SPECTRAL INVERSION DETECTION FOR POLARIZATION-DIVISION MULTIPLEXED OPTICAL TRANSMISSION

Abstract

Disclosed herein is a modulator (50) for polarization-division multiplexing (PDM) transmission. The modulator (50) comprises first and second DP-MZMs (12, 28) associated with first and second polarizations, each DP-MZM (12, 28) having an input for an in-phase and a quadrature driving signal for modulating the in-phase and quadrature components of an optical signal according to respective transfer functions, and a detector (58) suitable for detecting light comprising at least a portion of the light outputted by the first DP-MZM (12) and a portion of the light outputted by the second DP-MZM (28). The modulator (50) is adapted to superimpose a first pilot signal on one of the in-phase and quadrature driving signals of the first DP-MZM (12) and on one of the in-phase and quadrature driving signals of the second DP-MZM (28), and a second pilot signal on the respective other of the in-phase and quadrature driving signals of the first and second DP-MZMs (12, 28). Further, the first and second pilot signals are chosen such that the signal detected by said detector (58) is indicative as to whether the slopes of the transfer functions are different for the in-phase and quadrature components of one of the first and second DP-MZMs (12, 28) and identical for the other of the first and second DP-MZMs (12, 28).

Claims

1. A modulator for polarization-division multiplexed (PDM)signals, comprising first and second Dual Parallel Mach-Zehnder-Modulators (DP-DP-MZMs) associated with first and second polarizations, each DP-MZM having an input for an in-phase and a quadrature driving signal for modulating in-phase and quadrature components of an optical signal according to respective transfer functions, and a detector suitable for detecting light comprising at least a portion of light outputted by the first DP-MZM and a portion of light outputted by the second DP-MZM, wherein said modulator is adapted to superimpose a first pilot signal on one of the in-phase and quadrature driving signals of the first DP-MZM and on one of the in-phase and quadrature driving signals of the second -DP-MZM, and a second pilot signal on the respective other of the in-phase and quadrature driving signals of the first and second DP-MZMs, and wherein the first and second pilot signals are chosen such that the signal detectable by said detector is indicative as to whether the slopes of the transfer functions are different for the in-phase and quadrature components of one of the first and second DP-MZMs and identical for another of the first and second DP-MZMs.

2. The modulator of claim 1, wherein said first and second pilot signals are chosen such as to induce a beating component in said signal detectable by said detector.

3. The modulator of claim 1, wherein the first and second pilot signals are periodic signals.

4. The modulator of claim 2, wherein a frequency of the beating component is related to frequencies of said first and second pilot signals.

5. The modulator of claim 4, wherein the frequencies of the first pilot signal and the second pilot signal are at least approximately identical, and wherein a frequency of the beating component is essentially twice the frequency of the first and second pilot signals.

6. The modulator of claim 1, wherein the first and second pilot signals are essentially orthogonal signals.

7. The modulator of claim 1, wherein the light outputted by the second DP-MZM is subjected to a polarization conversion.

8. The modulator of claim 7, wherein at least part of the light outputted by the first DP-MZM and at least part of the polarization converted light outputted by the second DP-MZM are combined in a combined output.

9. The modulator of claim 8, wherein said at least part of the light outputted by the first and second DP-MZMs and received at said detector have non-orthogonal polarizations.

10. The modulator of claim 7, wherein said at least part of the light outputted by the second DP-MZM is directed to said detector prior to its polarization conversion.

11. The modulator of claim 7, wherein a polarization beam splitter is provided for splitting at least a portion of said combined output into first and second components of differing polarizations.

12. The modulator of claim 11, wherein the detector is arranged to detect said first component.

13. The modulator of claim 10, wherein the detector or a further detector is arranged to detect superposition of said first component and second components.

14. The modulator of claim 1, wherein the detector comprises a photodetector for detecting the intensity of the received light.

