JP4933505B2 - Interference reduction method and interference reduction apparatus - Google Patents

Description

  The present invention relates to an interference removal reducing method and an interference removal reducing apparatus for compensating for a signal delay caused by chromatic dispersion in optical communication.

In optical communication, chromatic dispersion caused by an optical fiber as a propagation path is a big problem. Conventionally, signal delay due to chromatic dispersion has been compensated by using a dispersion compensating fiber having inverse characteristics (see, for example, Non-Patent Document 1).
“Next-Generation Ultra-High-Speed Optical Communication Technology -Technical Issues and Overcoming Solutions for Optical Device Development”, First Edition, Technical Information Association, Inc., June 27, 2003, p. 112-p. 118

  However, in the above-described prior art, it is necessary to install an optimum dispersion compensating fiber for each optical communication path, which hinders flexible network design such as path change.

  The present invention has been made in consideration of such circumstances, and its purpose is interference that can reduce interference power caused by chromatic dispersion and frequency-dependent circuit characteristics without hindering flexible network design. An object of the present invention is to provide a reduction reduction method and an interference reduction reduction apparatus.

  The present invention relates to an interference reduction method in optical communication using an optical fiber, an analog / digital conversion step for converting an optical signal into a digital signal, and the digital signal is branched into two or more, and each of the branched digital signals is divided. On the other hand, a serial / parallel conversion step for converting into a parallel signal after giving a delay amount proportional to the order of branching, and a Fourier transforming into a signal in the frequency domain by performing a Fourier transform on the parallel signal. A distance information output step of estimating or storing and outputting in advance a chromatic dispersion amount caused by the optical fiber based on the conversion step, the distance propagated in the optical fiber and the characteristics of the optical fiber with respect to the signal light wavelength; and the wavelength Based on the amount of dispersion, the difference in arrival time due to frequency for the Fourier-transformed frequency domain signal A coefficient multiplying step for multiplying a first correction coefficient to be corrected, and an inverse Fourier transform for converting the frequency domain signal multiplied by the first correction coefficient into a time domain signal by performing an inverse Fourier transform. A step of selecting a time domain signal excluding a guard band portion from the time domain signal subjected to the inverse Fourier transform, and arranging the selected time domain signals in the order of branching. An interference reduction method comprising: an output step of outputting in place.

  The present invention is also a method for reducing interference in optical communication using an optical fiber, a band dividing step for separating an optical signal into signals of different frequency bands, and converting the band-divided signals into digital signals, respectively. An analog-to-digital conversion step for adding a delay amount proportional to the divided order, and the digital signal to which the delay amount is added are further branched into two or more, and each branched digital signal is branched. A serial / parallel conversion step that converts a delay signal proportional to the order and then converts the signal into a parallel signal, and a parallel signal from the different frequency band is synthesized, and a Fourier transform is performed on the frequency domain signal obtained by the synthesis. A Fourier transform step to convert the signal into a frequency domain signal, and the distance and signal light wavelength propagated by the optical fiber. A distance information output step for estimating or storing and outputting the chromatic dispersion amount generated by the optical fiber based on the characteristics of the optical fiber to be used, and the delay amount used for each of the chromatic dispersion amount and the branched frequency band. A frequency multiplying step of multiplying the Fourier-transformed frequency domain signal by a first correction coefficient for correcting an arrival time difference due to frequency; and a frequency domain signal multiplied by the first correction coefficient. An inverse Fourier transform step for transforming into a time domain signal by performing an inverse Fourier transform on the signal, and a signal for selecting a time domain signal excluding a guard band portion from the time domain signal subjected to the inverse Fourier transform A cut-out step, and an output step of outputting the selected signals in the plurality of time domains by rearranging them in the order of branching. Which is the interference reduction method characterized.

  The present invention is also a method for reducing interference in optical communication using an optical fiber, a band dividing step for separating an optical signal into signals of different frequency bands, and converting the band-divided signals into digital signals, respectively. An analog-to-digital conversion step that adds a delay amount proportional to the divided order, and a digital signal to which the delay amount is added is further branched into two or more, and a specific delay amount is given to each branched digital signal In addition, a serial / parallel conversion step for converting into a parallel signal, a Fourier transform step for converting into a frequency domain signal by performing a Fourier transform on each parallel signal, and a distance propagated by the optical fiber. And the amount of chromatic dispersion produced by the optical fiber based on the characteristics of the optical fiber with respect to the signal light wavelength, or The difference in arrival time due to frequency is calculated for the Fourier-transformed frequency domain signal based on the distance information output step for storing and outputting, and the chromatic dispersion amount and the delay amount used in each of the branched frequency bands. A coefficient multiplying step for multiplying a first correction coefficient to be corrected, and an inverse Fourier transform for converting the frequency domain signal multiplied by the first correction coefficient into a time domain signal by performing an inverse Fourier transform. A step of selecting a time domain signal excluding a guard band portion from the time domain signal subjected to the inverse Fourier transform, and arranging the selected time domain signals in the order of branching. An interference reduction method comprising: an output step of outputting in place.

  Further, in the interference reduction method of the present invention, the coefficient multiplying step is performed on the Fourier-transformed frequency domain signal in addition to the first correction coefficient for correcting the arrival time difference due to the estimated chromatic dispersion amount. A second correction coefficient for correcting the frequency pass characteristic of the circuit is multiplied.

  In the interference reduction method of the present invention, before executing the band dividing step, the analog / digital conversion step converts an optical signal into a digital signal, and the band dividing step converts at the analog / digital conversion step. The digital signal is separated into two or more different frequency band signals by a digital filter.

  In the interference reduction method of the present invention, the analog-digital conversion step converts the band-divided signal into a low-frequency signal by performing frequency conversion according to a center frequency, and converts the low-frequency signal. The signal is converted into a digital signal, and the inverse Fourier transform step rearranges the signal in the frequency domain multiplied by the first correction coefficient in the order of the original frequency according to the center frequency, By performing inverse Fourier transform, the signal is converted into a time domain signal.

  The present invention also provides an interference reduction apparatus for optical communication using an optical fiber, a band dividing circuit for separating an optical signal into signals of different frequency bands, and converting the band-divided signals into digital signals, respectively. An analog / digital conversion circuit that adds a delay amount proportional to the divided order, and a digital signal to which the delay amount is added are further branched into two or more, and a specific delay amount is given to each branched digital signal. In addition, a serial / parallel conversion circuit that converts the signal into a parallel signal, a Fourier transform circuit that converts the parallel signal into a frequency domain signal by performing a Fourier transform on the parallel signal, and a distance propagated by the optical fiber, A distance information output circuit for estimating the amount of chromatic dispersion generated by the optical fiber based on the characteristics of the optical fiber with respect to the signal light wavelength; Based on the obtained chromatic dispersion amount, the Fourier-transformed frequency domain signal is multiplied by a coefficient correction circuit that multiplies a first correction coefficient for correcting an arrival time difference due to frequency, and the first correction coefficient. An inverse Fourier transform circuit that converts the signal in the frequency domain into a time domain signal by performing an inverse Fourier transform, and a time domain in which the guard band portion is removed from the inverse Fourier transformed time domain signal An interference reduction apparatus, comprising: a signal cut-out circuit that selects a signal of the above; and an output circuit that rearranges and outputs the selected plurality of time-domain signals in an appropriate order.

  According to the present invention, the received signal is divided into parallel signals, a necessary time slot is extracted by delay processing, and the correction is performed by multiplying the correction coefficient based on the chromatic dispersion amount of the optical fiber and the frequency dependence characteristic of the circuit. As a result, interference power caused by chromatic dispersion and frequency-dependent circuit characteristics can be reduced. In addition, by performing the above processing after dividing each frequency band, the amount of computation in one processing block can be suppressed by parallel signal processing divided in the time domain and the frequency domain. The influence of interference can be reduced.

  Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

A. First Embodiment First, a first embodiment of the first embodiment of the present invention will be described.
FIG. 1 is a block diagram showing a configuration of a transmission unit according to the first embodiment of the present invention. In FIG. 1, 101 is an analog / digital conversion circuit, 102-1 to 102-K are serial / parallel conversion circuits, 103-1 to 103-K are Fourier transform circuits, 104-1 to 104-K are coefficient multiplication circuits, Reference numerals 105-1 to 105-K denote inverse Fourier transform circuits, 106-1 to 106-K denote signal extraction circuits, 107 denotes an output circuit, and 109 denotes a distance information output circuit. Here, K indicates the number of branches.

When the optical signal is received, the analog / digital conversion circuit 101 converts it into a digital signal. At this time, the optical signal may be directly converted into a digital signal that is an electric signal, or the optical signal may be converted into an electric signal and then converted into a digital signal. After being converted to a digital signal, the digital signal is branched into K pieces and input to serial / parallel conversion circuits 102-1 to 102-K. The serial / parallel conversion circuits 102-1 to 102-K give delays of _Td (1) to _Td (K) to the input signals, respectively, and convert serial signals at different positions into parallel signals.

