(40) In case of Time Duplexing Division (TDD) operation where downstream and upstream communications operate over the same frequency band but in distinct and non-overlapping time slots, the hybrid can be advantageously omitted as the transmitter and receiver operate in alternate mode: the receive circuitry is switched OFF (or the receive signal is discarded) while the transmit circuitry is active, and the way around, the transmit circuitry is switched OFF while the receive circuitry is active.

(41) The AFEs 112 and 212 further comprise impedance-matching circuitry for adapting to the characteristic impedance of the transmission medium, clipping circuitry for clipping any voltage or current surge occurring over the transmission medium, and isolation circuitry (typically a transformer) for DC-isolating the transceiver from the transmission medium.

(42) The DSPs 111 and 211 are configured to operate downstream and upstream communication channels for conveying user traffic over a transmission medium. The transmission medium can be of any type, including one or more Unshielded Twisted Pairs (UTP), a co-axial cable, a power line, etc.

(43) The DSPs 111 and 211 are further configured to operate downstream and upstream control channels that are used to transport control traffic, such as diagnosis or management commands and responses. Control traffic is multiplexed with user traffic over the transmission medium.

(44) More specifically, the DSPs 111 and 211 are for encoding and modulating user and control data into digital data symbols, and for de-modulating and decoding user and control data from digital data symbols.

(45) The following transmit steps are typically performed within the DSPs 111 and 211: data encoding, such as data multiplexing, framing, scrambling, error correction encoding and interleaving; signal modulation, comprising the steps of ordering the carriers according to a carrier ordering table, parsing the encoded bit stream according to the bit loadings of the ordered carriers, and mapping each chunk of bits onto an appropriate transmit constellation point (with respective carrier amplitude and phase), possibly with Trellis coding; signal scaling; Inverse Fast Fourier Transform (IFFT); cyclic Prefix (CP) insertion; and possibly time-windowing.

(46) The following receive steps are typically performed within the DSPs 111 and 211: CP removal, and possibly time-windowing; Fast Fourier Transform (FFT); Frequency EQualization (FEQ); signal de-modulation and detection, comprising the steps of applying to each and every equalized frequency sample an appropriate constellation grid, the pattern of which depends on the respective carrier bit loading, detecting the expected transmit constellation point and the corresponding transmit bit sequence, possibly with Trellis decoding, and re-ordering all the detected chunks of bits according to the carrier ordering table; and data decoding, such as data de-interleaving, error correction, de-scrambling, frame delineation and de-multiplexing.

(47) Some of these transmit or receive steps can be omitted, or some additional steps can be present, depending on the exact digital communication technology being used.

(48) The DSPs 111 are further configured to generate the respective components of the aforementioned pilot vectors X1 and X2, and thus acts as per the pilot generators 311 and 312 in FIG. 2. The DSPs 111 are configured by the VCU 130 with the respective orthogonal pilot sequences to be used for modulation of the pilot signals X1 and X2.

(49) As an illustrative embodiment, the pilot signals X1 and X2 are conveyed over common symbol positions, such as the so-called SYNC symbols, and over two dedicated set of carriers {k1} and {k2} that are interleaved together. The decimation factors for the sets {k1} and {k2} are chosen such as to be lower than the expected frequency coherence of the crosstalk channels so as to allow interpolation across frequency.

(50) The sets of carriers {k1} and {k2} are configured by the VCU 130, or alternatively the sets {k1} and {k2} are preliminary knows to the DSPs 111. The VCU 130 may also configure the symbol positions to be used for pilot transmission, or alternatively these symbol positions are preliminary known to the DSPs 111.

(51) The pilot signals X1 and X2 can also be conveyed over the same set of carriers yet over distinct symbol positions. A first set of symbol positions {m1} is reserved for transmission of the pilot signal X1, while another set of symbol positions {m2} is reserved for transmission of the pilot signal X2.

(52) The VCU 130 would then configure the DSP 111 with the sets of symbol positions {m1} and {m2} to be used for transmission of the pilot signals X1 and X2 respectively, or alternatively the sets {m1} and {m2} are preliminary known to the DSPs 111.

