Compensation system for adaptive equalization of an optical signal

[0001] When optical signals (pulses) are transmitted via an optical fiber, they are distorted. This distortion reduces the signal quality and considerably restricts the range without regeneration, especially at high bit rates. Wavelength-dependent attenuation can be compensated for via suitable amplifiers. Suitable measures are used in an attempt to compensate for other effects such as group delay time dispersion, polarization mode dispersion (PMD) and self phase modulation (SPM).

[0002] A method for PMD compensation via an optical first-order compensator and by via an adjustable electrical transversal filter are compared in Optics Communications, Volume 182, No. 1-3, pages 135-141.

[0003] Until now, compensation for group delay time dispersion, which is referred to as dispersion for short in the following text, mainly has been carried out by using special fibers whose group delay time characteristics are the inverse of those of the transmission fibers.

[0004] An adjustable electrical transversal filter which is used to compensate for PMD and dispersion is described in ECOC'99 Vol. 2, pp. 138-139, H. Buhlow et al.

[0005] The International Patent Application WO 00/50944 discloses an arrangement for dispersion compensation in which a Bragg grating filter is controlled as a function of the dispersion. By way of example, a dispersion analyzer assesses the eye opening or the bit error rate. The determined signal quality is used as a closed-loop control criterion. The arrangement likewise can be used for PMD compensation.

[0006] European Patent Application EP 0 740 173 A2 discloses the use of an optical transversal filter for dispersion compensation in individual channels, or for all channels at the same time, in a wavelength-division multiplexed system. This reduces the complexity in comparison to separate filters or compensation arrangements for individual channels. Compensation via a periodic filter has the same compensation effect for all channels but is not optimum in the presence of wavelength-dependent dispersion (higher-order dispersion), in which the gradient of the dispersion profile changes.

[0007] Furthermore, it is known, per se, for dispersion fundamental compensation to be carried out via an appropriate fiber, and for this to be combined with fine compensation using electrical filters.

[0008] An object of the invention is to derive an arrangement for adaptive signal equalization.

[0009] The advantage of the present invention is the capability of the optical filter to adapt itself to distortion which has been caused by dispersion, PMD or SPM. Compensation in the optical domain leads to better results than compensation in the electrical domain, since the equalization process is carried out before optoelectrical conversion. The photodiode used as an electrooptical converter is a nonlinear element with a square-law characteristic, which emits an envelope of the transmitted amplitude-modulated signal as a demodulated signal. The carrier and any phase information are lost during the conversion to an electrical signal.

[0010] In order to allow satisfactory compensation to be carried out, it is necessary to set an optimum complex transfer function for the compensation filter. As such, complex coefficients are set in a transversal filter. Although, in principle, both FIR filters (finite impulse response) and IIR filters (infinite impulse response) or a combination of both filter types can be used, transversal filters are, however, preferable since they can be adjusted well, possibly in combination with a recursive filter type.

[0011] In principle, all types of distortion can be compensated for, irrespective of their cause. A distinction is drawn between adaptive compensation and compensation via an inverse system. With an adaptive approach, optimum equalization is carried out without simulating the inverse system as a transfer function.

[0012] In an inverse system, the aim is to form the inverse transfer function. Initially, this is set such that it is fixed, but it also may include an adaptive component.

[0013] Dispersion compensation allows fundamental compensation to be carried out via a dispersion-compensating fiber (DCF) or a further (adjustable) compensation filter, thus reducing the required adjustment range of the filter.

[0014] A number of adaptive compensation devices also may be connected in series; for example, dispersion compensation for a transmission band having a number of wavelength-division multiplexed channels, and a channel-specific PMD compensation arrangement. The signal quality, such as the eye opening or bit error rate, or the spectrum is preferably assessed for open-loop compensation control, and the compensation is optimized by varying the coefficients.

[0015] Compensation likewise can be carried out using optical or electrical filters.

