WO2018198118A1

WO2018198118A1 - A system and method for coordinated silencing for alien noise mitigation in a multi-users communication system

Info

Publication number: WO2018198118A1
Application number: PCT/IL2018/050452
Authority: WO
Inventors: Itshak Bergel; Amir Leshem
Original assignee: Bar-Ilan University
Priority date: 2017-04-26
Filing date: 2018-04-25
Publication date: 2018-11-01

Abstract

The present invention relates to a method and system for suppressing alien noise in multi- users communication systems. Alien noise is caused by sources that are outside the system. As such, the operator has no control and/or limited knowledge on the alien noise. Focusing on linear receivers, at high SNR, when the alien noise is significant, the receiver needs to use part of its resources to suppress this noise. As the number of received lines is typically equal to the number of signals to detect, the system does not have resources to handle both alien suppression and signal detection. The present invention discloses a method that allows better signal detection. This is done by silencing certain users (using a time and/or frequency scheduling) i.e., ordering them not to transmit anything at certain time slots and over certain frequency-bins.

Description

A SYSTEM AND METHOD FOR COORDINATED SILENCING FOR ALIEN NOISE MITIGATION IN

A MULTI-USERS COMMUNICATION SYSTEM

Field of the Invention

The present invention relates generally to the field of multi-channel communication. More specifically, the present invention relates to coordinated silencing for alien noise mitigation in multi-users communication scheme.

Background of the invention

Digital Subscriber Line (DSL) provides high bandwidth data transmission and reception to customer premises over copper twisted-pair telephone wires. Because the wires are unshielded, they are vulnerable to various types of interference, including crosstalk, both from other subscribers to the same DSL service and other entities outside the DSL system (e.g., lines that belong to other operators sharing the same binder as the operator of the system or RF signals from various sources). Noise on DSL lines due to these latter entities is referred to as "alien noise." Such alien noise cannot generally be controlled by the DSL operator and is difficult to model for purposes of signal processing. Alien noise can therefore cause substantial degradation of the DSL signal/noise ratio (SNR).

Alien noise can also cause a significant capacity loss for DSL systems, especially at a high SNR. A centralized receiver can gain back much of the capacity loss by sophisticate processing of several lines. However, such sophisticated joint receiver for multiple lines is considered infeasible. Hence, most implementations use linear receivers. However, even the optimal linear receiver cannot cope well with alien noise, and a significant rate loss is experienced.

A linear receiver is known to be sufficient in systems that use measurements from inactive lines. However, in many scenarios, the network does not have enough (or does not have any) such measurements in order to allow reasonable signal detection.

It is therefore an object of the invention to provide a system which deals with a situation where there are not enough inactive lines by actively silencing lines (using time-frequency scheduling) and then use these lines as such pseudo signals.

Further objects and advantages of the invention will become apparent as the description proceeds.

Summary of the Invention

An apparatus for suppressing alien noise in multi-users communication system, comprising: a. multiple client devices which transmit signals and data to modems and disables upstream transmissions according to a central station receiver instructions;

b. a central station receiver, which instructs one or more of said client devices to

disable their upstream transmission according to a predefined schedule; said central station processes the data received over active communication links and disabled communication links; and

c. communication links between said multiple client devices and said central station receiver.

In an embodiment, the transmission scheme is a multicarrier transmission scheme and the disabling of the upstream transmission is performed over part of the frequency bins of each user according to a predefined schedule.

In an embodiment, the central station receiver comprises: a noise covariance matrix estimator for estimating the alien noise covariance matrix; an active user channel matrix estimator for estimating a channel matrix; a receiver for receiving the transmitted signals from said client devices together with the additive alien noise, the estimated noise covariance matrix and the estimated channel matrix from said active user channel matrix estimator, and suppressing the alien noise and the interference between said client devices; a silenced communication line selector which receives the estimated noise covariance matrix and the estimated channel matrix from said active user channel matrix estimator, and selects accordingly communication lines to be silenced by ordering the client devices to silence said selected communication lines for a predefined schedule.

In an embodiment, the number of disabled client devices is obtained from the eigenvalues of the covariance matrix of the noise.

In an embodiment, the receiver uses linear processing for the interference suppression.

In an embodiment, the receiver is a Linear Minimum Mean Square (LMMSE) receiver.

In an embodiment, the receiver is a zero forcing (ZF) receiver.

In an embodiment, the receiver is designed adaptively from the received signals.

In an embodiment, disabling upstream transmissions is done by silencing the upstream transmission or by transmitting known symbols.

