JP6985597B2 - Optical receiver and frequency offset estimation method - Google Patents

Description

The present invention relates to an optical receiver instrumentation置及Beauty frequency offset estimation method.

Currently, the PON (Passive Optical Network) method standardized by the IEEE (The Institute of Electrical and Electronics Engineers, Inc) and ITU-T ((International Telecommunication Union Telecommunication Standardization Sector)) is widely used in optical subscriber systems. There is.

As shown in FIG. 11, in the PON optical transmission system 100, one station-side device, the OLT (Optical Line Terminal) 120, and a plurality of subscriber devices, the ONU (Optical Network Unit) 110-1 to 1, Communication is performed with 110-3. In the downlink communication of the PON system, the optical signal transmitted from the OLT 120 is branched into a plurality of paths by the optical splitter 130, and each ONU 110-1 to 110-3 receives the optical signal. At this time, ONU 110-1 to 110-3 select and receive only the signal of the time slot assigned to them.

On the other hand, in uplink communication, ONU110-1 to 110-3 transmit signals only to the time slots assigned to them, and collide with signals transmitted from other ONU110-1 to 110-3. I'm avoiding it. As described above, in the PON method, it is possible to economically provide a high-speed optical access service by sharing the same optical fiber with a plurality of subscribers and suppressing the equipment cost.

In an access network using the PON method, it is required to extend the maximum transmission distance and increase the number of branches by the optical splitter 130 in order to reduce the equipment cost. The reason why it is required to extend the maximum transmission distance is that it is possible to consolidate lines in a wide area in one station building. Further, the reason why it is required to increase the number of branches of the optical splitter 130 is that by consolidating many lines in one OLT 120, it is possible to reduce the number of OLT 120s and station buildings per line, and it is possible to reduce the number of office buildings. This is because the installation and operation costs can be reduced.

However, extending the transmission distance or increasing the number of branches of the optical splitter 130 causes an increase in optical loss, so that the signal light intensity at the time of reception decreases, the signal-to-noise ratio decreases, and a code error occurs. It will worsen the rate. In addition to this, the extension of the transmission distance causes the waveform deterioration of the signal due to the wavelength dispersion, so that the code error rate due to the wavelength deterioration also deteriorates. Therefore, in order to prolong the transmission distance and increase the number of branches of the optical splitter 130, a technique for compensating for a decrease in the signal-to-noise ratio and waveform deterioration in the receiver is important.

In order to improve the decrease in the signal-to-noise ratio, it is conceivable to install an optical amplifier in front of the receiver to amplify the signal light. However, this method has a problem that the effect of amplification is limited to about 10 dB due to spontaneous emission light noise (ASE (Amplified Spontaneous Emission) noise) generated at the time of optical amplification.

On the other hand, when a method of installing an optical amplifier in the middle of a transmission line is adopted, the influence of ASE noise can be suppressed to a relatively small value. However, this method requires a new relay facility to be installed in the middle of the transmission line, which causes a problem of increased cost.

Further, in order to compensate for the waveform deterioration, a method of inserting a compensating optical fiber in the middle of the transmission line can be considered. However, the distance between the subscriber device that transmits the signal and the station building equipment varies depending on the subscriber, and the characteristics of the required compensation fiber also change accordingly, so that there is a problem that the cost increases even in this method. be.

In order to solve these problems, the introduction of a digital coherent method for access networks is being considered. In the conventional intensity modulation-direct detection (IM-DD) method, the signal-to-noise ratio decreases due to the thermal noise of the receiver. In contrast, in the digital coherent method, coherent reception is performed. By doing so, the thermal noise generated in the receiver is reduced, and reception with a signal-to-noise ratio near the shot noise limit becomes possible. In the digital coherent method, the received signal is propagated by performing digital signal processing. It is possible to compensate for the waveform deterioration caused by this.

Andreas Leven, Noriaki Kaneda, Ut-Va Koc, Young-Kai Chen, “Frequency Optimization in Intradyne Reception”, IEEE PHOTONICS TECHNOLOGY LETTERS, VOL. 19, NO. 6, MARCH 15, 2007 Lei Li, Zhenning Tao, Shoichiro Oda, Takeshi Hoshida, Jens C. Rasmussen, “Wide-range, Accurate and Simple Digital Frequency Offset Compensator for Optical Coherent Receivers”, OFC / NFOEC 2008 Xian Zhou, Xue Chen, Keping Long, “Wide-Range Frequency Offset Optimization Algorithm for Optical Coherent Systems Using Training Sequence”, IEEE PHOTONICS TECHNOLOGY LETTERS, VOL. 24, NO. 1, JANUARY 1, 2012

The digital coherent receiving side device performs coherent reception that detects the phase component of the signal light by measuring the intensity of the interference waveform between the transmitted optical signal and the locally oscillated light (hereinafter, also referred to as local emission). conduct. In the received signal at the time of coherent reception, a frequency offset component that changes with the frequency of the difference between the frequency of the signal light and the frequency of the station emission is superimposed on the modulation component corresponding to the transmission code sequence. In order to reproduce the modulation component corresponding to the transmission code sequence on the receiving side, it is necessary to remove the frequency offset component from the received signal. Therefore, the receiving side is required to estimate the frequency offset amount.