Here, it is assumed that Fourier transform is performed for each _Nf- point received signal, and the input signal is represented as s (t). t is a reception timing. Here, if the delay T _d (i) given by the i-th serial / parallel conversion circuit 102-i is defined as T _d0 × (i−1) using the delay step width T _d0 , the i-th serial / parallel conversion circuit 102-i is defined. parallel conversion circuit 102-i _{is, s (T 0 + (i} -1) T d0 -T g) ~s (T 0 + (i-1) T d0 + N f -T g -1) of _{N f} signal Are output to the Fourier transform circuit 103-i as parallel signals.

Here, N _f is the number of points to be considered in the Fourier transform, T _g is the guard band length used in the clipping circuit, and satisfies N _f −2T _g ≧ T _d0 . T ₀ can take an arbitrary number. In this description, it is assumed that the reception timing is the head position of a signal in which interference is reduced.

The Fourier transform circuit 103-i performs Fourier transform on the input parallel signal, and the frequency domain signal s _f (T ₀ + (i−1) T _d0 , F ₀ − (N _f / 2-1). _{F w), ..., s f} (T 0 + (i-1) T d0, F 0), ..., s f (T 0 + (i-1) T d0, F 0 + (N f / 2) F _w ). Here, s _f (t, f) is a signal in the frequency region corresponding to the head position reception timing t excluding the guard band and the frequency f. F ₀ is the center frequency and F _w is the frequency bandwidth of the frequency channel.

  The distance information output circuit 109 estimates the amount of chromatic dispersion using the transmission distance information of the optical signal and the information on the type of optical fiber. The shift in arrival time with respect to the wavelength of light, that is, the amount of phase rotation, can be calculated from the type of transmission distance and optical fiber. Can be given. The calculation of the amount of phase rotation is described in, for example, the document “Govind P. Agrawal,“ Nonlinear fiber optics, ”Academic press, 2006”.

Here, L is the transmission distance [km], λ is the wavelength [nm], c is the speed of light 3 × 10 ⁻⁷ [km / ps], D is the chromatic dispersion coefficient [ps / nm / km], and D _slope is the dispersion slope. coefficient ^{[ps / nm 2 / km]} , f c is the frequency of the optical carrier. Since the chromatic dispersion and the dispersion slope are specific values depending on the type of optical fiber, the phase rotation can be estimated if the wavelength λ of the light used and the transmission distance L of the optical fiber are known. In equation (1), the phase rotation can be estimated with D _slope = 0 for a fiber that is greatly affected by chromatic dispersion, or the phase rotation can be estimated with D = 0 for a fiber that is greatly affected by wavelength slope. . The distance information output circuit 109 outputs the phase rotation estimated in this way to the coefficient multiplication circuits 104-1 to 104-K. The distance information output circuit 109 may output the chromatic dispersion amount stored in advance in a storage unit (not shown).

  The coefficient multiplier circuit 104-i corrects the arrival time difference due to the light frequency by multiplying the frequency domain signal by using the complex conjugate of the phase rotation g (f) input from the distance information output circuit 109 as a correction coefficient. The obtained frequency domain signal can be expressed by the following equation (2).

The superscript * represents a complex conjugate number. S _f ′ (T ₀ + (i−1) T _d0 , F ₀ − (N _f / 2-1) F _w ) corrected by Expression (2),..., S _f ′ (T ₀ + (i− 1) T _d0 , F ₀ ),..., S _f ′ (T ₀ + (i−1) T _d0 , F ₀ + (N _f / 2) F _w ) are output to the inverse Fourier transform circuit 105-i.

Inverse Fourier transform circuit 105-i, an inputted signal _{s f '(T 0 + (} i-1) T d0, F 0 - (N f / 2-1) F w), ..., s f' (T _{0 + (i-1) T} d0, F 0), ..., s f '(T 0 + (i-1) T d0, F 0 + a _{(N f / 2) F w} ), again, the time domain The signal is converted into a signal, s ′ (T ₀ + (i−1) T _d0 −Tg) to s ′ (T ₀ + (i−1) T _d0 + N _f −T _g −1) is calculated, and a signal extraction circuit 106-i.

Signal extracting circuit 106-i deletes the data corresponding to the front and rear a _{T g} from the input signal (guard band _{portion), s'(T 0 + (} i-1) T d0) ~s' (T 0 A signal of N _f −2T _g of + (i−1) T _d0 + N _f −2T _g −1) is output to the output circuit 107.

The output circuit 107 combines the signals input from the signal cutout circuits 106-1 to 106-K and converts them into a serial signal. When N _f −2T _g = T _d0 is satisfied, the signal extraction circuits 106-1, 106-2,..., 106-K have s ′ (T ₀ ) to s ′ (T ₀ + T _d0 −1). _{, s'(T 0 + T d0} ) ~s' (T 0 + 2T d0 -1), ..., s'(T 0 + (K-1) T d0) ~s' (T 0 + KT d0 -1) is input Then, by rearranging these signals into serial signals, s ′ (T ₀ ) to s ′ (T ₀ + KT _d0 −1) can be obtained. Since the obtained signal has been corrected for arrival delay due to chromatic dispersion, it is output after the influence of chromatic dispersion is removed.

Further, in order to process the signal of KT _d0 at a time, T ₀ + KT _d0 −T _{g to} T ₀ + N _f + (K−1) KT _d0 + KT _d0 −T _g −1 with the guard band length T _g added A signal corresponding to the reception timing can be continuously processed by outputting to the serial / parallel conversion circuits 102-1 to 102-K.

FIG. 2 is a sequence diagram showing the flow of received signals when K = 2, T _g = 4, T _d0 = 8, and N _f = 16 in the first embodiment. FIG. 2 shows a state in which chromatic dispersion of a signal after reception timing T ₀ is compensated. Received _{signal, s (T 0) ~s (} T 0 +15) a first _{block, s (T 0 +16) ~s} (T 0 +31) a second _{block, s (T 0 +32) ~s} (T 0 the +47) as a third block, and it shall be sequentially processed for each _{N f} signal. In the illustrated reception signal series, the reception signal is inclined obliquely in order to simply show that the arrival time varies depending on the frequency component due to chromatic dispersion. Assuming that the chromatic dispersion coefficient is positive, it indicates that the high frequency component has arrived first. For example, when attention is paid to the reception timing T ₀ , the signal of the 0th block and the signal of the 1st block are mixed. It is received and this causes interference, which degrades the signal quality.

First, paying attention to the correction of the received signal of the first block, step S1-1 obtains s (T ₀ -4) to s (T ₀ +11) for the received signal sequence. Step S2-1 adds a delay _{T d0} = 8 in the same received sequence _signal to obtain the _{s (T 0 +4) ~s (} T 0 +19). In step S1-2 and step S2-2, _{respectively,} and _{s (T 0 -4) ~s (} T 0 +11), s (T 0 +4) and to perform Fourier transform _~s (T 0 +19), Frequency obtain domain signal _{s f (T 0, -8F w} ) ~s f (T 0, 7F w), s f (T 0 + 8, -8F w) ~s f a (T 0 + 8,7F _w), respectively. Here, the center frequency is represented as F ₀ = 0.

In step S1-3 and step S2-3, the phase rotation correction calculated from equation (1) is multiplied as in equation (2), and the corrected frequency domain signal s _f ′ (T ₀ , -8Fw) is obtained. _{~s f '(T 0, 7F} w), s f' (T 0 + 8, -8F w) ~s f ' get the (T 0 + 8,7F _w). In step S1-4 and step S2-4, respectively, performs an inverse Fourier _{transform, s'(T 0 -4) ~s'} (T 0 +11) and _{s'(T 0 +4) ~s' (} T 0 +19) is calculated.

In step S1-5 and step S2-5, it removes the guard band _T g = _4, s'and _{(T 0) ~s' (T 0} +7), s'(T 0 +8) ~s' (T 0 +15) respectively. By converting the obtained signal into a serial signal, s ′ (T ₀ ) to s ′ (T ₀ +7) and s ′ (T ₀ +8) to s ′ (T ₀ +15) can be obtained. . Further, by continuously processing the received signal of the second block as T ₀ = T ₀ +16, it is possible to continuously correct all received signal sequences. Further, by setting N _f −2T _g > T _d0 , it is possible to overlap the signal areas to be cut out.

FIG. 3 is a conceptual diagram showing a signal overlap region when _Tg is 2 in the configuration example of the interference reduction method shown in FIG. In the overlap _{region, s'(T 0 +5) ~s'} (T 0 +8) is first signal extracting circuit (e.g., signal extraction circuit 106-1 in FIG. 1) and the second signal extraction circuit (e.g., Fig. 1 and the signal extraction circuit 106-2). Here, in order to distinguish, the signal obtained from the first signal clipping circuit is represented by s ₁ ′ (T ₀ +5) to s ₁ ′ (T ₀ +8), and the signal obtained from the second signal clipping circuit is represented by This is expressed as s ₂ ′ (T ₀ +5) to s ₂ ′ (T ₀ +8). If there is an overlap area, the output circuit 107 adds these signals according to the following equation (3) and outputs the result.

Here, w _{k, 1} and w _{k, 2} are weighting values, and are given as w _{k, 1} = w _{k, 2} = 1 / √2 in equal power synthesis. Or, the closer to the boundary area is affected by noise and interference, so a small weight is given to the signal that is close to the boundary with the received signal (shaded part) to be discarded, and a large weight is given to the signal far from the boundary Can do.