15. An optical transmitter comprising a light source a modulator for polarization-division-multiplexed (PDM) signals as recited in claim 1 for modulating light outputted by said light source, and a control unit, said control unit being configured to determine, based on a signal detected by the detector of said modulator, whether slopes of transfer functions are different for the in-phase and quadrature components of one of the first and second DP-MZMs and identical for the other of the first and second DP-MZMs, and, if this is determined to be the case, to adjust said modulator or the input to said modulator such that a sign of the slope of the transfer function of one of the in-phase or quadrature components in one of the first and second DP-MZMs is effectively reversed.

16. The optical transmitter of claim 15, wherein the control unit is adapted to effectively reverse a sign of said transfer function by effecting a complex conjugation of the output signal of one of the first or second DP-MZM.

17. A method of controlling an optical transmitter said optical transmitter comprising a modulator for polarization-division-multiplexing (PDM) signals comprising first and second Dual Parallel Mach-Zehnder-interferometers (DP-MZMs) associated with first and second polarizations, each DP-MZM having an input for an in-phase and a quadrature driving signal for modulating in-phase and quadrature components of an optical signal according to a respective transfer function, said method comprising the following steps: superimposing a first pilot signal on one of the in-phase and quadrature driving signals of the first DP-MZM and on one of the in-phase or quadrature driving signals of the second DP-MZM, superimposing a second pilot signal on the respective other of the in-phase and quadrature components of the first and second DP-MZMs, detecting light comprising at least a portion of the light outputted by the first DP-MZM and a portion outputted by the second DP-MZM, determining, from the detected light, whether the slopes of the respective transfer functions are different for the in-phase and quadrature components of one of the first and second DP-MZMs and identical for the other of the first and second DP-MZMs, and if this is determined to be the case, adjusting the modulators or the input to said modulators such that the sign of the slope of the transfer function of one of the in-phase or quadrature component in one of the first and second DP-MZMs is effectively reversed.

18. The method of claim 17, wherein the first pilot signal is superimposed on the respective in-phase driving signals of the first and of the second DP-MZM and the second pilot signal is superimposed on the respective quadrature driving signals of the first and of the second DP-MZM.

19. The method of claim 17, wherein the first and second pilot signals are superimposed on driving signals representing training sequences, wherein the training sequences corresponding to the first DP-MZM and the training sequence corresponding to the second DP-MZM are orthogonal to each other.

20. The method of claim 17, wherein said step of determining is based on a detection of a beating component in said detected light.

21. The method of one of claim 17, wherein the first and second pilot signals are periodic signals.

22. The method of claim 20, wherein a frequency of the beating component is related to the frequencies of said first and second pilot signals.

23. The method of claim 22, wherein the frequencies of the first pilot signal and the second pilot signal are at least approximately identical, and wherein the frequency of the beating component is essentially twice the frequency of the first and second pilot signals.

24. The method of claim 17, wherein the first and second pilot signals are orthogonal signals.

25. The method of one of claim 17, wherein the step of adjusting the modulators or the input to said modulators such as to effectively reverse the sign of said transfer function comprises effecting a complex conjugation of the output signal of one of the first or second DP-MZMs.

26. The method of one of claim 17, wherein the optical transmitter is a transmitter according to claim 15.

Description

SHORT DESCRIPTION OF THE FIGURES

[0038] FIG. 1 shows a modulator for polarization-division multiplexed signals according to the prior art,

[0039] FIG. 2 shows the transfer function of a Mach-Zehnder modulator,

[0040] FIG. 3 shows a transmitter according to an embodiment of the present invention,

[0041] FIG. 4 shows a transmitter according to another embodiment of the present invention, and

[0042] FIG. 5 shows a flow diagram illustrating a method according to an embodiment of the invention.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0043] For the purposes of promoting an understanding of the principles of the invention, reference will now be made to the preferred embodiments illustrated in the drawings and specific language will be used to describe the same. It will nevertheless be understood that no limitation of the scope of the invention is thereby intended, such alterations and further modifications in the illustrated device and method and such further applications of the principles of the invention as illustrated therein being contemplated therein as would normally occur now or in the future to one skilled in the art to which the invention relates.

