WO2005015814A1

WO2005015814A1 - Signal quality monitoring method with signal averaging

Info

Publication number: WO2005015814A1
Application number: PCT/IL2003/000646
Authority: WO
Inventors: Dror Meiri; Menachem Tipris; Varda Herskowits; Tomer Segev; Yosi Shani
Original assignee: Aelis Photonics Israel Ltd
Current assignee: Aelis Photonics Israel Ltd
Priority date: 2003-08-06
Filing date: 2003-08-06
Publication date: 2005-02-17
Anticipated expiration: 2006-02-06
Also published as: AU2003247155A1

Abstract

A method of accessing quality of pulsed signals, possibly containing noise and/or Inter Symbol Interference, the method comprising: (a) providing a signal for measurement; and (b) averaging the signal for times related to the period for which the signal is below a given threshold level to produce a first metric.

Description

SIGNAL QUALITY MONITORING SYSTEM FIELD OF THE INVENTION The present invention relates the filed of signal quality and in particular to the field of monitoring of signal quality. BACKGROUND OF THE INVENTION Communication systems of many types are degraded by the presence of noise, inter symbol interference (ISI) or other problems that result in deterioration of signals. In general for digitally transmitted signals, a symbol duration is a time slot during which a symbol is transmitted. During the time slot a pulse is either transmitted (or not transmitted) during a period called the "pulse duration". The portion of a time slot during which a pulse is not transmitted (hereinafter, a "quiet time") should ideally be free of any signal. Noise or ISI cause some signal to be received during the quiet time. If this received signal is wide enough (and in the case of ISI if it spills over into the following symbol duration), it may be mistaken for a pulse, leading to errors in detection of the transmitted signal. One way of defining such deterioration of signals is by Bit Error Rate (BER). It is an object of many prior art systems to determine the BER for actual systems. However, measurement of BER is not always simple, especially in the presence of ISI or other disturbances. Furthermore, when the BER is relatively low, it may require a long time to collect enough errors to determine the BER. This makes it difficult to get a "real-time" measure of problems on the line. Furthermore, it is not always sufficient to know the value of BER. The underlying cause of the BER is also needed in some cases. For example, while equalization techniques can improve BER when the cause of BER degradation is ISI, they will not improve BER when the cause is high noise (i.e., low signal to noise ratios (SNR)). In some prior art systems the detected signal is sampled to estimate the BER in the quiet time. However, this requires synchronization with the signal timing of the received signal and/or fast sampling, which is often difficult or impossible. Other methods of finding the noise are also known, but may be difficult or expensive to implement. Other prior art systems define methods of improving the operation of transmission systems by adjustments in the systems. Various methods of correcting for degradation on optical transmission lines are shown in as yet unpublished PCT application PCT/IL02/00664, filed 12 August 2002, and entitled "Dynamic Broadband Equalizer, the disclosure of which is incorporated herein by reference. SUMMARY OF THE INVENTION An aspect of some embodiments of the invention is concerned with finding a metric related to BER. In an embodiment of the invention, the method is applied to signals which comprise "modulation with zeros (MWZ)". MWZ modulation methods include the common Amplitude Shift-Keying (ASK), On-Off Keying (OOK) and Pulse-Position-Modulation (PPM) schemes as well as other systems in which the pulse duration is shorter than the symbol duration. In an exemplary embodiment of the invention, the signal is detected and thresholded.