(53) The VPU 120 is for jointly processing the components of the transmit vector U in order to pre-compensate an estimate of the coming crosstalk. The VPU 120 comprises a first non-linear precoding stage 121 serially coupled to a second linear precoding stage 122.

(54) The first precoding stage 121 is configured to process the input vector, presently the transmit vector U, by multiplication with a first lower-triangular and unit-diagonal precoding matrix L, and includes a modulo function. Each row of the triangular matrix L is multiplied with the input vector, and the result is processed through the modulo function and substitutes for the original unprecoded sample before the next row of L is processed.

(55) The second precoding stage 122 is configured to process the input vector, presently the partially-precoded transmit vector X=(L)U or the unprecoded pilot vector X1, by multiplication with a second unitary (or almost unitary) matrix Q, to yield the fully-precoded transmit vector X=Q(L)U or a partially-precoded pilot vector QX1.

(56) The transmit vector U is input to the VPU 120 together with the pilot vectors X1 and X2. The fully-precoded transmit vector X, the partially-precoded pilot vector QX1, and the unprecoded pilot vector X2 are returned to the DSP 111i for Inverse Fast Fourier Transform (IFFT) and further transmission over the respective subscriber lines.

(57) The VPU 120 is configured by the VCU 130 with the respective set of carriers {k1} and {k2} that are used by the pilot signals X1 and X2, as well as the symbol positions used for pilot transmission. The frequency samples used for X1 pilot transmission, that is to say the frequency samples belonging to the set {k1} and corresponding to the symbol positions that have been marked as being reserved for pilot transmission, shall not be processed through the first non-linear precoding stage 121, but shall be passed transparently to the second linear precoding stage 122 for further processing (see dashed line in FIG. 3). The frequency samples used for X2 pilot transmission, that is to say the frequency samples belonging to the set {k2} and corresponding to the symbol positions that have been marked as being reserved for pilot transmission, shall not be processed by any of the first non-linear precoding stage 121 and second linear precoding stage 122, and shall be transparently returned unprecoded to the DSPs 111.

(58) Alternatively, the DSPs 111 may not send any frequency samples for the set {k2} during the pilot symbol positions as the VPU 120 is not expected to process them, or may send idle frequency samples for the set {k2} during the pilot symbol positions, and overwrite the precoded frequency samples with their initial value to yield the unprecoded pilot signal.

(59) The VCU 130 is basically for configuring the pilot signals X1 and X2, and for initializing or updating the precoding matrices L and Q that are used by the first precoding stage 121 and the second precoding stage 122 respectively.

(60) The VCU 130 configures the set of carriers {k1} and {k2}, the symbol positions and the pilot sequences to be used by the respective pilot signals X1 and X2. The pilot digit transmitted over a particular transmission line Li at frequency index k during a given symbol period m is denoted as w.sub.i.sup.m(k). The pilot sequences are mutually orthogonal, and comprises M pilot digits {w.sub.i.sup.m(k)}.sub.1 . . . M to be transmitted over M symbol periods (with M larger than the size of the vectoring group).

(61) Next, the VCU 130 gathers respective slicer errors as measured during the detection of the pilot signals X1 and X2 by the remote transceivers 210. The slicer error as measured by the transceiver 210i over a victim line Li at frequency index k during symbol period m is denoted as e.sub.i.sup.m(k). The transceivers 210 report both the real and imaginary part of the slicer error to the DPU 100 and further to the VCU 130.

(62) Alternatively, the transceivers 210 may return the raw received frequency samples prior to slicing to the DPU 100. The VCU 130 then determines the slicer errors based on the exact knowledge of the respective transmit samples.

(63) Finally, the VCU 130 correlates the M error measurements {e.sub.i.sup.m(k)}.sub.1 . . . M as measured over the victim line Li over a complete acquisition cycle with the M respective pilot digits {w.sub.i.sup.m(k)}.sub.1 . . . M of the pilot sequence transmitted over a disturber line Lj so as to obtain a crosstalk estimate from the disturber line Lj into the victim line Li at frequency index k. As the pilot sequences are mutually orthogonal, the contributions from the other disturber lines reduce to zero after this correlation step.