[0044] FIG. 3 shows a transmitter 46 according to an embodiment of the present invention. The transmitter comprises a laser 48, a modulator 50 and a control unit 52. The modulator 50 has generally a similar structure as the prior art modulator of FIG. 1, and likewise comprises first and second DP-MZMs 12, 28 as well as a TE/TM polarization conversion unit 42 and a polarization beam splitter 44. Like components in the modulator 50 of FIG. 3 are designated with the same reference signs as in FIG. 1, and these components are not described again.

[0045] However, in addition to the components of the modulator 10 of FIG. 1, the modulator 50 of the invention comprises a pilot tone generator 54 for generating a first pilot signal that is super-imposed on the in-phase driving voltages V.sub.HI and V.sub.VI of the first and second DP-MZMs 12, 28, and a second pilot signal that is superimposed on the quadrature driving voltages V.sub.HQ and V.sub.VQ of the first and second DP-MZMs 12, 28. In the embodiment of FIG. 3, the pilot signals are superimposed on the respective driving signals using adders 56. In the embodiment shown, the first and second pilot signals are both sinusoidal and shifted with respect to each other by 90°. More particularly, the first and second pilot signals are of the form A.Math.cos(2πf.sub.pt) and A.Math.sin(2πf.sub.pt), respectively, where A is the signal amplitude, f.sub.p is the pilot tone frequency and t is the time. As is further shown in FIG. 3, the output 16 of the first DP-MZM 12 and the output 30 of the second DP-MZM 28 are tapped, the tapped signals are combined using a coupler 56, and the combined signal is directed to a detector 58, which in the embodiment of FIG. 3 is formed by a photodiode. Accordingly, the detector 58 is suitable and arranged for detecting light comprising a portion of the light outputted by the first DP-MZM 12 and a portion of the light outputted by the second DP-MZM 28. The output of the detector 58 is connected with a control unit 52.

[0046] As will be explained below, in case of “inconsistent spectral polarity”, the output signal of the detector 58 will include a beating component at a frequency 2f.sub.p, i.e. at twice the frequency of the pilot signals, which can be detected by the control unit 52. Herein, as mentioned before, the “inconsistent spectral polarity” refers to a situation in which the slopes of the transfer functions are different for the in-phase and quadrature components of one of the first and second DP-MZMs 12, 28, and identical for the other of the first and second DP-MZMs 12, 28. If this is determined to be the case, the control unit 52 adjusts the modulator 50 or the input to said modulator 50 such that the sign of the slope of the transfer function of one of the in-phase or quadrature components in one of the first and second DP-MZMs 12, 28 actually or virtually is reversed. This can for example be achieved by effecting a complex conjugation of the output signal of one of the first or second DP-MZMs 12, 28, for example by reversing the sign of the control voltage V .sub.HQ or V.sub.VQ. As mentioned above, this would amount to a “virtual” sign change of the corresponding transfer function. A purely digital implementation of this complex conjugation is preferred. However, other ways to adjust the modulator or the input to said modulator accordingly are likewise possible. For example, it would be possible to apply a suitable bias voltage to one of the inner MZMs 22, 24, 36, 38 or the like, in which case the sign of the slope of the transfer function could be “actually” reversed.

[0047] Next, the functioning of the transmitter 46 including the modulator 50 according to an embodment of the invention shall be explained. In the embodiment shown in FIG. 3, it is assumed that each of the inner MZMs 22, 24, 36, 38 have the following transfer function:

[00001]$\begin{array}{cc}E=Z.Math.E_{\max }.Math.\sin \left(\pi .Math.\frac{V}{2.Math.V_{\pi }}\right),& \left(4\right)\end{array}$

where E is the amplitude of the electric field of the optical signal, E.sub.max is the maximum amplitude, 2V.sub.π is the half-period of the transfer function and

[00002]$\begin{array}{cc}Z=\{\begin{array}{cc}1& \mathrm{if}.Math..Math.\mathrm{the}.Math..Math.\mathrm{biasing}.Math..Math.\mathrm{point}.Math..Math.\mathrm{has}.Math..Math.\mathrm{positive}.Math..Math.\mathrm{slope}\\ -1& \mathrm{if}.Math..Math.\mathrm{the}.Math..Math.\mathrm{biasing}.Math..Math.\mathrm{point}.Math..Math.\mathrm{has}.Math..Math.\mathrm{negative}.Math..Math.\mathrm{slope}\end{array}.& \left(5\right)\end{array}$