Taking FIG. 3 as an example, s ₁ ′ (T ₀ +5), s ₁ ′ (T ₀ +6), s ₁ ′ (T ₀ +7), and s ₁ ′ (T ₀ +8) are respectively received by truncation. The distance from the signal (shaded area) is 4, 3, 2, 1, and s ₂ ′ (T ₀ +5), s ₂ ′ (T ₀ +6), s ₂ ′ (T ₀ +7), s ₂ ′ (T ₀ +8) has distances 1, 2, 3, and 4 from the reception signal (hatched portion) to be discarded. The signal s ′ (T ₀ + k) is given by the following equation (4).

Here, d _{1, k} and d _{2, k} are the distance from the boundary of the signal cut out from the first signal cutout circuit and the distance from the boundary of the signal cut out from the second signal cutout circuit, respectively. If w (d _{1, k} ) ² + w (d _{2, k} ) ² = 1 and a> b, then w (a)> w (b).

  As a method of determining w (d), the weight can be determined in advance according to the above rules, or can be adaptively changed according to the communication conditions. By using the transmitted known signal part and the decoded signal, the noise and interference power distribution near the cut-out boundary is actually measured, and the signal synthesis by the maximum ratio synthesis is performed from the signal-to-interference noise ratio (SINR). It can also be used. Alternatively, several weight patterns can be determined in advance, a rough estimation of the signal-to-noise interference ratio can be performed, and the weight to be used can be selected.

Hereinafter, a second embodiment of the first embodiment of the present invention will be described with reference to FIG. When the optical signal is received, the analog / digital conversion circuit 101 converts it into a digital signal. At this time, the optical signal may be directly converted into a digital signal, or the optical signal may be converted into an electric signal and then converted into a digital signal. After being converted to a digital signal, the digital signal is branched into K pieces and input to serial / parallel conversion circuits 102-1 to 102-K. Serial-parallel conversion circuit 102-1 to 102-K gives the delay with the deviation of each _{T 0} of 0~ (k-1) _{T 0} on the input signal, a parallel signal the serial signal of the delay _{T 0} min Convert to

Here, the input signal is represented as s (t). t is a reception timing. The i-th serial / parallel conversion circuit 102-i uses the T ₀ signal from s (T ₀ + (i−1) T _d0 ) to s (T ₀ + iT _d0 −1) as a parallel signal, and performs a Fourier transform circuit 103. Output to -i.

The Fourier transform circuit 103-i adds 0 before and after the input T ₀ parallel signal, and [0,..., 0, s (T ₀ + (i−1) T _d0 ) _{,. , s (T 0 + iT d0} -1), 0, ···, performing a Fourier transform on the _{N f} points of the signal 0]. T ₀ can take an arbitrary number. In this description, it is assumed that the reception timing is the head position of the signal from which interference is removed.

The Fourier transform circuit 103-i performs Fourier transform on the input parallel signal, and the frequency domain signal s _f (T ₀ + (i−1) T _d0 , F ₀ − (N _f / 2-1). _{F w), ..., s f} (T 0 + (i-1) T d0, F 0), ..., s f (T 0 + (i-1) T d0, F 0 + (N f / 2) F _w ). Here, s _f (t, f) is a signal in the frequency region corresponding to the head position reception timing t excluding the guard band and the frequency f. F ₀ is the center frequency and F _w is the frequency bandwidth of the frequency channel.

  The distance information output circuit 109 estimates the amount of chromatic dispersion using the transmission distance information of the optical signal and the information on the type of optical fiber. Subsequently, the coefficient multiplication circuit 104-i multiplies the frequency domain signal by a coefficient as shown in Equation (2) to compensate for chromatic dispersion.

S _f ′ (T ₀ + (i−1) T _d0 , F ₀ − (N _f / 2-1) F _w ) corrected by Expression (2),..., S _f ′ (T ₀ + (i− 1) T _d0 , F ₀ ),..., S _f ′ (T ₀ + (i−1) T _d0 , F ₀ + (N _f / 2) F _w ) are output to the inverse Fourier transform circuit 105-i.

Inverse Fourier transform circuit 105-i, an inputted signal _{s f '(T 0 + (} i-1) T d0, F 0 - (N f / 2-1) F w), ..., s f' (T _{0 + (i-1) T} d0, F 0), ..., s f '(T 0 + (i-1) T d0, F 0 + a _{(N f / 2) F w} ), again, the time domain The signal is converted into a signal, s ′ (T ₀ + (i−1) T _d0 −T _d1 ) to s ′ (T ₀ + (i−1) T _d0 + T _d2 −1) is calculated, and the signal clipping circuit 106 − output to i. Here, T _d1 and T _d2 are the number of 0 added by the Fourier transform circuit 103-i before the signal and the number of 0 added after the signal, respectively.

In the second form of the first embodiment, the signal clipping circuit 106-i outputs the input signal to the output circuit 107. The output circuit 107 synthesizes and outputs the input signals. Here, when the signal input from the i-th signal extraction circuit 106-i is s _i ′ (t), the output circuit 107 has s ′ (t) = (s ₁ ′ (t) + s ₂ ′ (t). +... + S _K ′ (t)) is calculated and output. Here, when there is no corresponding t, s _i ′ (t) is calculated as 0. Since the i-th signal extraction circuit 106-i outputs a signal at time t corresponding to T ₀ + (i−1) T _{d0 to} T ₀ + (i−1) T _d0 + T _d2 −1, K signals of t are not necessarily inputted, and K ′ ≦ K. Since the obtained signal has been corrected for arrival delay due to chromatic dispersion, it is output after the influence of chromatic dispersion is removed.

In addition, in order to process the signal of KT _{d0 at} a time, a signal corresponding to the reception timing of T ₀ + KT _{0 to} T ₀ + 2KT ₀ −1 is output to the serial / parallel conversion circuits 102-1 to 102-K. Can be processed continuously.

FIG. 4 shows the flow of a received signal when K = 2, T _d0 = 8, N _f = 16, and T _d1 = T _d2 = 4 in the second form of the first embodiment. It is a sequence diagram. FIG. 4 shows a state in which the chromatic dispersion of the signal after the reception timing T ₀ is compensated. Received _{signal, s (T 0) ~s (} T 0 +15) a first _{block, s (T 0 +16) ~s} (T 0 +31) a second _{block, s (T 0 +32) ~s} (T 0 the +47) as a third block, and it shall be sequentially processed for each _{N f} signal.

First, paying attention to the correction of the received signal of the first block, step S1-1 obtains s (T ₀ ) to s (T ₀ +7) for the received signal sequence, and further four each before and after. 0 signal is added. Step S2-1 adds a delay _{T d0} = 8 in the same received sequence _signal, obtains the _{s (T 0 +8) ~s (} T 0 +15) adds a 0 signal by four back and forth. In steps S1-2 and S2-2, [0, ..., 0, s (T ₀ ), ..., s (T ₀ +7), 0, ..., 0] and [ 0,..., 0, s (T ₀ +8),..., S (T ₀ +15), 0,..., 0], and the frequency domain signal s _f (T ₀ , obtaining _{-8F w) ~s f (T 0} , 7F w), s f (T 0 + 8, -8F w) ~s f a (T 0 + 8,7F _w), respectively. Here, the center frequency is represented as F ₀ = 0.

In step S1-3 and step S2-3, the phase rotation correction calculated from equation (1) is multiplied as in equation (2), and the corrected frequency domain signal s _f ′ (T ₀ , -8Fw) is obtained. _{~s f '(T 0, 7F} w), s f' (T 0 + 8, -8F w) ~s f ' get the (T 0 + 8,7F _w). In step S1-4 and step S2-4, respectively, performs an inverse Fourier _{transform, s'(T 0 -4) ~s'} (T 0 +11) and _{s'(T 0 +4) ~s' (} T 0 +19) is calculated.

In step S1-5, signals corresponding to the same time are added to obtain s ′ (T ₀ ) to s ′ (T ₀ +7). Signal _{s'(T 0 +8) ~s' (} T 0 +15) again performs the process from step S1-1 for the second block, _s'in the second block (T 0 +8) ~s' (T ₀ + 15) and adding.

However, as can be seen from FIG. 3, in the second embodiment of the first embodiment, the high frequency components of the signals of s ′ (T ₀ ) and s ′ (T ₀ +1) are leaking into the 0th block. Therefore, in order to obtain the same effect as that of the first embodiment of the first embodiment without considering the 0th block, it is necessary to set T ₀ smaller than that of the first embodiment of the first embodiment.

B. Second Embodiment Next, a second embodiment of the present invention will be described.
FIG. 5 is a block diagram showing a configuration of a transmission unit according to the second embodiment of the present invention. In FIG. 5, 200 is a band dividing circuit, 201-1 to 201-B are analog / digital conversion circuits, 202-1-1 to 202-KB are serial / parallel conversion circuits, and 203-1 to 203-K are Fourier transform circuit, 204-1 to 204-K are coefficient multiplication circuits, 205-1 to 205-K are inverse Fourier transform circuits, 206-1 to 206-K are signal extraction circuits, 207 is an output circuit, and 209 is distance information. This is an output circuit.