In an embodiment, the receiver applies a Least Mean Square (LMS) filter. The invention also encompassed a method for suppressing alien noise in multi-users communication system, comprising the steps of: a. establishing respective communication links between multiple client devices and a central station receiver using a multi-users transmission scheme;

b. disabling upstream transmissions by one or more of the client devices over the respective communication links according to a predefined schedule;

c. estimating, based on said disabling of said upstream transmissions, an effect of alien noise and interference between said client devices on said upstream transmissions received by said central station receiver; and

d. suppressing the alien noise and interference between said client devices by applying said estimated effect in processing the upstream transmissions received by the central station.

In an embodiment, establishing the respective communication links comprises establishing Digital Subscriber Line (DSL) connections over telephone wires between the client devices and the central station.

A method according to claim 10, wherein different client devices are disabled in different bins and at different times.

In an embodiment, estimating of the effect of the alien noise is done by using a noise covariance matrix estimator.

In an embodiment, estimating of the interference between said client devices is based on channel estimation.

In an embodiment, the interference suppression is done by linear processing.

In an embodiment, the receiver is a Linear Minimum Mean Square (LMMSE) receiver. In an embodiment, the receiver is a zero forcing (ZF) receiver. In an embodiment, the receiver is designed adaptively from the received signals. In an embodiment, the receiver applies a Least Mean Square (LMS) filter. Brief Description of the drawings

Fig. 1 illustrates an uplink channel of N CPEs of the DSL system according to an embodiment of the invention;

Fig. 2 depicts the uplink channel matrix H[k] and the channel matrix estimator (for a certain frequency bin k) according to an embodiment of the invention;

Fig.3 depicts the noise covariance matrix estimator (for a certain frequency bin), according to an embodiment of the invention;

Fig. 4 shows the joint uplink receiver at the CO, according to an embodiment of the invention;

Fig. 5 depicts an exemplification of the silencing system, according to an embodiment of the invention;

Fig. 6 depicts the silenced frequency bins selector and the control channel as well as the Multi Carrier Modulator (MCM) block diagram adapted to the silencing system at the CPE, according to an embodiment of the invention; and

Fig. 7 schematically discloses the method of the invention according to an embodiment of the invention.

Detailed Description of the Embodiment of the Invention

One of the main problems of communication systems is the alien noise. A system and a method for suppressing the alien noise in multi-user communication systems are disclosed herein.

Common DSL standards, such as VDSL and G.fadt, use a multicarrier transmission scheme known as discrete multi-tone modulation (DMT), in which the upstream and downstream frequency bands are divided into smaller frequency channels, commonly referred to as "bins." During the initial setup of a DSL link, the modems test each of the bins to measure its SNR and accordingly decide on the number of bits to modulate on each bin (referred to as "bit loading") and signal these decisions to one another. Very noisy bins may not be used at all. If noise conditions change, the modems can signal a change in the bits-per-bin allocations.

In an embodiment of the invention, the system and method take advantage of the multi- carrier modulation scheme and signaling protocols on DSL links to achieve better suppression of alien noise. In the disclosed embodiments, a central station, such as the DSL access multiplexer (DSLAM), instructs the various client devices, such as customer premises modems (CPEs), to silence their upstream transmissions in certain frequency bins, according to a predefined schedule. Typically, different customer premises modems are silenced in different bins and at different times. This silencing procedure can be carried out using standard DSL signaling (as would be used if the bins in question were to be shut off due to poor SN ), and allows the customer premises modems to continue upstream transmissions on the remaining bins without interruption. Silencing known bins at known times, however, enables the DSLAM to estimate the alien signals through measurements on these bins and then cancel the alien signal from the received signals of other users that are active in these bins. Thus, this coordinated silencing allows the DSLAM to mitigate the alien signal and achieve higher overall throughput. Fig. 7 schematically describes the method disclosed herein. In the first step 701, respective communication links are established between multiple client devices, and a central station, using a multi-carrier transmission scheme. Then in step 702, the upstream transmissions are disabled by one or more of the client devices over the respective communication links at specified carrier frequencies according to a predefined schedule, while continuing the upstream transmissions at other frequencies of the multi-carrier transmission scheme. In step 703, an effect of the alien noise on the upstream transmissions received by the central station is estimated. At last, in step 704 the estimated effect is applied in processing the upstream transmissions received by the central station, so as to suppress the alien noise.

In an embodiment the disabling of the upstream transmissions is done by silencing the upstream transmission.

In another embodiment, the disabling of the upstream transmissions is done by transmitting known symbols (the known symbols can be subtracted at the receiver, and hence become equivalent to silence).

In Fig. 5 a schematic exemplary of the system is shown according to an embodiment of the invention. The silencing system 500 comprises multiple client devices which are the CPEs 101.5, 101.6, 101.7. The CPEs are connected to a central station receiver 550 through communication links, and transmits signals and data to modems 105.5, 105.6, 105.7. The CPEs 101.5-101.7 silence the upstream transmissions according to a central station receiver 550 instructions. The central station 550 receiver instructs one or more of the CPEs to silence their upstream transmission in certain frequency bins according to a predefined schedule.