[0048] The optical signal s.sub.H generated at the output 16 of the first DP-MM 12, in presence of the pilot signals, reads

[00003]$\begin{array}{cc}{s}_H=Z_{\mathrm{HI}}.Math.E_{\max }.Math.\sin \left(\pi .Math.\frac{V_{\mathrm{HI}}+A.Math.\cos \left(2.Math.\pi .Math..Math.f_p.Math.t\right)}{2.Math.V_{\pi }}\right)+j.Math.Z_{\mathrm{HQ}}.Math.E_{\max }.Math.\sin \left(\pi .Math.\frac{V_{\mathrm{HQ}}+A.Math.\sin \left(2.Math.\pi .Math..Math.f_p.Math.t\right)}{2.Math.V_{\pi }}\right).& \left(6\right)\end{array}$

[0049] Herein, the frequency f.sub.p of the pilot signals is much lower than the symbol rate of the PDM QAM transmit signal, and is preferably in a range of 100 Hz to 50 kHz, more preferably in a range of 1 to 5 kHz. Further, to avoid excessive disturbance of the transmitted signal, the amplitude A of the pilot signals is chosen to be small with respect to 2V.sub.π. As a consequence, s.sub.H can be well approximated as

[00004]$\begin{array}{cc}{s}_H\cong u_H+p_H,.Math.\mathrm{where}& \left(7\right)\\ u_H=E_{\max }.Math.\left[Z_{\mathrm{HI}}.Math.\sin \left(\pi .Math.\frac{V_{\mathrm{HI}}}{2.Math.V_{\pi }}\right)+j.Math.Z_{\mathrm{HQ}}.Math.\sin \left(\pi .Math.\frac{V_{\mathrm{HQ}}}{2.Math.V_{\pi }}\right)\right].& \left(8\right)\end{array}$

is the useful signal and

[00005]$\begin{array}{cc}{p}_H=\frac{\pi .Math..Math.A}{2.Math.V_{\pi }}.Math.E_{\max }.Math.\hspace{1em}\left[Z_{\mathrm{HI}}.Math.\cos \left(2.Math.\pi .Math..Math.f_p.Math.t\right).Math.\cos \left(\pi .Math.\frac{V_{\mathrm{HI}}}{2.Math.V_{\pi }}\right)+j.Math.Z_{\mathrm{HQ}}.Math.\sin \left(2.Math.\pi .Math..Math.f_p.Math.t\right).Math.\cos \left(\pi .Math.\frac{V_{\mathrm{HQ}}}{2.Math.V_{\pi }}\right)\right]& \left(9\right)\end{array}$

is the contribution from the pilot tones. In the same manner, the V-polarization component s.sub.V can be expressed as:

[00006]$\begin{array}{cc}.Math.s_V\cong u_V+p_V,.Math..Math.\mathrm{where}& \left(10\right)\\ .Math.u_V=E_{\max }.Math.\left[Z_{\mathrm{VI}}.Math.\sin \left(\pi .Math.\frac{V_{\mathrm{VI}}}{2.Math.V_{\pi }}\right)+j.Math.Z_{\mathrm{VQ}}.Math.\sin \left(\pi .Math.\frac{V_{\mathrm{VQ}}}{2.Math.V_{\pi }}\right)\right].Math..Math..Math.\mathrm{and}& \left(11\right)\\ p_V=\frac{\pi .Math..Math.A}{2.Math.V_{\pi }}.Math.E_{\max }.Math.\hspace{1em}\left[Z_{\mathrm{VI}}.Math.\cos \left(2.Math.\pi .Math..Math.f_p.Math.t\right).Math.\cos \left(\pi .Math.\frac{V_{\mathrm{VI}}}{2.Math.V_{\pi }}\right)+j.Math.Z_{\mathrm{VQ}}.Math.\sin \left(2.Math.\pi .Math..Math.f_p.Math.t\right).Math.\cos \left(\pi .Math.\frac{V_{\mathrm{VQ}}}{2.Math.V_{\pi }}\right)\right].& \left(12\right)\end{array}$