Time is then divided into a period in which the signal is above the threshold and a period during which the signal is below the threshold. A measure of the signal during the quiet time is determined by averaging the signal during the time period at which the signal is below the threshold. The present inventors have found that, depending on the threshold, this measure, referred to herein as the indicator "M" correlates well with the BER, if the threshold is chosen properly. Computer simulations show that M correlates to the BER caused by a combination of noise and ISI. In an exemplary embodiment of the invention, an M indicator for noise only can be derived. This derivation takes into account that the ISI is present mainly within a short time before and after a given pulse, even if the ISI spills into the following symbol duration. Thus, in this embodiment, the ISI is excluded by averaging only over periods that are reduced to avoid the times at which ISI is expected to be maximized. This reduced time M indicator is referred to herein as M_NOISE- The difference between M and M_NOISE is designated Mκι since the difference is believed to be caused mainly by the ISI. The inventors have found that Mτ_Sτ correlates well with the BER caused in ISI situations only. In an aspect of some embodiments of the invention, the time of duration (hereinafter "D") of the quiet time is used as a measure of the ISI. The shorter this time, the greater the ISI that is indicated. Computer simulation shows that D correlates well with BER when the threshold is carefully chosen. In an aspect of some embodiments of the invention, one or more of the measures described above is defined and used for adaptive equalization of the transmission system using techniques that have been reported. Other uses of the measures are also described. There is thus provided, in accordance with an exemplary embodiment of the invention, a method of accessing quality of pulsed signals, possibly containing noise and/or Inter Symbol Interference, the method comprising: (a) providing a signal for measurement; and (b) averaging the signal for times related to the period for which the signal is below a given threshold level to produce a first metric. Optionally, the pulsed signals are optical signals and wherein providing comprises detection of the signals. Optionally, the detection is square law detection. In an embodiment of the invention, averaging comprises: thresholding the signal; averaging the signal for measurement for those periods for which the signal is below the threshold to produce said first metric. In an embodiment of the invention, averaging comprises: thresholding the signal; estimating the rise and fall times of an undistorted pulse; averaging the signal for measurement for those periods for which the signal is below the threshold, said averaging period being further reduced, based on said estimated rise and fall times to produce said first metric. Optionally, the method comprises estimating a signal quality metric, based on the averaged signal. Optionally, the signal quality metric is an estimate of Q, BER or eye opening. In an embodiment of the invention, the method includes: estimating a time surrounding a pulse during which ISI distortion is prominent; further averaging the signal for measurement for those periods for which the signal is below the threshold, said averaging period being further reduced, based on said estimated time for which ISI distortion is prominent, to produce a second metric, such that the second metric is predominantly a measure of noise. Optionally, the method includes finding a third metric based on the first and second metrics, said third metric being predominantly representative of the amount of ISI in the signal. In an embodiment of the invention, the method comprises: estimating a time surrounding a pulse during which ISI distortion is prominent; and further averaging the signal for measurement for those periods for which the signal is below the threshold, said averaging period being restricted to a time period surrounding the positions at which the signal crosses the threshold, said time period being based on said estimated time for which ISI distortion is prominent, to produce a third metric, said third metric being representative of the amount of ISI in the signal. There is further provided, in accordance with an embodiment of the invention, a method of accessing quality of pulsed signals, possibly containing noise and/or Inter Symbol Interference, the method comprising: (a) providing a signal for measurement; and (b) determining average time that the signal is below a given threshold level to produce a first metric. Optionally, the method includes estimating a signal quality metric, based on the first metric. Optionally the metric is an estimate of Q, BER or eye opening. In accordance with an embodiment of the invention the method includes adjusting an element in a transmission train of said signals, responsive to the first metric. Alternatively or additionally, the method includes adjusting an element in a transmission train of said signals, responsive to the first metric. Optionally, the element comprises the generation of the signals being transmitted. Optionally, the element comprises the detection of the signals. In an embodiment of the invention, the element comprises an optic fiber link filter. Optionally, the element comprises an adaptive optical equalizer (AOE) for said optic fiber link filter. In an embodiment of the invention, the element is placed downstream of a position in the transmission train at which said signal is detected. In an embodiment of the invention, the element is placed upstream of a position in the transmission train at which said signal is detected. Optionally, the element comprises a wide band adaptive optical equalizer (WAOE) for said optic fiber link filter. Optionally the method includes determining a plurality first and/or third metrics for said signal for a plurality of transmission bands utilizing said transmission chain and adjusting said equalizer responsive to metrics derived from a plurality of said bands. BRIEF DESCRIPTION OF FIGURES Exemplary, non-limiting embodiments of the invention are described in the following description, read with reference to the figures attached hereto. In the figures, identical and similar structures, elements or parts thereof that appear in more than one figure are generally labeled with the same or similar references in the figures in which they appear. Dimensions of components and features shown in the figures are chosen primarily for convenience and clarity of presentation and are not necessarily to scale. The attached figures are: Fig. 1 is a schematic representation of measurement circuitry suitable for determining measures of transmission quality degraded by noise and/or ISI, according to a first embodiment of the invention; Fig. 2 is a schematic representation of measurement circuitry suitable for determining a measure of transmission quality degraded by ISI, according to a second embodiment of the invention; Fig. 3 shows values of a metric of signal quality M, as a function of Q factor, in a computer simulation, according to the first method of the invention, for ISI degradation only, for various values of threshold (tsh) and constant SNR; Fig. 4 shows values of a metric of signal quality D, as a function of Q factor, in a computer simulation, according to the second method of the invention, for ISI degradation only, for various values of threshold (tsh) and constant SNR; Fig. 5 shows values of a metric of signal quality M, as a function of Q factor, in a computer simulation, according to the first method of the invention, for noise degradation only, for various values of threshold (tsh), without ISI; Fig. 6 shows values of a metric of signal quality, as a function of Q, in a computer simulation, according to the first method of the invention, for combined SNR, DGD and CD degradation, for various values of threshold (tsh); Figs. 7-9 show values of eye opening as a function of Q, in a computer simultaion, for various combinations of optical SNR, ISI, DGD and CD; and Figs. 10A-10E are schematic diagrams of circuitry of various applications of a signal quality monitor, according to exemplary embodiments of the invention; and DESCRIPTION OF EXEMPLARY EMBODIMENTS Fig. 1 is a schematic representation of measurement circuitry 100 suitable for determining measures of signal quality degraded by noise and or ISI, according to a first embodiment of the invention. The present invention is described with respect to an optical transmission system, for which it is especially suited, due to the difficulty of making measurements, the principles of invention are applicable to other transmission systems. The invention is especially suited for determining measures of signal or transmission quality for very high rate systems for example symbol rates of 10⁹ and up to more than 10¹⁰ per second since measurements on these systems are relatively difficult to perform. However, the present invention is also applicable to other data transmission systems at higher and lower rates, especially high rate systems, generally with only minor changes. A signal to be measured is first detected, as for example by a square law detector 102