In the access network, since an inexpensive light source is used as the light source of the transmitting device from the viewpoint of cost reduction, it is expected that the frequency variation of each transmitting device is larger than that of the core network. Therefore, it is expected that the range of the magnitude of the frequency offset amount generated on the receiving side will be widened. Therefore, the technique for estimating the frequency offset amount in the access network needs to be a technique having a wide frequency range to be estimated.

Further, in the uplink communication of the PON system, the signal received by the OLT 120 is a burst signal, and in each time slot, the OLT 120 receives the signals transmitted from different ONU 110-1 to 110-3. Therefore, it is necessary to estimate the frequency offset amount for each time slot on the receiving side, and the frequency of the estimation is higher than that in the case of the core network.

As a technique for estimating the amount of frequency offset, for example, there is a method of estimating the amount of frequency offset by analyzing a code sequence given in a certain section of a transmission signal on the receiving side. In an access network where the frequency offset amount is frequently estimated, when the method of estimating the frequency offset amount by analyzing the code sequence is applied, the processing time required for the estimation is required for efficient transmission. It needs to be reduced. Therefore, it is important to reduce the number of code sequences for estimating the amount of frequency offset that does not contribute to data transmission.

From the above, the technique for estimating the amount of frequency offset in the access network is required to have a wide frequency range to be estimated and a small number of code sequences used for estimation.

For example, as a technique for estimating the frequency offset amount, a technique using the M-multiplication method (see, for example, Non-Patent Document 1), a technique using a feedback loop (see, for example, Non-Patent Document 2), and the like have been proposed.

The M-th power method is a method of calculating the one-symbol delay differential of a received signal, removing the frequency offset component by multiplying it M times, and extracting the modulation component (where M is a multi-valued number and BPSK (). In the case of Binary Phase Shift Keying), M = 2). The frequency range in which the frequency offset amount can be estimated in the M-th power method is −B / 2M ≦ f _IF <B / 2M (where B is the baud rate of the symbol in the transmission signal, and f _IF is the signal light and the station. It is the frequency difference from the light emission). Therefore, when the M-th power method is used, there is a problem that the range of estimable frequencies decreases as the multi-valued number M increases.

On the other hand, in the technique using the feedback loop, the range of the estimable frequency is −B / 2 ≦ f _IF <B / 2, so that it can be estimated by the size of the multivalued number M as in the M-multiplication method. The frequency range is not reduced. However, in this technique, there is a problem that it takes a long time to converge the feedback, and therefore the code sequence becomes long.

In addition, by estimating the timing based on the correlation of general training signals and subtracting the training signal from the received signal, the range of frequencies that can be estimated by the above M-multiplication method is widened, and -B / 2≤f _IF <B / 2. (For example, see Non-Patent Document 3). Here, a technique for estimating the frequency offset amount using the training code sequence will be described with reference to the technique shown in Non-Patent Document 3.

FIG. 12 is a schematic block diagram showing a configuration of an optical transmission system 200 that performs digital coherent transmission. In the digital coherent transmission method, the optical coherent receiving unit 300 included in the optical receiving device 220 performs coherent reception for measuring the interference waveform between the received signal light and the local emission output from the local oscillator 301.

When coherent reception is performed, a polarization diversity receiver is often used, but FIG. 12 shows an example in which a phase diversity receiver is applied as the optical coherent receiver 300 in order to simplify the explanation of the configuration. There is. Assuming that the frequencies of the signal light and the local emission are f ₀ and f _LO , respectively, the I (In-Phase) component and the Q (Quadrature) component of the signal current output by the optical coherent receiver 300 are the following equations (1) and Q (Quadrature), respectively. It is expressed by the following equation (2).

In the equations (1) and (2), φ _m (t) is a phase modulation component of the signal light. α indicates an offset component of the phase difference between the local emission and the signal light, and is constant in time. f _IF is the frequency difference between the signal light and the local emission, and is obtained by the following equation (3).

From equations (1) to (3), the complex amplitude of the received signal is expressed as the following equation (4).

The ADC (Analog-to-Digital Converter) 302 converts an analog output represented by a complex amplitude output by the optical coherent receiving unit 300 into a digital signal. The digital signal processing unit 303 is, for example, a DSP (Digital Signal Processor), and includes an equalizer 310, a frequency offset compensation unit 311 and a phase offset compensation unit 312. The equalizer 310 performs adaptive equalization processing on the digital signal output by the ADC 302 to compensate for waveform distortion caused by propagation, such as wavelength dispersion, polarization mode dispersion, polarization rotation, and the like.