[0050] As is seen from FIG. 3, a small portion of the output of the first DP-MZM 12 and a small portion of the output of the second DP-MZM 28—prior to TE/TM polarization conversion—are added in the combiner 56 and subjected to an electro-optical conversion by the detector 58, i.e. a photodiode. After low-pass filtering, the resulting electrical signal is

[00007]$\begin{array}{cc}y=\frac{\stackrel{\_}{|s_H+s_V.Math.{|}^2}}{{\left(\frac{\pi .Math..Math.A}{2.Math.V_{\pi }}\right)}^2}\cong \frac{\stackrel{\_}{|u_H+u_V.Math.{|}^2}}{{\left(\frac{\pi .Math..Math.A}{2.Math.V_{\pi }}\right)}^2}+\frac{\stackrel{\_}{|p_H+p_V.Math.{|}^2}}{{\left(\frac{\pi .Math..Math.A}{2.Math.V_{\pi }}\right)}^2},& \left(13\right)\end{array}$

where the overbar denotes a time averaging over a time scale that is long as compared to the symbol rate, but considerably shorter than the period of the pilot tone, and where an immaterial scaling factor has been applied. By substituting equations (9) and (12) into (13), the following expression for the electrical signal is obtained

[00008]$\begin{array}{cc}y\cong \frac{\stackrel{\_}{|u_H+u_V.Math.{|}^2}}{{\left(\frac{\pi .Math..Math.A}{2.Math.V_{\pi }}\right)}^2}+\left[\stackrel{\_}{{\cos }^2\left(\pi .Math.\frac{V_{\mathrm{HI}}}{2.Math.V_{\pi }}\right)}+\stackrel{\_}{{\cos }^2\left(\pi .Math.\frac{V_{\mathrm{VI}}}{2.Math.V_{\pi }}\right)}\right].Math.{\cos }^2\left(2.Math.\pi .Math..Math.f_p.Math.t\right)+\left[\stackrel{\_}{{\cos }^2\left(\pi .Math.\frac{V_{\mathrm{HQ}}}{2.Math.V_{\pi }}\right)}+\stackrel{\_}{{\cos }^2\left(\pi .Math.\frac{V_{\mathrm{VQ}}}{2.Math.V_{\pi }}\right)}\right].Math.{\sin }^2\left(2.Math.\pi .Math..Math.f_p.Math.t\right)+2.Math.Z_{\mathrm{HI}}.Math.Z_{\mathrm{VI}}.Math.\stackrel{\_}{\cos \left(\pi .Math.\frac{V_{\mathrm{HI}}}{2.Math.V_{\pi }}\right).Math.\cos \left(\pi .Math.\frac{V_{\mathrm{VI}}}{2.Math.V_{\pi }}\right)}.Math.{\cos }^2\left(2.Math.\pi .Math..Math.f_p.Math.t\right)+2.Math.Z_{\mathrm{HQ}}.Math.Z_{\mathrm{VQ}}.Math.\stackrel{\_}{\cos \left(\pi .Math.\frac{V_{\mathrm{HQ}}}{2.Math.V_{\pi }}\right).Math.\cos \left(\pi .Math.\frac{V_{\mathrm{VQ}}}{2.Math.V_{\pi }}\right)}.Math.{\sin }^2\left(2.Math.\pi .Math..Math.f_p.Math.t\right)& \left(14\right)\end{array}$