When an optical signal is received, the band dividing circuit 200 branches the received signal into B pieces, and each corresponds to the 1st to Bth frequency bands using any of a high pass filter, a band pass filter, and a low pass filter. A reception signal is formed and output to the analog / digital conversion circuits 201-1 to 201-B. At this time, a filter may be used for the optical signal, or the filter may be used after the optical signal is converted into an electric signal. The analog / digital conversion circuits 201-1 to 201-B determine delays T _B (1) to T _B (B) to be added to the digital signal according to the distance input from the distance information output circuit 209.

The analog / digital conversion circuits 201-1 to 201-B convert an input analog signal into a digital signal, shift the signal by a delay amount, branch the digital signal into K pieces, and a serial / parallel conversion circuit 202-1. Output to −1 to 202-KB. The i-th serial / parallel conversion circuits 202-i-1 to 202 -i-B in each frequency band further give a delay of T _d (i) to the input signal, and convert serial signals at different positions into parallel signals. Convert.

Here, the delay T _B (m) added by the mth analog-digital conversion circuit 201-m is T _B0 × (m−1), and the i-th serial / parallel conversion circuit 202-i− in each frequency band. Assuming that the delay T _d (i) added by 1 to 202-i-B is T _d0 × (i−1), the i-th serial / parallel conversion circuit 202-im in the m-th frequency band is _{s (T 0 + (m-} 1) T B0 + (i-1) T d0 -T g) ~s (T 0 + (m-1) T B0 + (i-1) T d0 + N f -T g the signal of _{N f} -1) as a parallel signal, and outputs the Fourier transform circuit 203-i. Here, N _f is the number of points to be considered in the Fourier transform, T _g is the guard band length used in the clipping circuit, and satisfies N _f −2T _g ≧ T _d0 . T ₀ can take an arbitrary number. In this description, it is assumed that the reception timing is the head position of a signal in which interference is reduced.

Fourier transform circuit 203-i combines the B-number of parallel signals inputted from the serial-parallel conversion circuit 202-i-1~202-i- B, performs Fourier transform to obtain _{N f} number of parallel signals , Frequency domain signals s _f (T ₀ + (i−1) T _d0 , F ₀ − (N _f / 2-1) F _w ),..., S _f (T ₀ + (i−1) T _d0 , F ₀ ),..., S _f (T ₀ + (i−1) T _d0 , F ₀ + (N _f / 2) F _w ). Here, s _f (t, f) is a reception signal corresponding to the head position reception timing t excluding the guard band and the frequency f. F ₀ is the center frequency and F _w is the frequency bandwidth of the frequency channel.

  The distance information output circuit 209 estimates the chromatic dispersion amount using the transmission distance information of the optical signal and the information on the type of the optical fiber. The shift in arrival time with respect to the wavelength of light, that is, the amount of phase rotation can be calculated from the transmission distance and the type of optical fiber, and the phase rotation in the carrier of frequency f is given by equation (1). Can do.

  In the second embodiment, since the Fourier transform position differs for each frequency band, it is necessary to further multiply the Fourier transform position correction coefficient. The correction coefficient h (f, m) can be expressed by the following equation (5).

  The coefficient multiplication circuit 204-i corrects the complex conjugate of the phase rotation g (f) input from the distance information output circuit 209 and the delay coefficient given by the analog / digital conversion circuit 201-i h (f, m). Is multiplied as a correction coefficient by the frequency domain signal to correct the arrival time difference due to the light frequency. The obtained frequency domain signal is expressed by the following equation (6).

The superscript * represents a complex conjugate number. S _f ′ (T ₀ + (i−1) T _d0 , F ₀ − (N _f / 2-1) F _w ) corrected by Expression (6),..., S _f ′ (T ₀ + (i− 1) T _d0 , F ₀ ),..., S _f ′ (T ₀ + (i−1) T _d0 , F ₀ + (N _f / 2) F _w ) are output to the inverse Fourier transform circuit 205-i.

Inverse Fourier transform circuit 205-i, an inputted signal _{s f '(T 0 + (} i-1) T d0, F 0 - (N f / 2-1) F w), ..., s f' (T _{0 + (i-1) T} d0, F 0), ..., s f '(T 0 + (i-1) T d0, F 0 + a _{(N f / 2) F w} ), again, the time domain The signal is converted, s ′ (T ₀ + (i−1) T _d0 −T _g ) to s ′ (T ₀ + (i−1) T _d0 + N _f −T _g −1) is calculated, and signal extraction is performed. It outputs to the circuit 206-i.

Signal extracting circuit 206-i deletes the data corresponding to the front and rear a _{T g} from the input signal (guard band _{portion), s'(T 0 + (} i-1) T d0) ~s' (T 0 The signal of N _f −2T _g of + (i−1) T _d0 + N _f −2T _g −1) is output to the output circuit 207.

The output circuit 207 combines the signals input from the signal cutout circuits 206-1 to 206-K and converts them into a serial signal. When N _f −2T _g = T _d0 is satisfied, the signal extraction circuits 206-1, 206-2,..., 206-K have s ′ (T ₀ ) to s ′ (T ₀ + T _d0 −1). _{, s'(T 0 + T d0} ) ~s' (T 0 + 2T d0 -1), ..., s'(T 0 + (K-1) T d0) ~s' (T 0 + KT d0 -1) is input Then, by rearranging these signals into serial signals, s ′ (T ₀ ) to s ′ (T ₀ + KT _d0 −1) can be obtained. Since the obtained signal has been corrected for arrival delay due to chromatic dispersion, it is output after the influence of chromatic dispersion is removed. Moreover, for processing signals of KT _d0 at a time, by decoding the signals from T 0 ₊ KT _d0, it is possible to process the received signal continuously.

By using the second embodiment described above, it is possible to reduce the temporal spread of the signal due to chromatic dispersion. Therefore, the distance information output circuit 209 depending on the transmission distance can specify the size of the delay _T B (m) given by the analog-to-digital converter 201-1 to 201-B. For example, according to the transmission distance, a table that gives the corresponding delay T _B (m) is prepared in advance, and T _B (m) can be determined according to the rules of this table.

Also, the number B of frequency bands to be divided by the band dividing circuit 200 can be designated by the distance information output circuit 209. Therefore, the distance information output circuit 209 depending on the transmission distance can specify the size of the delay _T B (m) given by the analog-to-digital converter 201-1 to 201-B. For example, according to the transmission distance, a table giving the number of band divisions B is prepared in advance, and B can be determined according to the rules of this table. Alternatively, a table that gives the number of band divisions B and the delay T _B (m) is prepared in advance, and B and T _B (m) can be determined according to the rules of this table.

FIG. 6 shows the flow of received signals in the second embodiment when K = 2, B = 2, T _g = 4, T _d0 = 8, T _B0 = 1, and N _f = 16. FIG. FIG. 6 shows a state in which chromatic dispersion of a signal after reception timing T ₀ is compensated. The received signal _{s (T 0) ~s (T} 0 +15) a first _{block, s (T 0 +16) ~s} (T 0 +31) a second _{block, s (T 0 +32) ~s} (T 0 +47 ) As a third block, and is sequentially processed for each _Nf signal. In the illustrated reception signal series, the reception signal is inclined obliquely in order to simply show that the arrival time varies depending on the frequency component due to the chromatic dispersion.

First, paying attention to correction of the received signal of the first block, step S0 separates the band into B pieces, here two, for the received signal series, and the analog / digital conversion circuits 201-1 and 201- 2 and output. In steps S1-0-1 and S2-0-1, the first analog-digital conversion circuit (for example, the analog-digital conversion circuit 201-1 in FIG. 5) has a frequency band other than the first frequency band portion. The signals s ₁ (T ₀ -4) to s ₁ (T ₀ +11) whose components are attenuated are sent to the serial-parallel conversion circuit 202-1-1, and s ₁ (T ₀ +4) to s ₁ (T ₀ +19). Are respectively output to the serial / parallel conversion circuit 202-2-1.

In steps S1-0-2 and S2-0-2, the analog-digital conversion circuit 201-2 performs serial / parallel conversion of s ₂ (T ₀ -3) to s ₂ (T ₀ +12) whose bandwidth is narrowed. to the conversion circuit _202-1-2, to _{s 2 (T 0 -3) ~s} 2 (T 0 +12) serial-parallel conversion circuit 202-2-2, and outputs respectively. In step S1-1, the Fourier transform circuit 203-1 synthesizes the signals input from the serial / parallel transform circuits 202-1-1 and 202-1-2, and the synthesized signal s _s (T ₀ -4). ˜s _s (T ₀ +11) are converted into signals (s ₁ (T ₀ −4) + s ₂ (T ₀ −3)) / √2 to (s ₁ (T ₀ +11) + s ₂ (T ₀ +12)) / Get as √2.