The central station receiver 550 comprises: a) a noise covariance matrix estimator 103.1, which estimates the correlation of the alien noise over the different lines; b) an active user channel matrix estimator 102.2, which estimates the channel matrix for each frequency bin; c) a receiver 104.1 which receives the signals and data transmitted from the CPEs and receives the estimated alien noise covariance matrix and the estimated channel matrix for each frequency bin. The receiver processes all the received data to estimate and suppresses accordingly the alien noise and the interference between the CPEs (client devices). In addition, a silenced frequency bins selector 105.4 also receives the estimated alien noise covariance matrix and the estimated channel matrix for each frequency bin and selects accordingly frequency bins to be silenced. The selector 105.4 instructs the CPEs to silence said selected frequency bins accordingly.

Fig. 1 schematically illustrates an uplink channel 100 of N Customers Premises Equipment (CPEs) 101.2 to 101.7 of a DSL system according to an embodiment of the invention. The alien noise 120 is caused by the Far end crosstalk (FEXT) of the CPEs inside the binder 101.8 (connected to a Service Provider 130 (Hereinafter SP) from outside the system), and by Near end crosstalk (NEXT) from other SP connected to the binder and by alien transmissions 101.1 from outside the binder . The transmission in time n at frequency bin k of the CPEs i=l,2,...,N is depicted by x " [k] and the received signals by y " [k] i=l,2,...,N.

In the following a detailed formal mathematical description of the proposed method is given: The multi-user upstream channel model in DSL systems in the frequency bin k and time n can be described as: ) Yk,n = ¾τΑ,η + ^vk,n k = 1, 2, ... , K,

Where X_{k n} =

.... xfi [k]] is the vector of the transmissions from all the CPEs (101.5,101.6 and 101.7), H_{k n} is the NxN upstream channel matrix (102.1) (with input links 101.10,101.11, 101.12 and output links 101.14, 101.15 and 101.16) and v_fe is the total noise vector at the receiver in frequency bin k (including the alien noise). For simplicity of presentation, in the following, the method is described for a certain frequency bin, where the operation for the other bins is similar and independent. Therefore, the frequency bin index is omitted. Moreover, it is assumed that the channel matrix varies very slowly and is considered to be constant for a quite long time. Hence, equation (1) can be rewritten as:

(2) Yn ^ HX^ + v^.

In the initiation of the silencing system all users are assumed to be active and synchronized. As can be seen in Fig. 2, it is wished to estimate the channel matrix H 102.1, for each frequency bin, separately. Using orthogonal training sequences of duration t_c , an Active Users Channel Matrix Estimator 102.2 estimates the entry i,j of H 102.1. This is obtained by:

At a second stage, all CPEs can be silenced for T_c samples in order to allow a simple estimation of the Noise Covariance matrix. Thus, the noise covariance in Fig. 3 depicts the noise covariance matrix estimator (for a certain frequency bin). Matrix Estimator 103.1 can estimate the covariance matrix of the total noise for example using the sample covariance matrix:

as an estimation of the covariance matrix, where Y = ^∑„=i Y_n . Without loss of generality, v_n can be written as (5) v_n = Z_n + v%

Where Z_n is the received alien noise vector (that is caused by either 101.8 FEXT from CPEs from outside the system, 101.13 NEXT from SP from outside the system in the same binder or by transmissions outside the binder 101.1) and v is the AWGN in the receiver. Looking on the number of dominant eigenvalues of the covariance matrix C_v, the number of alien sources Λ½ (that have a major influence on the system's performance) can be estimated.

Using the control channel, at each time a set of N CPEs is muted at each frequency bin (note that the number of alien sources may be frequency dependent). In this

exemplification, a round-robin pattern was used for the selection of the silenced bins of each CPE. However, it might be that the silencing policy is based on the channel state. Fig. 5. discloses the silencing system 105.4 with the control channel (105.1,105.2,105.3 and 105.6).

As can be seen in Fig. 6 the CPE has a slightly modified MCM system (composed of 106.1-6) where the bit loading component 106.1 that is controlled by the feedback channel 105.2,allocates no data in the silenced bins.

Once the chosen CPEs has been silenced, the estimation of the channel matrix of the activate users can be improved. For example by using the same procedure described earlier. Denote the estimation of the channel matrix of the active CPEs by H_s.

The uplink receiver 104.1 at the DSLAM in Fig. 4, can be implemented using the estimation of the a linear receiver such as the LMMSE receiver that is given by

(6) W = H»(a_xH_sH» + C)^"1.