[0051] It can then be further assumed that

[00009]$\begin{array}{cc}\stackrel{\_}{{\cos }^2\left(\pi .Math.\frac{V_{\mathrm{HI}}}{2.Math.V_{\pi }}\right)}+\stackrel{\_}{{\cos }^2\left(\pi .Math.\frac{V_{\mathrm{VI}}}{2.Math.V_{\pi }}\right)}=\stackrel{\_}{{\cos }^2\left(\pi .Math.\frac{V_{\mathrm{HQ}}}{2.Math.V_{\pi }}\right)}+\stackrel{\_}{{\cos }^2\left(\pi .Math.\frac{V_{\mathrm{VQ}}}{2.Math.V_{\pi }}\right)}\equiv 2.Math.C_2.Math..Math.\mathrm{and}& \left(15\right)\\ .Math.\stackrel{\_}{\cos \left(\pi .Math.\frac{V_{\mathrm{HI}}}{2.Math.V_{\pi }}\right).Math.}.Math.\stackrel{\_}{\cos \left(\pi .Math.\frac{V_{\mathrm{VI}}}{2.Math.V_{\pi }}\right)}=\stackrel{\_}{\cos \left(\pi .Math.\frac{V_{\mathrm{HQ}}}{2.Math.V_{\pi }}\right)}.Math.\stackrel{\_}{.Math.\cos \left(\pi .Math.\frac{V_{\mathrm{VQ}}}{2.Math.V_{\pi }}\right)}\equiv C_1^2,& \left(16\right)\end{array}$

such that one obtains

[00010]$\begin{array}{cc}y\cong \frac{\stackrel{\_}{|u_H+u_V.Math.{|}^2}}{{\left(\frac{\pi .Math..Math.A}{2.Math.V_{\pi }}\right)}^2}+2.Math.C_2+2.Math.Z_{\mathrm{HI}}.Math.Z_{\mathrm{VI}}.Math.C_1^2.Math.\left[{\cos }^2\left(2.Math.\pi .Math..Math.f_p.Math.t\right)+\frac{Z_{\mathrm{HQ}}}{Z_{\mathrm{HI}}}.Math.\frac{Z_{\mathrm{VQ}}}{Z_{\mathrm{VI}}}.Math.{\sin }^2\left(2.Math.\pi .Math..Math.f_p.Math.t\right)\right].& \left(17\right)\end{array}$

[0052] Equation (17) shows that the electrical output signal of the photodiode 58 provides a suitable criterion to identify inconsistent spectral polarity, because it contains a spectral line of frequency 2f.sub.p if and only if

[00011]$\begin{array}{cc}\frac{Z_{\mathrm{HQ}}}{Z_{\mathrm{HI}}}\ne \frac{Z_{\mathrm{VQ}}}{Z_{\mathrm{VI}}}.& \left(18\right)\end{array}$

[0053] Note that equation (18) holds exactly if the slopes of the transfer functions are different for the in-phase and quadrature components of one of the first and second DP-MZMs 12, 28 and identical for the other of the first and second DP-MZMs 12, 28. Accordingly, the control unit 52 is configured to look for the presence of a spectral line of frequency 2f.sub.p to detect inconsistent spectral polarity at the transmitter 46. If such inconsistent spectral polarity is detected, the correction is implemented via a complex conjugation of exactly one transmitter polarization (either s.sub.H or s.sub.V). As discussed above, the remaining ambiguity can then be detected and if necessary corrected at the receiver by conjugating both received polarizations.

[0054] Note that a similar result is achieved if equation (13) is generalized in that each of s.sub.H and s.sub.V have an arbitrary phase, i.e. if s.sub.H and s.sub.V would be multiplied with a corresponding complex number. In this general case too, the criterion (18) for the beating signal remains valid.

[0055] In some circumstances, the two DP-MZMs 12, 28 including the four inner MZMs 22, 24, 36, 38 may be integrated in a single component that does not allow to tap the optical tributaries as shown in FIG. 3.

[0056] FIG. 4 shows an alternative embodiment of the transmitter 46 in which essentially a prior art modulator 10 as shown in FIG. 1 is employed, which amounts to the components included in the dashed box. As is seen in FIG. 4, a small portion of the combined optical output, i.e. downstream of the polarizer PBS 44, is tapped and split through a further PBS 60 into two orthogonal polarizations X and Y. Generally, these polarization planes will each contain a linear combination of the original transmit polarizations H and V. Therefore, after photo-detection both X and Y will give rise to the beating between the pilot tones in the same way as explained above. Accordingly, using photodetector 62 shown in FIG. 4, the beating signal can be detected for example from the X-component, since this X-component comprises a portion of light outputted by the first DP-MZM 12 and a portion of light outputted by the second DP-MZM 28.