In step S1-2, Fourier transform is performed on s _s (T ₀ -4) to _{s s} (T ₀ +11), and frequency domain signals s _f (T ₀ , -8Fw) to s _f (T ₀ , 7F _w ) are obtained. , Get. Here, the center frequency is represented as F ₀ = 0. In step S1-3, the phase rotation correction calculated from Equation (1) and Equation (5) is multiplied as shown in Equation (6), and the corrected frequency domain signal s _f ′ (T ₀ , −8F _w ) ˜ s _f ′ (T ₀ , 7F _w ) is obtained. In step S1-4, inverse Fourier transform is performed to calculate s ′ (T ₀ −4) to s ′ (T ₀ +11). In step S1-5, the guard band T _g = 4 is removed, and s ′ (T ₀ ) to s ′ (T ₀ +7) are obtained.

In step S2-1, the Fourier transform circuit 203-2 synthesizes the signals input from the serial / parallel transform circuits 202-2-1 and 202-2-2, and the synthesized signal s _s (T ₀ +4). ) To _{s s} (T ₀ +19) are converted into signals (s ₁ (T ₀ +4) + s ₂ (T ₀ +5)) / √2 to (s ₁ (T ₀ +19) + s ₂ (T ₀ +20)) / √ Get as 2. In step S2-2, Fourier transform is performed on s _s (T ₀ +4) to _{s s} (T ₀ +19), and the frequency domain signal s _f (T ₀ +8, −8F _w ) to s _f (T ₀ +8, 7F). _w ). Here, the center frequency is represented as F ₀ = 0.

In step S2-3, the phase rotation correction calculated from Equation (1) and Equation (5) is multiplied as shown in Equation (6), respectively, and the corrected frequency domain signal s _f ′ (T ₀ +8, −8F _{w ) ~s f '(T 0 +} 8,7F w), s f' (T 0 + 8, get a _{-8F w) ~s f '(T} 0 + 8,7F w).

In step S2-4, inverse Fourier transform is performed to calculate s ′ (T ₀ +4) to s ′ (T ₀ +19). In step S2-5, the guard band T _g = 4 is removed, and s ′ (T ₀ +8) to s ′ (T ₀ +15) are obtained. Therefore, the results of step S1-5 and step S2-5 are converted into serial signals by the output circuit 207, thereby obtaining s ′ (T ₀ ) to s ′ (T ₀ +15).

Further, by continuously processing the received signal of the second block as T ₀ = T ₀ +16, it is possible to continuously correct all received signal sequences.

  Also in the second embodiment, as in the second embodiment of the first embodiment, 0 signals are added in the Fourier transform circuits 203-1 to 203-K, and the signal clipping circuits 206-1 to 206- are added. The same effect can be obtained by adding and outputting signals corresponding to the same time in the output circuit 207 without using K.

C. Third Embodiment Next, a third embodiment of the present invention will be described.
FIG. 7 is a block diagram showing a configuration of a transmission unit according to the third embodiment of the present invention. In FIG. 7, 300 is a band dividing circuit, 301-1 to 301-B are analog / digital conversion circuits, 302-1-1 to 302-KB are serial / parallel conversion circuits, and 303-1-1 to 303-. K-B is a Fourier transform circuit, 304-1 to 304-K are coefficient multiplication circuits, 305-1 to 305-K are inverse Fourier transform circuits, 306-1 to 306-K are signal extraction circuits, 307 is an output circuit, Reference numeral 309 denotes a distance information output circuit.

When the optical signal is received, the band dividing circuit 300 divides the received signal into B pieces, and each corresponds to the 1st to Bth frequency bands using any of a high-pass filter, a band-pass filter, and a low-pass filter. A reception signal is formed and output to the analog / digital conversion circuits 301-1 to 301 -B. At this time, a filter may be used for the optical signal, or the filter may be used after the optical signal is converted into an electric signal. The analog / digital conversion circuits 301-1 to 301-B determine delays T _B (1) to T _B (B) to be added to the digital signal according to the distance input from the distance information output circuit 309.

  The analog / digital conversion circuits 301-1 to 301-B convert an input analog signal into a digital signal, shift the signal by a delay amount, branch the digital signal into K pieces, and a serial / parallel conversion circuit 302-1. -1 to 302-KB. The i-th serial / parallel conversion circuits 302-i-1 to 302 -i-B in each frequency band further delay the input signal by T (i), and convert serial signals of different timings into parallel signals. To do.

The delay _T B a (m) _{T B0} × to be added by an analog-digital converter 301-1~301-B (m-1) , a serial-parallel converter circuit 302-1-1~302-K- Assuming that the delay T _d (i) added at B is T _d0 × (i−1), the i-th serial-to-parallel conversion circuit 302-im in the m-th frequency band is s (T ₀ + ( _{m-1) T B0 + (} i-1) T d0 -T g) ~s (T 0 + (m-1) T B0 + (i-1) T d0 + N f -T g -1) of _{N f} Are output as parallel signals to the Fourier transform circuit 303-i. Here, N _f is considered points in the Fourier transform, T _g is the guard band length for use in extracting circuit. T ₀ can take an arbitrary number. In this description, T ₀ is a reception timing that is a head position of a signal output from the output circuit 307 in the subsequent stage.

The Fourier transform circuit 303-im in the m-th frequency band performs a Fourier transform on the B parallel signals input from the serial / parallel transform circuit 302-im, and the frequency domain signal s _{f, m} ( T ₀ + (i−1) T _d0 , F ₀ − (N _f / 2-1) F _w ),..., S _{f, m} (T ₀ + (i−1) T _d0 , F ₀ ) _,. s _{f, m} (T ₀ + (i−1) T _d0 , F ₀ + (N _f / 2) F _w ) is obtained. Here, s _{f, m} (t, f) is a reception signal corresponding to the head position reception timing t and frequency f excluding the guard band in the m-th frequency band. F ₀ is the center frequency and F _w is the frequency bandwidth of the frequency channel. Among the signals in the frequency domain, the signal s _{f, m} (T ₀ + (i−1) T _d0 , F _{m, s} ),..., S _{f, m} (T ₀ + (i) corresponding to the mth frequency band. -1) _Td0 , _Fm , _f ) is output to the coefficient multiplier circuit 304-i. Here, F _{m, s} , F _{m, f} are the center frequency of the lowest frequency channel and the center frequency of the highest frequency channel of the m-th frequency band, respectively.

The coefficient multiplying circuit 304-i receives s _{f, 1} (T ₀ + (i−1) T _d0 , F _{1, s} ),... Input from the Fourier transform circuits 303-i-1 to 303-i-B. _{s f, 1 (T 0 +} (i-1) T d0, F 1, f), ..., s f, B (T 0 + (i-1) T d0, F B, s), ..., s f _{, B} (T ₀ + (i−1) T _d0 , F _{B, f} ) is multiplied by a correction coefficient.

  The distance information output circuit 309 estimates the chromatic dispersion amount using the transmission distance information of the optical signal and the information on the type of the optical fiber. The shift in arrival time with respect to the wavelength of light, that is, the amount of phase rotation can be calculated from the transmission distance and the type of optical fiber, and the phase rotation in the carrier of frequency f is given by equation (1). Can do. Since the Fourier transform position differs for each frequency band, it is necessary to further multiply the Fourier transform position correction coefficient. The correction coefficient h (f, m) is obtained by the mathematical formula (5).

  The coefficient multiplication circuit 304-i corrects the complex conjugate of the phase rotation g (f) input from the distance information output circuit 309 and the delay coefficient given by the analog / digital conversion circuit 301-i h (f, m). Is multiplied by the frequency domain signal as a correction coefficient to correct the arrival time difference due to the light frequency. The obtained frequency domain signal is expressed by the following equation (7).

The superscript * represents a complex conjugate number. S _f ′ (T ₀ + (i−1) T _d0 , F ₀ − (N _f / 2-1) F _w ),..., S _f ′ (T ₀ + (i− 1) T _d0 , F ₀ ),..., S _f ′ (T ₀ + (i−1) T _d0 , F ₀ + (N _f / 2) F _w ) are output to the inverse Fourier transform circuit 305-i.

Inverse Fourier transform circuit 305-i, an inputted signal _{s f '(T 0 + (} i-1) T d0, F 0 - (N f / 2-1) F w), ..., s f' (T _{0 + (i-1) T} d0, F 0), ..., s f '(T 0 + (i-1) T d0, F 0 + a _{(N f / 2) F w} ), again, the time domain The signal is converted, s ′ (T ₀ + (i−1) T _d0 −T _g ) to s ′ (T ₀ + (i−1) T _d0 + N _f −T _g −1) is calculated, and signal extraction is performed. Output to circuit 306-i.

Signal extracting circuit 306-i deletes the data corresponding to the front and rear a _{T g} from the input signal (guard band _{portion), s'(T 0 + (} i-1) T d0) ~s' (T 0 A signal of N _f −2T _g of + (i−1) T _d0 + N _f −2T _g −1) is output to the output circuit 307.

The output circuit 307 combines the signals input from the signal cutout circuits 306-1 to 306-K and converts them into a serial signal. When N _f −2T _g = T _d0 is satisfied, the signal extraction circuits 306-1, 306-2,..., 306-K have s ′ (T ₀ ) to s ′ (T ₀ + T _d0 −1). _{, s'(T 0 + T d0} ) ~s' (T 0 + 2T d0 -1), ..., s'(T 0 + (K-1) T d0) ~s' (T 0 + KT d0 -1) is input Then, by rearranging these signals into serial signals, s ′ (T ₀ ) to s ′ (T ₀ + KT _d0 −1) can be obtained. Since the obtained signal has been corrected for arrival delay due to chromatic dispersion, it is output after the influence of chromatic dispersion is removed. Further, at a time for processing the signals of _{KT d0,} by decoding the signals from T 0 ₊ KT _d0, it is possible to process the received signal continuously.