(SLD) to provide a function r(t). This function contains the signal, together with noise and various types of distortion characteristic of the transmission system. In many systems the noise and distortions are comprised of fixed noise and distortion as well as time and signal dependent noise and distortion. High frequency elements in r(t) are optionally removed by a low pass filter (LPF) 104 to produce a filtered signal r'(t). The filtered signal enters a slicer 106, which subjects the signal to a threshold V_tsh to produce a signal I(t). I(t) is equal to zero for those time periods for which the signal is above V_tsh and equal to a convenient constant, typically 1 for those time periods for which the signal is below V_tsh. The signal I(t) acts as a gating signal for r'(t), in a gate 108. In this way I(t) is believed to select those portions of r'(t) that are not actual signal, but are mainly the result of noise, ISI and other degrading factors of the transmission system. The gated signal is then integrated (or otherwise averaged), for example in an integrator 110, over a time Tm to yield a measure M, which, as will be shown below, is a good indicator of transmission and signal quality. In an exemplary embodiment of the invention, an M indicator for noise only can be derived. This derivation takes into account that the ISI is present mainly within time slots having non-zero signals and the next adjacent symbol duration periods. Thus, in this embodiment, the ISI is excluded by averaging only over periods that are reduced to reduce the effect of ISI by excluding times at which ISI is expected to be most significant. This reduced time M indicator is referred to herein as M_NOISE- Note that this time includes many time slots (signal duration periods) that contain no transmitted signal. The difference between M and M_NOISE is designated M_ISI since it is believed to be caused mainly by the ISI. It correlates well with the BER caused in ISI situations only. In an exemplary embodiment of the invention, the width of non-zero portions of signal I(t) is decreased for example in a processor 112, to provide a signal I'(t). The leading edge of the non-zero portion is moved forward by τ_pos and the trailing edge of the zero portion is moved backward by τ_pre so that the non-zero portion is decreased by an amount τ_pre + τ_pos- These two factors are chosen to estimate the extent of the ISI distortion, before and after the pulses. Signal I'(t) is used as a gate for r'(t). I'(t) is zero for periods which are believed to include both the actual signal and most or all of the ISI distortion. The signal I'(t) acts as a gating signal for r'(t), in a gate 114. In this way I'(t) is believed to select those portions of r'(t) that are not actual signal or ISI, but mainly due to noise in the transmission system. The gated signal is then integrated (or otherwise averaged), for example in an integrator 116, over a time Tn to yield a measure M_NOISE, which, as will be shown below, is a good indicator of transmission and signal quality, for the noise portion of the unwanted signal. Since the ISI interference is mainly in the time difference between the two signals, I(t) and I'(t), the difference between M and M_NOISE is designated as Mκι. Fig. 3 shows a scatter plot of a computer simulation of the relationship between M (the abscissa) and a quality factor Q for a simulation in which only ISI is present. For reference, the relationship between Q and BER is given by Q=(2)^{1 2} erfcinv(2*BER) where efrcinv is the inverse of the error function which is:

erfc(x) = j . dt

As is clearly seen in Fig. 3, V_tsh (normalized to the maximum value of the undistorted pulse signal) values between 0.4 and 0J or 0.8 give a relationship between the Q factor and M, with relatively small spread in the values. Other values of V_tsh could be used, however, use of about 0.5 or 0.6 gives a reasonable safety band in which the relationship is usable. Each of the plots comprises 500 realizations of links ISI generated by PMD (polarization mode dispersion) with mean DGD (differential group delay) between 0 and 40 ps and CD (chromatic dispersion) between 0 and 4000 ps/nm. The SNR is constant and fixed in all cases to a high value of 20 dB (very low noise). The modulation is NRZ (non return to zero) and the data rate is 10 Gbps. Fig. 5 shows a scatter plot of a computer simulation of the relationship between M and Q where only noise is present. As can be seen the similar values of V_tsh, namely 0.6 to 0.8 give substantially the same relationship between Q and M. Each of the plots comprises 200 realizations of links with variable SNR between 8 and 20 dB. The data rate is 10 Gbps. In practical situations, both noise and ISI can be expected to be present in a transmission channel or link, and to vary. Since treatment of the two problems is radically different (for example, noise can not be reduced by equalization of the signal and ISI can). However, the noise measure can be used for determination of quality of service, were ISI is small or not present. Fig. 6 shows a scatter plot of a computer simulation of the relationship between M and Q where each plot comprises 9000 realizations of links affected by an ensemble of variable distortions: SNR between 8 and 20 dB, ISI generated by PMD with mean DGD between 0 and 40 ps and CD between 0 and 4000 ps/nm. The modulation is NRZ and the data rate is 10 Gbps. While the plots show somewhat more scatter than those of Figs. 3 and 5, the correlation is still very strong. Since Fig. 6 covers a practical range of values of noise and ISI, the curves of Fig. 6 can be used to estimate Q for practical systems. The present inventors have found that when M and MNOISE both determined, MI_SI~M/M_NOISE- Note that for very high and very low thresholds, the relationship breaks down as it appears to do for situations where M_NOISE is very low. Figs. 7, 8 and 9 show, as a reference, the relationship between Q and normalized eye opening for the simulations, whose results are show, of Figs 3, 5 and 6 respectively. While the slopes of the curves (especially Fig. 7) are different, for all cases, eye opening increases for Q greater than about 10 dB. Eye opening is an accepted signal quality measure, but generally requires rapid sampling and synchronization. The above series of computer simulations shows that the new measures of M, M_NOISE and M_BI can be used as metrics for overall quality of the signal and quality in the presence of noise and or ISI respectively. Fig. 2 is a schematic representation of measurement circuitry 200, suitable for determining a measure of transmission quality degraded by ISI, according to a second embodiment of the invention. The system of Fig. 2, differs from that of Fig. 1 in that the signal I(t) itself (and not r'(t) gated by I(t) as with Fig. 1) is integrated. This results in a metric D, plotted in a scatter graph as shown in Fig. 4, produced under the same conditions as Fig. 3. As indicated, D is well correlated with Q and as a result with eye-opening and BER. However, simulations indicate that the sensitivity of the measurement for noise only (corresponding to Fig. 5 for M) is very low (but usable under certain conditions) and that when both distortion and noise are present, correlation is not good enough to be useful. However, when only one of noise and ISI varies, the various metrics of the D family, do appear to correlate with Q. Metric D appears to be useable under these limited conditions. Figs. 10A-10F show a variety of circuits that illustrate exemplary uses for the Signal Quality Monitor of the present invention, in single and multi channel applications. In Fig. 10 A, the monitor is used to measure transmission quality. As indicated above, this can result in an inexpensive monitor for transmission quality. Both the metric M and the metric D (when either the noise or the ISI is low or non-existent) can be used as an inexpensive monitor of Quality of Service. In Fig. 10B, the monitor is used to adjust the output of a transmitter utilizing, for example adaptive control, as is known in the art. This embodiment is of little practical use, since the transmitter is not generally available at the receiver. In Fig. IOC, the monitor is used to adjust the receiver, utilizing, for example, adaptive control, as known in the art. Figs. 10D and 10E show the monitor used to provide adaptive control to signal processing circuitry between the physical channel and the receiver. In Fig. 10D, the monitor precedes the signal processing. This requires a knowledge of the sources of the distortion and a algorithms for correcting it. In Fig. 10E the signal quality monitor follows the signal processing, so that the iterative and search techniques, as well as a neural network can be used to correct the distortion. Fig. 10E shows a monitor used in a circuit, such as is shown in PCT application PCT/IL02/00664. In this circuit multiple channels can be corrected at the same time using a wide band adaptive optical equalizer (WAOE) at the end of optic fiber link filter, in the configuration shown in Fig. 10F. A plurality of metrics M or D values for said signal for a plurality of transmission bands utilizing optical link and adjusting the equalizer responsive to metrics derived from a plurality of said bands. The invention has been described in the context of particular non-limiting embodiments. While many features are shown in the exemplary embodiments, some of these features, although desirable, are not essential. Additionally, features of various embodiments may be combined in other embodiments of the invention. As used in the claims the terms "comprise", "include" or "have" or their conjugates mean "including but not limited to".