[0016] The particular advantage of dispersion compensation is the individual configuration of the dispersion profile, which is the inverse of that of the transmission fiber, of the optical compensation filter. The period (free spectral rangeFSR) of the filter covers a number of transmission channels, and preferably all the channels of a wavelength-division multiplexed signal, so that all the WDM channels, or even an entire transmission band, preferably can be compensated for using only one filter.

[0017] In an extended embodiment, the filter is adjusted adaptively, with characteristic transmission channels, or all the transmission channels, supplying quality criteria. This makes it possible to react to a change in the system parameters.

[0018] The optical filters which are normally used these days frequently have an unsatisfactory transfer function; in particular a nonideal phase response. This can be equalized by using a correction filter. The filter characteristics that are desired overall, such as a specific frequency response or a wide adjustment range of the passband or cut-off frequency, cannot be provided solely by the desired correction filter (filter function). An ideal component can be created by combining the nonideal component, such as a Bragg grating, with an optical correction filter for equalization.

[0019] Additional features and advantages of the present invention are described in, and will be apparent from, the following Detailed Description of the Invention and the Figures.

[0020] FIG. 1 shows an outline circuit diagram of the compensation system.

[0021] FIG. 2 shows an outline circuit diagram of a transversal filter.

[0022] FIG. 3 shows an outline circuit diagram of an IIR filter.

[0023] FIG. 4 shows a PMD compensation system.

[0024] FIG. 5 shows an outline circuit diagram of the system for wavelength-dependent dispersion compensation.

[0025] FIG. 6 shows an outline circuit diagram for adaptive dispersion compensation.

[0026] FIG. 7 shows an outline circuit diagram of a multistage compensation system.

[0027] FIG. 1 illustrates the principle of a transmission system with the compensation system FT, OEC, RX, AN, CON according to the present invention. A transmitter TX emits an optical signal OS which is modulated with a binary signal, is amplified in an amplifier AMP, and is transmitted via an optical fiber FIB. Further amplifiers and dispersion compensation devices, generally dispersion-compensating fibers, can be inserted on the transmission path. The optical signal OSV, which has been distorted by dispersion and nonlinear effects, is supplied to a closed-loop control device, which essentially includes an optical compensation filter FT, an analyzer AN and an open-loop control device CON. Since the analyzer and the open-loop control device operate electrically, conversion via an optoelectrical converter OEC is required prior to this. The receiver RX which is connected downstream from this amplifies the electrical signal linearly and, at its output, emits the data signal DS, which has not yet been limited or sampled. An analyzer AN which is connected to the output of the receiver, or via a splitter to the output of the compensation filter and to a further optoelectrical converter, assesses the quality of the data signal. This may be done by assessing the eye opening, the bit error rate or the spectrum. A closed-loop control signal QS which is derived from the quality criterion is supplied to the open-loop control device CON. This uses an optimization algorithm to change the complex coefficients of the compensation filter F1 until a quality maximum is reached.

[0028] The compensation and pulse forming can be further improved via an electrical filter EFI, which is connected downstream from the optical filter and is either fixed or adaptively adjustable.

[0029] An optical transversal filter (FIRfinite impulse response) whose basic design is illustrated in FIG. 2, contains delay time elements TD, multipliers x and at least one adder circuit . The complex coefficients k₁jb₁ to k_mjb_m of the multipliers can be changed by adjusting the phase (delay time change) and the attenuation in the various branches. The phase can be fine tuned, by way of example, by changing the refractive index of optical elements by heating them. A number of implementation options are known, so that there is no need to describe these here. However, in principle, the filter structure is irrelevant.