By using the third embodiment described above, it is possible to reduce the temporal spread of the signal due to chromatic dispersion. Therefore, distance information output circuit 309, depending on the transmission distance can specify the size of the delay _T B (m) given by the analog-to-digital converter 301-1 through 301-B. For example, according to the transmission distance, a table that gives the corresponding delay T _B (m) is prepared in advance, and T _B (m) can be determined according to the rules of this table.

Also, the number B of frequency bands to be divided by the band dividing circuit 300 can be designated by the distance information output circuit 309. Therefore, distance information output circuit 309, depending on the transmission distance can specify the size of the delay _T B (m) given by the analog-to-digital converter 301-1 through 301-B. For example, according to the transmission distance, a table giving the number of band divisions B is prepared in advance, and B can be determined according to the rules of this table. Alternatively, a table that gives the number of band divisions B and the delay T _B (m) is prepared in advance, and B and T _B (m) can be determined according to the rules of this table.

FIG. 8 shows the flow of the received signal in the third embodiment when K = 2, B = 2, T _g = 4, T _d0 = 8, T _B0 = 1, and N _f = 16. FIG. FIG. 8 shows a state in which the chromatic dispersion of the signal after the reception timing T ₀ is compensated. The received signal _{s (T 0) ~s (T} 0 +15) a first _{block, s (T 0 +16) ~s} (T 0 +31) a second _{block, s (T 0 +32) ~s} (T 0 +47 ) As a third block, and is sequentially processed for each _Nf signal. In the illustrated reception signal series, the reception signal is inclined obliquely in order to simply show that the arrival time varies depending on the frequency component due to the chromatic dispersion.

First, paying attention to the correction of the received signal of the first block, the step S0 separates the band into B pieces, here two, for the received signal series, and the analog / digital conversion circuits 301-1 and 301- 2 and output. In steps S1-0-1 and S2-0-1, the first analog-digital conversion circuit (for example, the analog-digital conversion circuit 301-1 in FIG. 7) is set to a frequency band other than the first frequency band portion. The signals s ₁ (T ₀ -4) to s ₁ (T ₀ +11) whose components are attenuated are sent to the serial-parallel conversion circuit 302-1-1, and s ₁ (T ₀ +4) to s ₁ (T ₀ +19). Are output to the serial / parallel conversion circuit 302-2-1.

In steps S1-0-2 and S2-0-2, the analog-digital conversion circuit 301-2 performs serial / parallel conversion of s ₂ (T ₀ -3) to s ₂ (T ₀ +12) whose bandwidth is narrowed. to the conversion circuit _{302-1-2, s 2 (T 0 +5} ) ~s 2 a (T 0 +20) to the serial-parallel converter circuit 302-2-2, and outputs respectively. In steps S1-1-1 and S1-1-2, the Fourier transform circuits 303-1-1 and 303-1-2 are connected to the serial / parallel transform circuits 302-1-1 and 302-1-2. the input signal, respectively, performs Fourier _{transform, s f, 1 '(T} 0, -8F w) ~s f, 1' and _{(T 0, 7F w),} s f, 2 '(T 0, -8F _{w) ~s} f, 2 _'get the _(T 0, 7F _w).

In step S1-2, signals corresponding to the respective frequency bands are extracted, synthesized, and frequency domain signals s _{f, 1} ′ (T ₀ , −8F _w ),..., S _{f, 1} ′ (T ₀ , − F _w ), s _{f, 2} ′ (T ₀ , 0) to s _{f, 2} ′ (T ₀ , 7F _w ) are obtained. In step S1-3, the phase rotation correction calculated from Equation (1) and Equation (5) is multiplied as shown in Equation (7), and the corrected frequency domain signal s _f ′ (T ₀ , −8F _w ) is obtained. ˜s _f ′ (T ₀ , 7F _w ). In step S1-4, inverse Fourier transform is performed to calculate s ′ (T ₀ −4) to s ′ (T ₀ +11). In step S1-5, the guard band T _g = 4 is removed, and s ′ (T ₀ ) to s ′ (T ₀ +7) are obtained.

In steps S2-1-1 and S2-1-2, the Fourier transform circuits 303-2-1 and 303-2-2 are input from the serial / parallel transform circuits 302-2-1 and 302-2-2. the signals, respectively performs a Fourier _{transform, s f, 1 '(T} 0 + 8, -8F w) ~s f, 1' and _{(T 0 + 8,7F w),} s f, 2 '(T 0 +8, -8F _{w) ~s f, 2} 'get the (T 0 + 8,7F _w). In step S1-2, it extracts a signal corresponding to the respective frequency bands, combining the signals in the frequency domain _{s f, 1 '(T 0} + 8, -8F w), ..., s f, 1' (T 0 +8 , −F _w ), s _{f, 2} ′ (T ₀ +8,0) to s _{f, 2} ′ (T ₀ + 8,7F _w ).

In step S2-3, the phase rotation correction calculated from Equation (1) and Equation (5) is multiplied as shown in Equation (7), and the corrected frequency domain signal s _f ′ (T ₀ +8, −8F _w ) To s _f ′ (T ₀ + 8,7F _w ). In step S2-4, inverse Fourier transform is performed to calculate s ′ (T ₀ +4) to s ′ (T ₀ +19). In step S2-5, the guard band T _g = 4 is removed, and s ′ (T ₀ +8) to s ′ (T ₀ +15) are obtained.

Further, by continuously processing the received signal of the second block as T ₀ = T ₀ +16, it is possible to continuously correct all received signal sequences.

In the third embodiment, the frequency domain is divided in the band dividing circuit 300 and then down-converted, so that the circuit clock can be kept low and the calculation load can be reduced. In this case, the number of points in the Fourier transform can be reduced. Assuming B = 4, K = 2, T _g = 4, T _d0 = 8, T _B0 = 1, N _f = 16 are the same, and the number of points of Fourier transform used in each frequency band is considered as N _f0 8. The received signal _{s (T 0) ~s (T} 0 +15) a first _{block, s (T 0 +16) ~s} (T 0 +31) a second _{block, s (T 0 +32) ~s} (T 0 +47 ) As the third block, and is sequentially processed for each _Nf signal. Paying attention to the correction of the received signal of the first block, step S0 separates the band into four for the received signal sequence, down-converts at the center frequency of each band, and analog-digital conversion circuit 301-1. To 301-4.

In steps S1-0-1 and S2-0-1, the analog-digital conversion circuit 301-1 attenuates the frequency band components other than the frequency band 1 portion, and the signals s ₁ (T ₀ -4) to s ₁ ( T corresponds to ₀ +11), downconverted _{[s L1 (T 0 -4)} , s L1 (T 0 -2), s L1 (T 0), s L1 (T 0 +2), s L1 (T ₀ +4), s _L1 (T ₀ +6), s _L1 (T ₀ +8), s _L1 (T ₀ +10)] to the serial / parallel conversion circuit 302-1-1 and s ₁ (T ₀ +4) to s. ₁ corresponds to _(T 0 +19), downconverted _{[s L1 (T 0 +4)} , s L1 (T 0 +6), s L1 (T 0 +8), s L1 (T 0 +10), s L1 ( _{T 0 +12), s L1 (} T 0 +14), s L1 T _{0 +16),} respectively output s _L1 a (T 0 +18)] to the serial-parallel converter circuit 302-2-1.

In steps S1-0-i and S2-0-i (2 ≦ i ≦ 4), the analog-digital conversion circuit 301-i has a band narrowed from s _i (T ₀ −4+ (i−1)) to _{s i (T 0 +11+ (i} -1)) corresponding to the _{[s Li (T 0 -4+ (} i-1)), s Li (T 0 -2+ (i-1)), s Li (T 0 + _{(i-1)), s} Li (T 0 +2+ (i-1)), s Li (T 0 +4+ (i-1)), s Li (T 0 +6+ (i-1)), s Li (T _{0 +8+ (i-1))} , s Li (T 0 +10+ (i-1))] to the serial-parallel conversion circuit _{302-1-i, s i (T} 0 +4+ (i-1)) ~s i _{(T 0 +19+ (i-1} )) corresponding to the _{[s Li (T 0 +4+ (} i-1)), s Li (T 0 +6+ (i-1)), s L _{(T 0 +8+ (i-1} )), s Li (T 0 +10+ (i-1)), s Li (T 0 +12+ (i-1)), s Li (T 0 +14+ (i-1)), s _Li (T ₀ +16+ (i−1)), s _Li (T ₀ +18+ (i−1))] are output to the serial / parallel conversion circuit 302-2-i, respectively.