Claims

1. A method of accessing quality of pulsed signals, possibly containing noise and/or Inter Symbol Interference, the method comprising: (a) providing a signal for measurement; and (b) averaging the signal for times related to the period for which the signal is below a given threshold level to produce a first metric.

2. A method according to claim 1 wherein the pulsed signals are optical signals and wherein providing comprises detection of the signals .

3. A method according to claim 2 wherein the detection is square law detection.

4. A method according to any of the preceding claims wherein averaging comprises: thresholding the signal; averaging the signal for measurement for those periods for which the signal is below the threshold to produce said first metric.

5. A method according to any of claims 1-3 wherein averaging comprises: thresholding the signal; estimating the rise and fall times of an undistorted pulse; averaging the signal for measurement for those periods for which the signal is below the threshold, said averaging period being further reduced, based on said estimated rise and fall times to produce said first metric.

6. A method according to any of the preceding claims and comprising: estimating a signal quality metric, based on the averaged signal.

7. A method according to claim 6 wherein the signal quality metric is an estimate of Q.

8. A method according to claim 6 wherein the signal quality metric is an estimate of BER.

9. A method according to claim 6 wherein the signal quality metric is an estimate of eye opening.

10. A method according to any of the preceding claims, comprising: estimating a time surrounding a pulse during which ISI distortion is promment; further averaging the signal for measurement for those periods for which the signal is below the threshold, said averaging period being further reduced, based on said estimated time for which ISI distortion is prominent, to produce a second metric, such that the second metric is predominantly a measure of noise.

11. A method according to claim 10 and including: finding a third metric based on the first and second metrics, said third metric being predominantly representative of the amount of ISI in the signal.

12. A method according to any of claims 1 -9, comprising: estimating a time surrounding a pulse during which ISI distortion is prominent; further averaging the signal for measurement for those periods for which the signal is below the threshold, said averaging period being restricted to a time period surrounding the positions at which the signal crosses the threshold, said time period being based on said estimated time for which ISI distortion is prominent, to produce a third metric, said third metric being representative of the amount of ISI in the signal.

13. A method of accessing quality of pulsed signals, possibly containing noise and/or Inter Symbol Interference, the method comprising: (a) providing a signal for measurement; and (b) determining average time that the signal is below a given threshold level to produce a first metric.

14. A method according to claim 13 and comprising: estimating a signal quality metric, based on the first metric.

15. A method according to claim 14 wherein the signal quality metric is an estimate of Q.

16. A method according to claim 14 wherein the signal quality metric is an estimate of BER.

17. A method according to claim 14 wherein the signal quality metric is an estimate of eye opening.

18. A method according to any of the preceding claims and including adjusting an element in a transmission train of said signals, responsive to the first metric.

19. A method according to claim 11 or claim 12 and including adjusting an element in a transmission train of said signals, responsive to the first metric.

20. A method according to claim 18 or claim 19 wherein the element comprises the generation of the signals being transmitted.

21. A method according to any of claims 18-20, wherein the element comprises the detection of the signals.

22. A method according to any of claims 18-21 wherein the element comprises an optic fiber link filter.

23. A method according to claim 22 wherein the element comprises an adaptive optical equalizer (AOE) for said optic fiber link filter.

24. A method according to claim 22 or claim 23 wherein said element is placed downstream of a position in the transmission train at which said signal is detected.

25. A method according to claim 22 or claim 23 wherein said element is placed upstream of a position in the transmission train at which said signal is detected.

26. A method according to claim 25 wherein the element comprises a wide band adaptive optical equalizer (WAOE) for said optic fiber link filter.

27. A method according to claim 26 and including determining a plurality first and/or third metrics for said signal for a plurality of transmission bands utilizing said transmission chain and adjusting said equalizer responsive to metrics derived from a plurality of said bands.