As shown in the equation (4), the phase of the received signal has a form in which the frequency offset component 2πf _{IF t is superimposed on the phase modulation component φ m (t).} Therefore, on the IQ plane, the phase of the symbol rotates with the passage of time, and the phase of the symbol cannot be detected correctly. If the period of the symbol is T, the amount of rotation per symbol is 2πf _IF nT (n is a positive integer including 0). This phase rotation amount 2πf _IF nT is the frequency offset amount described above, and is an amount proportional to _{the frequency difference f IF between the signal light and the station emission.}

In the optical receiver 220, the frequency offset compensating unit 311 performs digital signal processing to give the received signal a rotation of the same magnitude in the opposite direction to the _{frequency offset component 2πf IF nT on the IQ plane.} Compensate to enable phase detection. After that, the phase offset compensation unit 312 can reproduce the constellation of the transmission signal by performing a process of removing the phase offset component α which is constant in time. The output unit 304 decodes the output signal of the phase offset compensation unit 312 and outputs it to the outside.

In the method using a training signal, the frequency offset amount is estimated by analyzing the training code sequence transmitted to a specific section of the transmission signal on the receiving side. The training code sequence is, for example, a sequence of 0 or π in the case of BPSK.

In order to estimate the frequency offset amount by analyzing the training code sequence, it is necessary to detect the head position of the code sequence. As shown in FIG. 13, the format of each burst signal transmitted from the optical transmission device 210 is divided into fields of the preamble 400, the payload 401, and the end of burst 402 from the beginning.

The payload 401 contains a transmission data sequence. The preamble 400 contains data processed by the digital signal processing unit 303. A training code sequence 410 for timing detection and a training code sequence 411 for frequency offset estimation are provided in each of the two sections of the _{length L t} from the beginning of the field of the preamble 400 and the section of the length L _{f that follows.} included.

As shown in FIG. 14, in the frequency offset compensation unit 311 of the digital signal processing unit 303 of the optical receiver 220, the timing detection unit 500 first analyzes the timing detection training code sequence 410 to detect the timing, that is, Detects the beginning position of the training signal. After that, when the _{training code sequence 411 for frequency offset estimation existing after L t} from the head position is given to the multiplier 503, the analysis is performed for the code sequence and the frequency offset amount is obtained.

(Analysis of training code sequence for timing detection)
The analysis method of the training code sequence 410 for timing detection by the timing detection unit 500 will be described. The timing detection training code sequence 410 used for timing detection is s (n), and the head position of s (n) is n = n ₀ . At this time, s (n) is a code sequence satisfying the relationship shown in the following equation (5) in _{n 0} ≦ n ≦ n ₀ + L _{t / 2.}

As shown in equation (5), s (n) is the sequence length _{L t,} and the code sequence from the head to _L t / 2 _second, the code sequence from the _L t / 2-th to _{L t} th matching It is a code series to be used. The received signal including the frequency offset at the symbol time t = nT is represented by the following equation (6) as a function of n from the equation (4).

In the equation (6), φ _m (n) is s (n), that is, the phase modulation corresponding to the training code sequence 410 for timing detection, and in the case of BPSK, it is a value of 0 or π. The nth symbol of the received signal r (n) and _{the differential component d (n) of the n + L t} / second symbol are represented by the following equation (7) from the equation (6).

The "*" in r (n + L _t / 2) ^{* means the complex conjugate.} The following equation (8) is obtained by calculating the sum of d (n) from an arbitrary position n = n'to n = n'+ L _{t / 2.}

Here, from the equation (5), when n ₀ ≤ n ≤ n ₀ + L _t / 2, the following equation (9) holds.

Therefore, the equation (8) becomes the following equation (10).

In the equation (10), the first term on the right side is the sum of vectors having random phases, and there are components that cancel each other out. On the other hand, the second term is the sum of the same vectors that do not depend on n, and these components strengthen each other. Therefore, when the number of vectors included in the second term of the equation (10) is larger than the number of vectors included in the first term, the value of D (n') increases. When n'= n ₀ , the component of the first term on the right side of the equation (10) becomes 0, and the value of D (n') becomes the maximum.

By the above procedure, the timing detection unit 500 detects the maximum value of | D (n') | and calculates _{the head position n 0 of the timing detection training code sequence 410.} Since the sequence length of the timing detection training code sequence 410 is L _t , the head position of the frequency offset estimation training code sequence 411 can also be calculated by adding L _{t to n 0.} This enables the timing detection unit 500 to perform timing detection for outputting the received signal r (n) corresponding to the training code sequence 411 for frequency offset estimation to the multiplier 503. In the timing detection method, it is necessary to increase the value of the second term of the equation (10) in order to improve the accuracy of detecting the head position of the training signal, and for that purpose, the sequence of the training code sequence 410 for timing detection. there is a need to increase the length L _t.