In step S1-1-i (1 ≦ i ≦ 4), the Fourier transform circuit 303-1-i performs a Fourier transform of N _f0 = 8 on the signal input from the serial / parallel transform circuit 302-1-i, s _{f, i} ′ (T ₀ , −10 F _w + (i−1) × 4 F _w ) to s _{f, i} ′ (T ₀ , −2 F _w + (i−1) × 4 F _w ) are obtained. Here, the frequency is considered as the frequency before down-conversion. In step S1-2, signals corresponding to the respective frequency bands are extracted and synthesized. In this case, i-th Fourier transform circuit from _{s f, i '(T 0} , -8F w + (i-1) × 4F w) ~s f, i' (T 0, -8F w + i × 4F w) to obtain a signal _{s f,} 1 in the frequency domain _{'(T 0, -8F w)} , ···, s f, 1' (T 0, -F w), s f, 2 '(T 0, 0) to s _{f, 2} ′ (T ₀ , 7F _w ) can be obtained.

In step S1-3, respectively Equation 1, the phase rotation correction calculated from Equation 5, multiplied by Equation 7, the correction frequency domain signal _{s f '(T 0, -8F} w) ~s f' (T 0 , 7F _w ). In step S1-4, inverse Fourier transform is performed to calculate s ′ (T ₀ −4) to s ′ (T ₀ +11). In step S1-5, the guard band T _g = 4 is removed, and s ′ (T ₀ ) to s ′ (T ₀ +7) are obtained.

In steps S2-1-1 and S2-1-2, the Fourier transform circuits 303-2-1 and 303-2-2 are input from the serial / parallel transform circuits 302-2-1 and 302-2-2. performs each Fourier transform _{signal, s f, 1 '(T} 0 + 8, -8F w) ~s f, 1' (T 0 + 8,7F w) and _{s f, 2 '(T 0} + 8, -8F w ) To s _{f, 2} ′ (T ₀ + 8,7F _w ). In step S1-2, signals corresponding to the respective frequency bands are extracted, synthesized, and frequency domain signals s _{f, 1} ′ (T ₀ + 8, −8F _w ),..., S _{f, 1} ′ (T _{0 + 8, -F w),} s f, 2 '(T 0 +8,0) ~s f, 2' to obtain a (T 0 + 8,7F _w).

In step S2-3, the phase rotation correction calculated from Equation 1 and Equation 5 is multiplied as shown in Equation 7, respectively, and corrected frequency domain signals s _f ′ (T ₀ +8, −8F _w ) to s _f ′ (T ₀ + 8,7F _w) obtained. In step S2-4, inverse Fourier transform is performed to calculate s ′ (T ₀ +4) to s ′ (T ₀ +19). In step S2-5, the guard band T _g = 4 is removed, and s ′ (T ₀ +8) to s ′ (T ₀ +15) are obtained.

Further, by continuously processing the received signal of the second block as T ₀ = T ₀ +16, it is possible to continuously correct all received signal sequences.

  Next, a second embodiment of the third embodiment of the present invention will be described with reference to FIG. Similar to the second embodiment of the first embodiment, 0 signals are added in the Fourier transform circuits 303-1 to 303-K, and the signal cutout circuits 306-1 to 306-K are not used, and the output circuit 307 is used. A similar effect can be obtained by adding and outputting signals corresponding to the same time.

FIG. 9 shows a case where K = 2, B = 2, T _d0 = 8, T _B0 = 1, N _f = 16, and T _d1 = T _d2 = 4 in the second mode of the third embodiment. FIG. 3 is a sequence diagram showing the flow of a received signal in FIG. FIG. 9 shows a state in which chromatic dispersion of a signal after reception timing T ₀ is compensated. Received _{signal, s (T 0) ~s (} T 0 +15) a first _{block, s (T 0 +16) ~s} (T 0 +31) a second _{block, s (T 0 +32) ~s} (T 0 the +47) as a third block, and it shall be sequentially processed for each _{N f} signal. In the illustrated reception signal series, the reception signal is inclined obliquely in order to simply show that the arrival time varies depending on the frequency component due to chromatic dispersion.

First, paying attention to the correction of the received signal of the first block, the step S0 separates the band into B pieces, here two, for the received signal series, and the analog / digital conversion circuits 301-1 and 301- 2 and output. In step S1-0-1 and the step S2-0-1, analog-to-digital converter 301-1 is the signal _s 1 attenuated frequency band components other than the frequency band 1 part _{(T 0) ~s 1 (T} 0 0 signal is added to +7) and output to the serial / parallel conversion circuit 302-2-1.

In steps S1-0-2 and S2-0-2, the analog-digital conversion circuit 301-2 adds a 0 signal to s ₂ (T ₀ +1) to s ₂ (T ₀ +8) whose bandwidth has been narrowed. , to the serial-parallel converter circuit _{302-1-2, s 2 (T 0 +9} ) ~s 2 (T 0 +16) in with a zero signal, to the serial-parallel converter circuit 302-2-2 outputs respectively . In steps S1-1-1 and S1-1-2, the Fourier transform circuits 303-1-1 and 303-1-2 are input from the serial / parallel transform circuits 302-1-1 and 302-1-2. each performs Fourier transform on the _{signal, s f, 1 '(T} 0, -8F w) ~s f, 1' (T 0, 7F w) and _{s f, 2 '(T 0} , -8F w) ~ s _{f, 2} ′ (T ₀ , 7F _w ) is obtained.

In step S1-2, signals corresponding to the respective frequency bands are extracted, synthesized, and frequency domain signals s _{f, 1} ′ (T ₀ , −8F _w ),..., S _{f, 1} ′ (T _{0 , -F w), s f,} 2 '(T 0, 0) ~s f, 2' get the _(T 0, 7F _w). In step S1-3, respectively Equation 1, the phase rotation correction calculated from Equation 5, multiplied by Equation 7, the correction frequency domain signal _{s f '(T 0, -8F} w) ~s f' (T 0 , 7F _w ). In step S1-4, inverse Fourier transform is performed to calculate s ′ (T ₀ −4) to s ′ (T ₀ +11).

In steps S2-1-1 and S2-1-2, the Fourier transform circuits 303-2-1 and 303-2-2 are input from the serial / parallel transform circuits 302-2-1 and 302-2-2. performs each Fourier transform _{signal, s f, 1 '(T} 0 + 8, -8F w) ~s f, 1' (T 0 + 8,7F w) and _{s f, 2 '(T 0} + 8, -8F w ) To s _{f, 2} ′ (T ₀ + 8,7F _w ). In step S1-2, signals corresponding to the respective frequency bands are extracted, synthesized, and signals in the frequency domain s _{f, 1} ′ (T ₀ + 8, −8F _w ),..., S _{f, 1} ′ (T _{0 + 8, -F w),} s f, 2 '(T 0 +8,0) ~s f, 2' to obtain a (T 0 + 8,7F _w).

In step S2-3, the phase rotation correction calculated from Equation 1 and Equation 5 is multiplied as shown in Equation 7, respectively, and corrected frequency domain signals s _f ′ (T ₀ +8, −8F _w ) to s _f ′ (T ₀ + 8,7F _w) obtained. In step S2-4, inverse Fourier transform is performed to calculate s ′ (T ₀ +4) to s ′ (T ₀ +19). In step S1-5, signals corresponding to the same time among the signals input from steps S1-4 and S2-4 are added, and s ′ (T ₀ −4) to s ′ (T ₀ +19) are output. .

Further, by continuously processing the received signal of the second block as T ₀ = T ₀ +16, it is possible to continuously correct all received signal sequences. Also, among the _s'calculated in step _{S1-5 (T 0 -4) ~s' (} T 0 +19), s' and _{(T 0 -4) ~s' (T} 0 -1) s'(T ₀ + 16) to s ′ (T ₀ +19) overlap the output results of step S2-4 in the 0th block and step S1-4 in the second block, respectively. For this reason, the effect of chromatic dispersion can be compensated by adding the output results in the preceding and succeeding blocks as shown in FIG.

Further, after the optical signal is branched by the band dividing circuit 300 and the frequency band is divided, the chromatic dispersion can be compensated by using the first embodiment for the signal of each frequency band and adding the output. At this time, the delay time T _B (m) specific to the frequency band is used in the output of the band dividing circuit 300 or in the analog / digital circuits 301-1 to 301-B.

  5 and 7, the band dividing circuits 200 and 300 and the analog / digital conversion circuits 201-1 to 201-B and 301-1 to 301-B can be used interchangeably. In this case, the optical signal is converted into a digital signal by the analog / digital conversion circuits 201-1 to 201-B and 301-1 to 301-B, and after branching, the digital signal is divided by the band dividing circuits 200 and 300, Other than the corresponding frequency band is attenuated and output to the serial / parallel conversion circuits 202-1-1 to 202 -KB, 302-1-1 to 302 -KB, and Fourier transform circuits 203-1 to 203- K, 303-1-1 to 303-KB or lower processing can also be performed.

  Further, in the formulas (2), (6), and (7), the coefficient α (f) that compensates for the amplitude and phase generated by propagating through the circuit or filter can be multiplied. When an example in the case of Expression (7) is shown, a corrected signal can be obtained as the following Expression (8). In this way, frequency dependent phase rotation and amplitude reduction by the device can be compensated.

The m th delay T _B (m) given by the band dividing circuit 300 outputs the maximum arrival time difference T _max due to chromatic dispersion to the distance information output circuit 309 so that T _B (m) = T _max / B. Can do.