(64) Basically, the matrices L and Q are computed or updated based on these correlation results. Depending on the pilot injection point, the meaning of the correlation result is different. Considering the pilot signal X1, the normalized correlation yields the residual crosstalk R.sub.HQ of the concatenated channel S.sup.1D.sup.1HQ; considering the pilot signal X2, the normalized correlation yields the nominal crosstalk of the channel S.sup.1D.sup.1H, and further its QR matrix decomposition to yield a unitary matrix Q, a lower triangular matrix with unit diagonal L, and a scaling diagonal matrix S. The components r.sub.ii.sup.1 of the scaling matrix S.sup.1 shall be returned to the respective DSP 110 for further communication to the CPEs 200.

(65) In an alternative embodiment, the second pilot generator 312 can be replaced by, or combined with, a third pilot generator 313 that generates a third pilot signal X3 for processing by the first precoding stage 301 and next the second precoding stage 302 to yield a fully-precoded pilot signal Q(L)X3. The third pilot signal X3 is combined with the unprecoded signal U by means of a multiplexer 323 (see dashed blocks in FIG. 2).

(66) The correlation of the error samples with the respective pilot sequences would thus yields the residual crosstalk of the concatenated channel HP, and may be used to update the overall precoding P, e.g. by means of iterative update algorithms, and next to proceed with the QR matrix decomposition of the updated matrix P to yield a new unitary matrix Q, a new lower triangular matrix with unit diagonal L, and a new scaling diagonal matrix S.

(67) The VCU 130 is further configured to determine whether any of the row norms of the matrix Q exceeds the applicable transmit PSD mask (possibly including some additional margin). If so, the VCU 130 scales down one or more selected components of the transmit vector U so as to avoid the PSD mask violation. The VCU 130 selects the components to be scaled down based on a fairness criterion, and possibly on some further criteria such as a grade of service to be achieved, etc.

(68) If the required signal down-scaling is too important leading to an unacceptable data rate reduction on one or more subscriber lines, then a new QR matrix decomposition is triggered to restore the unitary properties of the matrix Q, and the scaling factors are restored to their initial value, typically 1.

(69) The scaling factors may either be implemented by the respective DSPs 111, or may form part of the VPU 120 as a new upfront diagonal matrix. Also, these scaling factors shall be communicated to the respective transceivers 200 like the factors r.sub.ii.sup.1.

(70) The initial scaling factors can be set to a value strictly lower than 1, e.g. 0.95, in order to let the matrix Q to deviate slightly from a unitary matrix without any signal down-scaling. The initial data rate would thus be sub-optimal, yet such a scheme is advantageous in that there is no signaling overhead with the remote transceivers 200.

(71) There is plotted in FIG. 4 the evolution of the SNRs of 20 vectored lines after successive updates of Q as per the present invention. As one can see, with a threshold of 2 allowing a 3 dB transmit power increase due to non-unitary Q, a new QR matrix decomposition for re-initializing both L and Q is required rather infrequently, presently at tracking cycle 182 and 457.

(72) So a new method is described in which channel variations or residual estimation errors are tracked using a single matrix Q, while keeping L fixed for a longer time period. This method proofs theoretically feasible. It requires two estimates: unprecoded pilots for preconditioning, and partially-precoded pilots between Q and L for tracking Q at a fast pace. L is then updated in a slower pace using QR matrix decomposition, e.g. whenever the row norm of Q exceeds a certain value.

(73) It is to be noticed that the term comprising should not be interpreted as being restricted to the means listed thereafter. Thus, the scope of the expression a device comprising means A and B should not be limited to devices consisting only of components A and B. It means that with respect to the present invention, the relevant components of the device are A and B.

(74) It is to be further noticed that the term coupled should not be interpreted as being restricted to direct connections only. Thus, the scope of the expression a device A coupled to a device B should not be limited to devices or systems wherein an output of device A is directly connected to an input of device B, and/or vice-versa. It means that there exists a path between an output of A and an input of B, and/or vice-versa, which may be a path including other devices or means.

(75) 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 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 inventor(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.

(76) The functions of the various elements shown in the figures may be provided through the use of dedicated hardware as well 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, a processor 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), etc. Other hardware, conventional and/or custom, such as read only memory (ROM), random access memory (RAM), and non volatile storage, may also be included.