(Analysis of training code sequence for frequency offset estimation)
The procedure for estimating the frequency offset amount performed by the frequency offset compensation unit 311 will be described with reference to FIG. 14. The training signal generation unit 501 shown in FIG. 14 generates and outputs a training signal from the training code sequence 411 for frequency offset estimation given in advance. The training signal generated from the symbols of the training code sequence is represented by the following equation (11) as a function of n from the equation (4).

On the other hand, the received signal r (n) corresponding to the training code sequence 411 for frequency offset estimation output by the timing detection unit 500 is as described in the above equation (6), and this equation is referred to as the following equation (12). Reposted as.

As described above, the timing detection unit 500 detects the head position of the training code sequence 411 for frequency offset estimation, and outputs the received signal r (n) including the symbols after the detected head position. However, in the timing detection by the timing detection unit 500, an error Δn (where Δn is an integer, ... -3, -2, -1, 0, 1, 2, 3, ... Is one of the values) may occur. Hereinafter, the description including the error Δn will be described.

The multiplier 503 multiplies the received signal r (n + Δn) including the position error by the complex conjugate t (n) ^{* of} the training signal t (n) output by the complex conjugate 502 to obtain the received signal r (n). The differential component d'(n) between n + Δn) and the training signal t (n) is calculated. The differential component d'(n) is expressed as the following equation (13).

The delay device 504 takes in the differential component d'(n) output by the multiplier 503 and outputs d'(n + 1) delayed by one symbol. The complex conjugate 505 takes in d'(n + 1) and outputs the ^{complex conjugate d'(n + 1) *.} The multiplier 506 multiplies the differential component d'(n) by the d'(n + 1) ^* output by the complex conjugate 505 to calculate the one-symbol delay differential component represented by the following equation (14). ..

In the equation (14), when the timing detection error Δn is 0, the following equation (15) is obtained.

Therefore, the frequency offset amount is obtained as the declination angle of the equation (15) as expressed by the following equation (16).

In actual transmission, the received signal is affected by noise due to transmission, so it is necessary to remove the influence of the noise, and therefore, a plurality of 1-symbol delay differential components d'(n) d'(n + 1). Calculate the vector average of ^*. Vector averager 507, a one-symbol delay differential component from the multiplier 506 is output, calculates a vector mean of formula (15) over the sequence length L _f of the frequency offset estimation training code sequence 411. The decliner 508 outputs the declination of the vector average calculated by the vector averager 507. As a result, the frequency offset amount is obtained as shown in the following equation (17).

The frequency offset amount can also be obtained by calculating the average of the declinations as in the following equation (18).

As described above, the frequency offset amount can be calculated using the training signal included in the received signal and the training signal generated from the training code sequence 411 for frequency offset estimation given in advance. The frequency offset compensation processing unit 509 of the frequency offset compensation unit 311 performs frequency offset of the received signal using the calculated frequency offset amount, and outputs the received signal compensated for the frequency offset.

By the way, in the above procedure, as can be seen from the equation (14), when the position error Δn at the time of timing detection occurs by one symbol or more, it becomes difficult to perform the calculation for calculating the frequency offset amount. Therefore, in the method shown in Non-Patent Document 3 using a training signal, high accuracy is required for the timing detection unit 500 to detect the head position of the timing detection training code sequence 410. However, in order to perform position detection with high accuracy, it is necessary to lengthen the sequence length of the training code sequence 410 for timing detection, and there is a problem that it is difficult to apply it as a technique for estimating the frequency offset amount in the access network.

In view of the above circumstances, it is an object of the present invention to provide a technique for estimating a frequency offset amount with a wide estimation frequency range and a small number of training code sequences in a digital coherent method.

One embodiment of the present invention receives a coherent optical signal including a first training signal generated based on a predetermined training code sequence including a training code sequence for frequency offset estimation , and obtains the coherent optical signal from the received coherent optical signal. An optical receiver that compensates for the frequency offset of the received signal using the first training signal included in the received signal, from the first training signal included in the received signal to the training code sequence. a section including a part or all of the frequency offset estimation training code sequence contained, the frequency offset estimation training code sequence of a sequence for frequency offset estimation with the first training signal of a length corresponding to the length The first is generated based on the timing detection unit detected as the received training signal of the above, the one-symbol delay differential component of the detected received training signal, and the training code sequence for frequency offset estimation stored in advance by the own device . The training code sequence for frequency offset estimation includes a one-symbol delay differential component of two training signals and a frequency offset estimation unit that calculates a frequency offset amount based on the two training signals, and the training code sequence for frequency offset estimation is composed of multi-valued phase modulation symbols. argument of the vector average of one symbol delay differential component of the signal that a code sequence Ru generated based on the frequency offset estimate for training code sequence is a light receiving device is a code sequence at a predetermined angle.