  In addition, after the band is cut out in the band dividing circuit 300 of FIG. 7, down-conversion is performed at a predetermined frequency in the cut-out frequency band, and the result is output to the analog / digital conversion circuits 301-1 to 301-B. The device clock can be lowered to reduce the computation load. In the case where the coefficient multiplication circuits 304-1 to 304-K do not use down-conversion by multiplying the coefficients based on the frequencies before down-conversion and input them to the inverse Fourier transform circuit in the order of the frequencies before down-conversion. The same effect can be obtained.

  Further, when down-conversion is used, a phase shift may occur due to down-conversion in each of the Fourier transform circuits 303-i-1 to 303-i-B that outputs a signal to the i-th coefficient multiplication circuit 304-i. The phase correction can also be performed on the input signal. As the reference signal, a signal used for down-conversion or a frequency domain signal overlapping in each Fourier transform circuit can be used.

  Further, as a frequency used for down-conversion, a center frequency in each frequency band can be used, or a frequency that is not a signal region in each frequency band can be used in order to prevent a DC component from being mixed.

  Further, as the transmission distance L used in the distance information output circuit, when predistortion is performed on the dispersion compensating fiber or the transmission side, the path length corresponding to the uncompensated path length can be substituted instead of the actual path length.

  According to the above-described embodiment, it is possible to reduce interference power caused by chromatic dispersion or frequency-dependent circuit characteristics by performing Fourier transform repeatedly and multiplying a frequency component by a coefficient.

It is a block diagram which shows the structure of the transmission part by the 1st Embodiment of this invention. It is a sequence diagram showing the flow of the received signal in the 1st Embodiment of this invention. In the 1st Embodiment of this invention, it is a conceptual diagram which shows the duplication area | region of a signal when _{Tg is set} to 2. FIG. It is a sequence diagram showing the flow of the received signal in the 1st Embodiment of this invention. It is a block diagram which shows the structure of the transmission part by the 2nd Embodiment of this invention. It is a sequence diagram showing the flow of the received signal in the 2nd Embodiment of this invention. It is a block diagram which shows the structure of the transmission part by the 3rd Embodiment of this invention. It is a sequence diagram showing the flow of the received signal in the 3rd Embodiment of this invention. It is a sequence diagram showing the flow of the received signal in the 3rd Embodiment of this invention.

Explanation of symbols

  101, 201-1 to 201-B, 301-1 to 301-B... Analog / digital conversion circuit, 102-1 to 102-K, 202-1-1 to 202-KB, 302-1- 1 to 302-KB ... serial-parallel conversion circuit, 103-1 to 103-K, 203-1 to 203-K, 303-1-1 to 303-KB ... Fourier transform circuit, 104-1 to 104 -K, 204-1 to 204 -K, 304-1 to 304 -K... Coefficient multiplication circuit, 105-1 to 105 -K, 205-1 to 205 -K, 305-1. 305-K: Inverse Fourier transform circuit, 106-1 to 106-K, 206-1 to 206-K, 306-1 to 306-K ... Signal extraction circuit, 107,207,307 ... Output Circuit, 109, 209, 309 ... Distance information output Circuit, 200, 300 ... band division circuit

Claims (7)

An interference reduction method in optical communication using an optical fiber,
An analog-digital conversion step for converting an optical signal into a digital signal;
A serial-parallel conversion step of branching the digital signal into two or more and converting each of the branched digital signals into a parallel signal after giving a delay amount proportional to the order of branching;
A Fourier transform step of transforming the parallel signal into a frequency domain signal by performing a Fourier transform on the parallel signal;
A distance information output step of estimating or pre-stored and output the amount of chromatic dispersion generated by the optical fiber based on the distance propagated in the optical fiber and the characteristics of the optical fiber with respect to the signal light wavelength;
A coefficient multiplication step of multiplying the Fourier-transformed frequency domain signal by a first correction coefficient for correcting the arrival time difference due to the frequency based on the chromatic dispersion amount;
An inverse Fourier transform step of performing an inverse Fourier transform on the frequency domain signal multiplied by the first correction coefficient to convert the signal into a time domain signal;
A signal cut-out step for selecting a time-domain signal excluding a guard band portion for the inverse Fourier-transformed time-domain signal;
An output step of rearranging and outputting the selected plurality of time domain signals in the order of branching. An interference reduction method in optical communication using an optical fiber,
A band division step for separating the optical signal into signals of different frequency bands;
An analog-digital conversion step of converting the band-divided signals into digital signals, respectively, and adding a delay amount proportional to the band-divided order;
Serial / parallel which branches the digital signal to which the delay amount is added into two or more, and converts each branched digital signal into a parallel signal after giving a delay amount proportional to the branching order. A conversion step;
A Fourier transform step of synthesizing parallel signals from the different frequency bands and performing a Fourier transform on a frequency domain signal obtained by the synthesis, thereby converting the signal into a frequency domain signal;
A distance information output step of estimating or pre-stored and output the amount of chromatic dispersion generated by the optical fiber based on the distance propagated in the optical fiber and the characteristics of the optical fiber with respect to the signal light wavelength;
Coefficient multiplication for multiplying the frequency domain signal subjected to Fourier transform by a first correction coefficient for correcting the arrival time difference due to frequency based on the chromatic dispersion amount and the delay amount used in each of the branched frequency bands. Steps,
An inverse Fourier transform step of performing an inverse Fourier transform on the frequency domain signal multiplied by the first correction coefficient to convert the signal into a time domain signal;
A signal cut-out step for selecting a time-domain signal excluding a guard band portion for the inverse Fourier-transformed time-domain signal;
An output step of rearranging and outputting the selected plurality of time domain signals in the order of branching. An interference reduction method in optical communication using an optical fiber,
A band division step for separating the optical signal into signals of different frequency bands;
An analog-digital conversion step of converting the band-divided signals into digital signals, respectively, and adding a delay amount proportional to the band-divided order;
A serial / parallel conversion step of further branching the digital signal to which the delay amount is added into two or more, giving a specific delay amount to each of the branched digital signals, and converting the digital signal to a parallel signal;
A Fourier transform step of transforming each parallel signal into a frequency domain signal by performing a Fourier transform;
A distance information output step of estimating or pre-stored and output the amount of chromatic dispersion generated by the optical fiber based on the distance propagated in the optical fiber and the characteristics of the optical fiber with respect to the signal light wavelength;
A coefficient for multiplying the frequency domain signal subjected to Fourier transform by a first correction coefficient for correcting the arrival time difference due to frequency based on the chromatic dispersion amount and the delay amount used in each of the branched frequency bands. Multiplication step;
An inverse Fourier transform step of performing an inverse Fourier transform on the frequency domain signal multiplied by the first correction coefficient to convert the signal into a time domain signal;
A signal cut-out step for selecting a time-domain signal excluding a guard band portion for the inverse Fourier-transformed time-domain signal;
An output step of rearranging and outputting the selected plurality of time domain signals in the order of branching. The coefficient multiplication step includes:
In addition to the first correction coefficient that corrects the arrival time difference due to the estimated amount of chromatic dispersion, the Fourier-transformed frequency domain signal is multiplied by a second correction coefficient that corrects the frequency pass characteristic of the circuit. The interference reduction method according to any one of claims 1 to 3. Prior to performing the band splitting step, the analog to digital conversion step converts an optical signal into a digital signal,
4. The band division step according to claim 2, wherein the digital signal converted in the analog-digital conversion step is separated into signals of two or more different frequency bands by a digital filter. The interference reduction method according to the item. The analog-digital conversion step converts the band-divided signal into a low-frequency signal by performing frequency conversion according to a center frequency, converts the low-frequency signal into a digital signal,
In the inverse Fourier transform step, the signal in the frequency domain multiplied by the first correction coefficient is rearranged in the order of the original frequency according to the center frequency, and then the inverse Fourier transform is performed. 4. The interference reduction method according to claim 3, wherein the interference reduction method is converted into a time domain signal. An interference reduction device for optical communication using an optical fiber,
A band dividing circuit for separating an optical signal into signals of different frequency bands;
An analog-digital conversion circuit that converts the band-divided signals into digital signals, and adds a delay amount proportional to the band-divided order;
A serial-to-parallel conversion circuit for further branching the digital signal to which the delay amount is added into two or more, giving a specific delay amount to each of the branched digital signals, and converting to a parallel signal;
A Fourier transform circuit for transforming the parallel signal into a frequency domain signal by performing a Fourier transform on the parallel signal;
A distance information output circuit that estimates the amount of chromatic dispersion caused by the optical fiber based on the distance propagated in the optical fiber and the characteristics of the optical fiber with respect to the signal light wavelength;
A coefficient multiplication circuit that multiplies the Fourier-transformed frequency domain signal by a first correction coefficient that corrects the arrival time difference due to the frequency based on the estimated chromatic dispersion amount;
An inverse Fourier transform circuit for transforming a signal in the time domain by performing an inverse Fourier transform on the signal in the frequency domain multiplied by the first correction coefficient;
A signal cut-out circuit that selects a time-domain signal excluding a guard band part with respect to the inverse Fourier-transformed time-domain signal;
An output circuit for rearranging and outputting the selected plurality of time domain signals in an appropriate order; and
An interference reduction apparatus comprising:

Images