[0030] FIG. 3 illustrates a further compensation filter (IIRinfinite impulse response) which is provided by ring resonators RR₁, . . . , RR_M. The ring resonators are coupled to the fiber FIR via first amplitude adjustment elements EA₁-EA_M, and the distorted signal OSV is fed into the fiber FIR. These adjustment elements govern the real coefficients, while the delay time, and hence the phase, that is to say the imaginary coefficient, is in each case changed via second phase adjustment elements EP₁-EP_N which are inserted into the fibers FR₁-FR_M of the ring resonators. This system requires fewer coefficients for compensation.

[0031] The compensation system always optimizes the data signal irrespective of the cause of the distortion. A system such as this may be used, for example, to compensate for chromatic dispersion or polarization mode dispersion. Self phase modulation also can be compensated for by using an inverse compensation function to compensate for the transmission fiber characteristics which cause the SPM. This is also true for distortion which has a number of causes.

[0032] The compensation system may in each case be provided for one transmission channel or, (and this is more economical for WDM systems) have only one filter for a number of channels, or for all the channels, in a transmission band, provided that the periodicity of the filter is chosen such that it corresponds to the channel interval in the WDM system. This system is particularly suitable for fundamental compensation for group delay time dispersion and/or self phase modulation.

[0033] FIG. 4 illustrates an advantageous system, especially for PMD compensation. The series circuit formed by a polarization controller POLCON and a polarization splitter PSPL is supplied with the distorted optical signal OSV, which is split into two orthogonal signal components (polarization planes) OSV1 and OSV2. The polarization controller uses closed-loop control, which is not illustrated, to ensure stable polarization. The signal components OSV1 and OSV2 are supplied to a respective optical compensation filter FIP1 and FIP2, both of which use closed-loop control. In principle, one of the compensation filters may be replaced by a filter with a fixed setting or by a fixed or controlled delay time element, although this restricts the control capabilities. The signal components OSV1 and OSV2 are equalized, and their delay time differences are compensated for, individually, and they are then combined via a combiner COM to form a compensated signal OSK. After conversion to an electrical data signal DS, the signal quality is checked in the analyzer AN, and the filter coefficients are adjusted by the open-loop control device CON in accordance with the closed-loop control signal QS. Further electrical filters also may be added to the compensation system, for channel-specific equalization of PMD, SPM or wavelength-dependent dispersion.

[0034] FIG. 5 shows an optical transmission system with channel-specific dispersion compensation. A transmitter TX transmits an optical wavelength-division multiplexed signal OS_1-N, which is amplified by an amplifier AMP. This signal is transmitted via a transmission fiber FIB, in which a number of amplifiers also may be connected. The nonideal characteristics of the transmission fiber FIB result in a distorted optical signal OSV_1-N being produced from the signal OS_1-N. As is likewise shown above the compensation filter in FIG. 1, the dispersion DIW does not change linearly with the wavelength; that is to say, it has a higher order. The profile of the higher-order nonlinear component is referred to as the wavelength-dependent dispersion component. The dispersion can be largely compensated for via an inversely dimensioned compensation filter FIW.

[0035] For initial compensation, the distorted optical multiplexed signal OSV_1-N is passed through a dispersion-compensating fiber DCF. However, the dispersion-compensating fiber cannot be produced to be precisely the inverse of the dispersion-causing characteristics of the transmission fiber. The nonlinear wavelength-dependent dispersion component still remains. This is compensated for via an appropriate inverse profile of the transfer function of the compensation filter FIW.

[0036] The period (FSRfree spectral range) of the dispersion compensation filter FIW covers a number of channels, preferably the channels of a WDM multiplexed signal. The complex transfer function of a filter must be optimized for optimum dispersion compensation. Both FIR (for example, transversal or lattice filters) and IIR filters (for example, filters with ring resonators), whose complex transfer functions/coefficients are adjusted, are suitable for this purpose.

[0037] The compensated optical WDM signal OSK_1-N is split via an optical demultiplexer DMUX into individual signals (channels), which are converted in optoelectronic converters OEW and, after amplification in the receiver RX, are emitted as electrical data signals DS₁ to DS_N. The optimized transfer function is adjusted via open-loop control which, if possible, should be upgraded to closed-loop control.