[0057] However, if X and Y should coincide exactly with H and V, there would be no beating between the two polarizations. To cope with this pathological case and to guarantee that the sensitivity of the detector does not vanish under any conditions, according to FIG. 4 both the signals corresponding to the polarization X and the sum of the signals corresponding to the polarizations X and Y are detected, using a further detector 64. It is seen that in case of inconsistent spectral polarity, at least one of these two signals will exhibit a line at the frequency 2f.sub.p after photodetection. To avoid replicating twice the same hardware, it would also be possible to check alternatively over time the X-signal and the (X+Y)-signal, which requires a single detector and an optical or electrical switch.

[0058] In FIG. 5, a flow diagram illustrating an embodiment for controlling an optical transmitter, such as the optical transmitter 46 of FIG. 3 or 4 is shown. In step 68, a first pilot signal is superimposed along with the bias voltage on the in-phase driving voltages V .sub.HI, V.sub.VI, and a second pilot signal is superimposed on the quadrature driving voltages V.sub.HQ, V.sub.VQ. While the process 66 is not limited to this, in many important applications the first and second pilot signals can for example be superimposed on driving signals representing data signals with training sequences, wherein the training sequences corresponding to the first DPMZM 12 and the training sequence corresponding to the second DPMZM 28 are orthogonal to each other.

[0059] In step 70, the electrical output signal of the detector 58 is analyzed for a spectral component at a frequency of 2f.sub.p. In step 72, it is checked whether a spectral component at 2f.sub.p is present. If this is not the case, then the process proceeds to step 74, in which the pilot signals are switched off, and ends at 76. In an alternative embodiment, the pilot signals do not need to be switched off In this case it would be possible to go from step 72-No back to step 70 in an infinite loop.

[0060] However, if in step 72 a spectral component at 2f.sub.p is detected in the output signal of the detector 58, this is indicative of an inconsistent spectral polarity. In order to correct for this inconsistent spectral polarity, the process proceeds to step 78, in which the sign of the control signal V.sub.HQ is reversed. This can be regarded as a complex conjugation of the input signal in the equivalent complex base band, and likewise leads to a complex conjugation of the optical output signal of the first DP-MZM 12 in the equivalent base band. Note that this also has the same effect as changing the sign of the transfer function of the corresponding second inner MZM 24. The process then returns to step 70, and the loop is repeated, until in step 72, the spectral component at 2f.sub.p is detected. Under normal operation, after the inversion of the sign of V.sub.HQ in step 78, in the next round the spectral component at 2f.sub.p should vanish, and the program should come to an end 76. While not shown in FIG. 6, the process may include a timeout function which stops the process and effects an error message if the end of the procedure is not reached within a given time period.

[0061] The embodiments described above and the accompanying figures merely serve to illustrate the method according to the present invention, and should not be taken to indicate any limitation of the method. The scope of the patent is solely determined by the following claims.

LIST OF REFERENCE SIGNS

[0062] 10 modulator for PDM signals

[0063] 12 first DP-MZM

[0064] 14 input to first DP-MZM 12

[0065] 16 output of first DP-MZM 12

[0066] 18 first arm of first DP-MZM 12

[0067] 20 second arm of first DP-MZM 12

[0068] 22 first MZM

[0069] 24 second MM

[0070] 26 phase shifter

[0071] 28 second DP-MZM

[0072] 30 output of second DP-MZM 28

[0073] 32 first arm of second DP-MZM 28

[0074] 34 second arm of second DP-MZM 28

[0075] 36 third MZM

[0076] 38 fourth MZM

[0077] 40 phase shifter

[0078] 42 TE/TM polarization converter

[0079] 44 polarization beam splitter

[0080] 46 transmitter

[0081] 48 laser

[0082] 50 modulator for polarization-division multiplexed signals

[0083] 52 control unit

[0084] 54 pilot tone generator

[0085] 56 adder

[0086] 58 photodiode

[0087] 60 polarization beam splitter

[0088] 62 photodiode

[0089] 64 photodiode

[0090] 66 to 78 steps of flow diagram