One aspect of the present invention is the above-described optical receiving apparatus, said frequency offset estimation training code sequence, said first and second training signals are generated based on those said frequency offset estimation training code sequence When the vector average of each one-symbol delay differential component is calculated, it is a code sequence in which the vector averages are strengthened without canceling each other and the declination angle of the vector averages is a predetermined angle.

One aspect of the present invention is the optical receiver described above, wherein the training code sequence has a redundant code sequence added to either the front or the back, or the front or the back of the training code sequence for frequency offset estimation. When the redundant code sequence is added before the training code sequence for frequency offset estimation, the redundant code sequence is a partial sequence including the end of the training code sequence for frequency offset estimation, and is the same as the redundant code sequence. When the code sequence is the same as a partial sequence having the same sequence length and the redundant code sequence is added after the training code sequence for frequency offset estimation, the redundant code sequence is the training code sequence for frequency offset estimation. It is a partial sequence including the first and has the same code sequence as the partial sequence having the same sequence length as the redundant code sequence.

One aspect of the present invention is the above-mentioned optical receiving device, and when the received signal is a burst signal, the timing detection unit receives the light based on the light intensity of the burst signal and a predetermined threshold value. Detect training signals.

One aspect of the present invention is a frequency offset estimation method in an optical receiver that compensates for the frequency offset of a received signal obtained from a received coherent optical signal, and is a predetermined training code including a training code sequence for frequency offset estimation. The coherent optical signal including the first training signal generated based on the sequence is received, and the first training signal included in the received signal obtained from the received coherent optical signal is included in the training code sequence. Receive training for frequency offset estimation of the first training signal having a length corresponding to the sequence length of the frequency offset estimation training code sequence, which is a section including a part or all of the frequency offset estimation training code sequence. was detected as a signal, a one-symbol delay differential component of the detected said received training signal, a second training said optical receiving device is generated based on distichum wavenumber offset estimation training code sequence before stored in advance The frequency offset amount is calculated based on the one-symbol delay differential component of the signal, and the frequency offset estimation training code sequence is a code sequence consisting of multi-valued phase modulation symbols and the frequency offset estimation training. This is a frequency offset estimation method, which is a code sequence in which the deviation angle of the vector average of the one-symbol delay differential component of the signal generated based on the code sequence is a predetermined angle.

According to the present invention, in the digital coherent method, the frequency range to be estimated is wide, and it is possible to estimate the frequency offset amount with a small number of training code sequences.

It is a block diagram which shows the structure of the optical transmission system in 1st Embodiment of this invention. It is a figure which shows the structure of the burst frame generated in the optical transmission apparatus of the same embodiment. It is a block diagram which shows the internal structure of the frequency offset compensation part of the same embodiment. It is a flowchart which shows the flow of processing by the timing detection part and the frequency offset estimation part of the same embodiment. It is a figure explaining the frequency offset amount in the IQ plane in the same embodiment. It is a figure (the 1) which shows the example of the redundant code sequence added to the training code sequence for frequency offset estimation in the same embodiment. It is a figure (No. 2) which shows the example of the redundant code sequence added to the training code sequence for frequency offset estimation in the same embodiment. It is a figure (the 1) explaining the frequency offset amount in the IQ plane in the 2nd Embodiment. It is a figure (the 2) explaining the frequency offset amount in the IQ plane in the same embodiment. It is a figure explaining the detection method of the head position in 3rd Embodiment. It is a block diagram which shows the structure of a PON system. It is a block diagram which shows the structure of the optical transmission system of a digital coherent system. It is a figure which shows the structure of the burst signal in the PON system. It is a block diagram which shows an example of the structure which estimates the frequency offset amount in a digital coherent type optical receiver.

(First Embodiment)
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a schematic block diagram showing a configuration of an optical transmission system S according to the first embodiment. The optical transmission system S includes an optical transmission device 1 and an optical reception device 2. The optical transmission device 1 and the optical reception device 2 are communicably connected via the optical transmission line 3.

The optical transmission device 1 includes a frame generation unit 10, an IQ signal generation unit 11, and an optical coherent transmission unit 12.
The frame generation unit 10 takes in transmission data given from the outside and generates the burst frame 10out shown in FIG. The burst frame 10out has each field of preamble, payload, and end of burst in order from the beginning. The frame generation unit 10 includes the transmission data given from the outside in the payload to generate the burst frame 10out.

Further, the frame generation unit 10 includes in the preamble a training code sequence 60 for timing detection _{having a sequence length of L t} and a training code sequence 70 for frequency offset estimation _{having a sequence length of L f} , which are stored in advance in the internal storage area. Burst frame 10out is generated. In the first embodiment, the method of the timing detection unit 500 described above is applied to the method of detecting the head position of the training signal performed by the optical receiver 2 described later. Therefore, the timing detection training code sequence 60 included in the preamble by the frame generation unit 10 is a code sequence that satisfies the relationship of the above equation (5).