[0038] In FIG. 5, the outputs of the wavelength-division demultiplexer DMUX are each followed by a correction filter KFI, in order to correct its nonideal transmission characteristics; in particular, its phase response. The correction filter may be an IIR (infinite impulse response) or FIR (finite impulse response) filter, whose complex transfer function is adjustable. Frequently, it is sufficient to compensate only for the phase response. The correction filter can be used wherever problems occur with undesirable filter characteristics. An open-loop or closed-loop controlled correction filter can react to changes in the filter and/or system parameters.

[0039] FIG. 6 shows an extended system which, in addition to the system known from FIG. 5, has a closed control loop for adjusting the dispersion compensation filter FIW. This includes an optical demultiplexer DMUX for splitting the multiplexed signal into individual signals, which are converted by optoelectrical converters OEW₁ to OEW_N into electrical data signals DS₁-DS_N.

[0040] An analyzer AN (or a number of analyzers) assesses (or assess) the signal quality of the electrical (not limited) data signals DS₁-DS_N of the transmission channels (for example in a chronological sequence) and emits a quality criterion QS_1-N to an open-loop control device CON, which adjusts the filter coefficients on the basis of an optimization algorithm. Fundamental equalization is carried out either via the dispersion-compensating fiber DCF or via an appropriately dimensioned, a fixed adjustable, or a variably adjustable periodic optical filter FID, which is connected upstream of or downstream from the compensation filter FIW.

[0041] FIG. 7 shows a compensation device for dispersion and PMD compensation for WDM signals, which contains a number of compensation systems, with a respective compensation filter FI1 and one of the filters FI2₁-FI2_N being arranged functionally in series for each channel and being controlled by associated open-loop compensation control devices CON2₁ to CON2_N and by a common open-loop dispersion control device CON.

[0042] A WDM signal OS_1-N is once again transmitted as the optical signal. The filters are preceded by a dispersion-compensating fiber DCF for initial compensation, since the control range of the filters is limited. The initial compensation likewise can be carried out jointly for all the channels via a further periodic filter.

[0043] The initial compensation in the system is carried out via the dispersion-compensating fiber DCF jointly for all the channels. The wavelength-dependent remaining dispersion is compensated for in the adjustable dispersion compensation filter FI1, preferably a transversal filter, whose periodicity (FSRfree spectral range) preferably covers all the WDM channels in a transmission band.

[0044] The quality criteria QS₁-QS_N (closed-loop control criteria) are derived from a number of channels, or from all of the channels, and are passed to the first open-loop control device CON1, which adjusts the coefficients of the first filter.

[0045] The dispersion compensation filter FI1 also may be a periodic filter, and may replace the dispersion-compensating fiber. In one variant, a filter 1, which is periodic with the channel interval, can be used for equalization in the presence of group delay time dispersion and/or self phase modulation compensation.

[0046] However, polarization mode dispersion should be compensated for channel-by-channel with a short time constant. The first compensation filter is therefore followed by an optical demultiplexer DEMUX, which separates the channels/individual signals in the WDM signal. The signals DS₁-DS_N which are transmitted in the individual channels are compensated for as far as PDM is concerned separately in the second compensation filters FI2₁ to FI2_N, with the quality criteria QS₁, . . . , QS_N for each channel being determined in a separate analyzer AN₁ to AN_N. Different analyzers also may be used for dispersion compensation and for PMD compensation.

[0047] This compensation device also may be supplemented by further fixed setting or adaptive elements for compensation, by a further adaptive compensation filter, or by electrical filters. Equivalent optical PMD compensation may be carried out, especially in the case of linear optoelectrical conversion.

[0048] Although the present invention has been described with reference to specific embodiments, those of skill in the art will recognize that changes may be made thereto without departing from the spirit and scope of the present invention as set forth in the hereafter appended claims.