Then, the non-silenced CPEs' transmission scan be estimated by

(7) X_Sin = W^HY_n.

Thus, each of the output signals 104.2, 104.3 and 104.4 of the LMMSE receiver 104.1 can be decoded using the x.DSL modems 105.5, 105.6 and 105.7(respectively).

In an embodiment the uplink receiver 104.1 can also be implemented as a zero forcing (ZF) linear receiver.

In another embodiment, the uplink receiver can be obtained using an adaptive (or iterative) equalizer as an alternative to the channel and the noise covariance matrix estimation and reception blocks described above. In such embodiment, the received signals are used to directly update the receiver over sufficient number of iterations. The adaptation is typically designed to minimize a cost function such that over enough time, the receiver can converge very close to the optimal linear receiver (e.g., using a Least Mean Square filter - LMS) and hence obtain the same alien noise cancellation as the blocks described above. The adaptation can be based on transmission of known (pilot) symbols, on decisions of data symbols (decision aided), or on the statistics of the transmitted signal (non-decision aided). While the above description contains many specifics, these should not be construed as limitations on the scope of the invention, but rather as exemplifications of the

embodiments. Those skilled in the art will envision other possible variations that are within the scope of the invention. Accordingly, the scope of the invention should be determined not by the embodiment illustrated, but by the appended claims and their legal equivalents.

Claims

1. An apparatus for suppressing alien noise in multi-users communication system,

comprising:

a. multiple client devices which transmit signals and data to modems and disables upstream transmissions according to a central station receiver instructions;

2. An apparatus according to claim 1, wherein the transmission scheme is a multicarrier transmission scheme and the disabling of the upstream transmission is performed over part of the frequency bins of each user according to a predefined schedule.

3. An apparatus according to claim 1, wherein disabling upstream transmissions is done by silencing the upstream transmission or by transmitting known symbols.

4. An apparatus according to claim 1, wherein the central station receiver comprises: a. a noise covariance matrix estimator for estimating the alien noise covariance matrix; b. an active user channel matrix estimator for estimating a channel matrix;

c. a receiver for receiving the transmitted signals from said client devices together with the additive alien noise, the estimated noise covariance matrix and the estimated channel matrix from said active user channel matrix estimator, and suppressing the alien noise and the interference between said client devices;

d. a disabled communication line selector which receives the estimated noise

covariance matrix and the estimated channel matrix from said active user channel matrix estimator, and selects accordingly communication lines to be disabled by ordering the client devices to disable said selected communication lines for a predefined schedule.

5. An apparatus according to claim 4, wherein the number of disabled client devices is obtained from the eigenvalues of the covariance matrix of the noise.

6. An apparatus according to claim 1, wherein the receiver uses linear processing for the interference suppression.

7. An apparatus according to claim 6, wherein the receiver is a Linear Minimum Mean Square (LMMSE) receiver.

8. An apparatus according to claim 6, wherein the receiver is a zero forcing (ZF) receiver.

9. An apparatus according to claim 1, wherein the receiver is designed adaptively from the received signals.

10. An apparatus according to claim 8, wherein the receiver applies a Least Mean Square (LMS) filter.

11. A method for suppressing alien noise in multi-users communication system, comprising the steps of:

a. establishing respective communication links between multiple client devices and a central station receiver using a multi-users transmission scheme;

12. A method according to claim 10, wherein the transmission scheme is a multicarrier transmission scheme and the disabling of the upstream transmission is performed over part of the frequency bins of each user according to a predefined schedule.

13. A method according to claim 10, wherein disabling upstream transmissions is done by silencing the upstream transmission or by transmitting known symbols.

14. A method according to claim 10, wherein establishing the respective communication links comprises establishing Digital Subscriber Line (DSL) connections over telephone wires between the client devices and the central station.

15. A method according to claim 10, wherein different client devices are disabled in

different bins and at different times.

16. A method according to claim 10, wherein estimating of the effect of the alien noise is done by using a noise covariance matrix estimator.

17. A method according to claim 16, wherein the number of disabled client devices is obtained from the eigenvalues of the covariance matrix of the noise.

18. A method according to claim 10, wherein estimating of the interference between said client devices is based on channel estimation.

19. A method according to claim 10, wherein the interference suppression is done by linear processing.

20. A method according to claim 10, wherein the receiver is a Linear Minimum Mean Square (LMMSE) receiver.

21. A method according to claim 10, wherein the receiver is a zero forcing (ZF) receiver.

22. A method according to claim 10, wherein the receiver is designed adaptively from the received signals.

23. A method according to claim 10, wherein the receiver applies a Least Mean Square (LMS) filter.