In the following description, the term "training code sequence" means a code sequence including both the training code sequence 60 for timing detection and the training code sequence 70 for frequency offset estimation. However, when the description "training code sequence" is used in the case where the preamble described later does not include the training code sequence 60 for timing detection, it means the training code sequence 70 for frequency offset estimation.

The IQ signal generation unit 11 takes in the burst frame 10out generated by the frame generation unit 10 and generates the modulation signal 11out of the I component and the Q component. In the first embodiment, for example, the IQ signal generation unit 11 shall generate a modulated signal by BSDK (Binary Phase shift Keying).

The optical coherent transmission unit 12 generates a transmission signal of a coherent optical signal by modulating the light output by the internal light source by using the modulation signals of the I component and the Q component generated by the IQ signal generation unit 11. The optical coherent transmission unit 12 transmits the generated transmission signal to the optical receiver 2 via the optical transmission path 3.

The optical receiving device 2 includes an optical coherent receiving unit 20, a local oscillator 21, an ADC 22, a digital signal processing unit 23, and an output unit 24.

The optical coherent receiving unit 20 performs coherent reception for measuring the interference waveform of the received light received via the optical transmission line 3 by the phase diversity method and the station emission output by the local oscillator 21, and is an analog signal of the I component and the Q component. Is output.

The ADC 22 converts the analog signals of the I component and the Q component output from the optical coherent receiving unit 20 into digital signals. The digital signal processing unit 23 is, for example, a DSP, and includes an equalizer 31, a frequency offset compensation unit 32, and a phase offset compensation unit 33. The equalizer 31 performs adaptive equalization processing on the digital signal output by the ADC 22, and compensates for waveform distortion caused by propagation, for example, wavelength dispersion, polarization mode dispersion, polarization rotation, etc., received signal 31out. Is output.

The frequency offset compensation unit 32 calculates the frequency offset amount using the training signal included in the received signal 31out output by the equalizer 31. Further, the frequency offset compensation unit 32 performs a reverse rotation calculation on the received signal 31out output by the equalizer 31 using the calculated frequency offset amount, and outputs the received signal 32out from which the phase rotation of the frequency offset is removed. do.

The phase offset compensation unit 33 removes a temporally constant phase offset component from the reception signal 32out output by the frequency offset compensation unit 32, and outputs a reception signal 33out in which the constellation of the transmission signal is reproduced. The output unit 24 decodes the received signal 33out and outputs it to the outside.

FIG. 3 is a block diagram showing an internal configuration of the frequency offset compensation unit 32. The frequency offset compensation unit 32 includes a timing detection unit 320, a training signal analysis unit 321, a frequency offset estimation unit 322, and a frequency offset compensation processing unit 323.

The timing detection unit 320 detects the position of the training signal included in the received signal 31out. In the first embodiment, as described above, the method by the timing detection unit 500 is applied as the method for detecting the head position of the training signal by the timing detection unit 320.

The training signal analysis unit 321 includes a training signal generation unit 41, a delay device 42, a complex conjugate device 43, a multiplier 44, a vector averager 45, and an argument device 46. The frequency offset estimation unit 322 includes a delay device 51, a complex conjugate device 52, a multiplier 53, a vector averager 54, a decliner 55, and a subtractor 56.

In the training signal analysis unit 321 the training signal generation unit 41 generates and outputs a training signal t (n) from the training code sequence 70 for frequency offset estimation given in advance. It is assumed that the training code sequence 70 for frequency offset estimation given in advance is stored in advance in the internal storage area of the training signal generation unit 41 by the following means. For example, the optical receiving device 2 estimates the frequency offset by means such as the optical receiving device 2 receiving from the optical transmitting device 1 or the optical receiving device 2 reading from the recording medium via another communication line having good communication quality. The information of the training code sequence 70 is taken in internally. The training signal generation unit 41 writes and stores the information of the captured training code sequence 70 for frequency offset estimation in the internal storage area in advance.

In the training signal analysis unit 321 and the frequency offset estimation unit 322, the delay devices 42 and 51 output a signal delayed by one symbol with respect to the captured signal. The complex conjugates 43 and 52 output the complex conjugate signal of the captured signal. The multipliers 44 and 53 multiply the two captured signals and output the result of the multiplication. The vector averager 54 obtains and outputs a vector average by calculating the sum of the vectors in a predetermined range. The decliners 46 and 55 calculate the declination of the vector. In the frequency offset estimation unit 322, the subtractor 56 calculates and outputs the difference between the two captured declinations.

The frequency offset compensation processing unit 323 compensates for the frequency offset with respect to the received signal 31out using the frequency offset amount calculated by the frequency offset estimation unit 322.

(Calculation of declination θ _{t of training signal)}
Before the frequency offset compensation unit 32 captures the received signal 31out, the training signal analysis unit 321 _{calculates the declination θ t} of the training signal. The training signal generation unit 41 generates and outputs a training signal t (n) from the training code sequence 70 for frequency offset estimation given in advance.

For example, in the case of BPSK, the training code sequence is a binary code sequence of 0 and π, and the training signals t (n) are Aexpj (ωt) and Aexpj (ωt + π). For reference, in the case of QPSK (Quadrature Phase Shift Keying), the training code sequence is a four-valued code sequence of + 1 / 4π, -1 / 4π, + 3 / 4π, and -3 / 4π, and the training signal t (n). Is Aexpj (ωt + 1 / 4π), Aexpj (ωt-1 / 4π), Aexpj (ωt + 3/4π), and Aexpj (ωt-3 / 4π).

The delay device 42 captures the training signal t (n) and outputs the signal t (n + 1) after one symbol delay of the captured training signal t (n). The complex conjugate 43 takes in the t (n + 1) output by the delayer 42 and outputs the complex conjugate t (n + 1) ^* of t (n + 1).

The multiplier 44 multiplies the training signal t (n) by the t (n + 1) ^* output by the complex conjugate 43 to calculate the one-symbol delay differential component Δt (n). The one-symbol delay differential component Δt (n) is expressed by the following equation (19) from the equation (11).

In addition, in the equation (19), the replacement shown in the equation (20) is performed.

The vector averager 45 calculates the vector sum Δt a (n) over the sequence length L _f of the frequency offset estimation training code sequence 70. The vector sum calculated by the vector averager 45 is expressed by the following equation (21).

The decliner 46 outputs the declination of the vector sum output by the vector averager 45, that is, θ _t .

(Flow of processing to estimate the amount of frequency offset)
Next, the flow of processing for calculating and estimating the frequency offset amount by the timing detection unit 320 and the frequency offset estimation unit 322 by the timing detection unit 320 and the frequency offset estimation unit 322 will be described with reference to the flowchart shown in FIG.

The timing detection unit 320 captures the reception signal 31out output by the equalizer 31, and detects the head position of the training signal included in the reception signal 31out as the first head position using the timing detection training code sequence 60. do. The sequence length of the training code sequence 60 for timing detection is L _t . Therefore, the timing detector 320, the head position of the training signal interval by adding a signal length corresponding to the sequence length L _t in the first start position corresponding to the frequency offset estimation training code sequence 70, that is, the second head Detect the position. The timing detection unit 320 outputs the reception signal r (n) including the detected symbols after the second head position as a reception training signal (step S101).

In the first embodiment, it is premised that the detection of the head position by the timing detection unit 320 includes an error, as will be described later. Therefore, the term "received training signal" includes the term "training signal" in it, but it is possible that some sequences other than the training code sequence 70 for frequency offset estimation are included due to a detection error. The following explanation will be given assuming that there is a property.

The delay device 51 captures the received signal r (n) output by the timing detection unit 320, and outputs the signal r (n + 1) after one symbol delay of the captured received signal r (n). The complex conjugate 52 takes in the r (n + 1) output by the delayer 51 and outputs the complex conjugate r (n + 1) ^* of r (n + 1). The multiplier 53 multiplies the received signal r (n) by the r (n + 1) ^* output by the complex conjugate 52 to calculate the one-symbol delay differential component Δr (n) (step S102). The one-symbol delay differential component Δr (n) is expressed by the following equation (22) from the equation (12).

From equation (22), the phase of the one-symbol delay differential component Δr between adjacent symbols (n) is expressed it can be seen that the sum of the modulation phase difference Δθ between the adjacent symbols and the frequency offset component 2 [pi] f _{IF T.} From this, the frequency offset amount can be obtained by the following equation (23) using the declination of the one-symbol delay differential component Δr (n).

The relationship shown in the equation (23) is shown in FIG. 5 (a) on the IQ plane. Further, when the frequency offset amount is estimated by the equation (23), the range of the frequency offset amount that can be estimated is as shown by the following equation (24).

Therefore, the frequency range in which the frequency offset amount can be estimated is as shown in the equation (25).

In the equation (25), B is the baud rate of the symbol in the transmission signal as described above.

To eliminate the effects of noise, the vector averager 54 calculates the vector sum over a temporal average, i.e. 1 sequence length L _f of the symbol delay differential component [Delta] r (n) the frequency offset estimation training code sequence 70 (Step S103). The vector sum calculated by the vector averager 54 is expressed by the following equation (26).

The decliner 55 calculates the declination of the vector sum output by the vector averager 54, that is, θ _r (step S104).

The subtractor 56 subtracts θ _{t from θ r} to calculate the frequency offset amount (step S105). The calculated frequency offset amount can be expressed as the following equation (27).

As shown in FIG. 5 (b), on the IQ plane, the frequency is obtained by subtracting _{θ t} output by the training signal analysis unit 321 from the declination angle of ΣΔr (n) which is the output of the vector averager 54, that is, θ _r. it can be determined 2 [pi] f _IF T offset amount.

In the above processing, as shown in equations (21) and (26), the interval in which the vector averaging unit 45 of the training signal analysis unit 321 obtains the vector average and the vector averaging unit 54 of the frequency offset estimation unit 322 are the vector averaging. It is necessary that the interval for which is obtained _{coincides with n = 0 to L f.}

When the detection error is included in the first head position detected by the timing detection unit 320, the error is also included in the second head position corresponding to the head position of the training code sequence 70 for frequency offset estimation. .. If this error is Δn (where Δn is an integer and is a value of any of -3, -2, -1, 0, 1, 2, 3, ...), Δn is generated. In this case, the equation (26) becomes the following equation (28).

From the equations (21), (26), and (28), it is necessary that the relationship shown in the following equation (29) is established in order to eliminate the influence of the timing detection error.

For example, if the relationship shown in the following equation (30) is established in the relationship between the training code sequence 70 for frequency offset estimation and the code sequences before and after the sequence, the condition of the equation (29) is approximately satisfied. be able to. As a result, it is possible to reduce the influence of the detection error by the timing detection unit 320.

The relationship shown by the equation (30) is that the sequence length L _f of the training code sequence 70 for frequency offset estimation is sufficiently long with respect to Δn and the vector average of the vectors constituting Δt (n) is calculated. , It is to satisfy that it is a code sequence having Δt (n) that does not cancel each other out and conversely strengthens each other.

In other words, in the configuration of the first embodiment, even if there is an error in the detection of the first head position by the timing detection unit 320, the pattern of the combination of symbols included in the training code sequence 70 for frequency offset estimation and the pattern By devising the series length, the influence of error can be suppressed. Therefore, the position detection accuracy of the timing detection unit 320 does not require the high accuracy required of the timing detection unit 500, and the timing detection unit 320 is the training signal corresponding to the training code sequence 70 for frequency offset estimation. The position should be roughly detected. Therefore, the sequence length of the timing detection training code sequence 60 can also be shortened, and the communication efficiency can be improved. Further, as the detection method by the timing detection unit 320, a detection method having a lower accuracy than the detection method using the training code sequence 60 for timing detection may be applied.

The argument on the left side of the equation (30) is θ _{t from the equation (21), and in order to calculate the frequency offset amount by the calculation of θ r} − θ _t as shown in the equation (27), the equation (30) is used. ), The length of the vector of the vector average of the one-symbol delay differential component Δt (n) does not become zero, and the argument θ _t needs to have a predetermined angle on the IQ plane. If the length of the vector of the vector average of the one-symbol delay differential component Δt (n) becomes zero, the relationship of the equation (30) will not be established if there is even a slight error. If the relationship of the equation (30) is not established, the noise component due to the error becomes large, so that θ _r becomes a random value, and it becomes difficult to calculate the frequency offset amount by calculation. The predetermined angle of the declination θ _t is the declination of the vector of the training code series that satisfies the relationship of the equation (30), and is the length of the composite vector that becomes a predetermined angle depending on the amount of the detection error. The required value increases or decreases.

An example in BPSK is shown below as an example of the training code sequence 70 for frequency offset estimation that strengthens each other without canceling each other. As described above, the code sequence of BPSK is a binary value of 0 and π, and is a code sequence in which π and 0 are alternately exchanged, for example, (π, 0, π, 0, π, 0, ...). Is set to the training code sequence 70 for frequency offset estimation. In this case, the one-symbol delay differential component Δt (n) is (+ π, −π, + π, −π, ...) From the equation (19). Since + π = −π, as a result, the one-symbol delay differential component Δt (n) becomes (+ π, + π, + π, + π, ...).

When the vector average is calculated on the IQ plane for this (+ π, + π, + π, + π, ...) code sequence, it strengthens each other without canceling each other, so it becomes larger than the noise component due to the error, and the equation It will have the relationship shown in (30). Therefore, it can be said that the code sequence is an appropriate code sequence as the training code sequence 70 for frequency offset estimation.

When the code sequence satisfying the following equation (31) is used in the equation (29), the condition of the equation (29) is completely satisfied, so that it is possible to eliminate the influence of the detection error by the timing detection unit 320. Become.

For example, by adding a redundant code sequence satisfying the equation (31) before and after the training code sequence 70 for frequency offset estimation, a code sequence completely satisfying the condition of the equation (29) can be realized. .. As an example of the redundant code sequence, for ease of understanding, FIGS. 6 and 7 show a case of a QSPC 4-valued code sequence instead of the BPSK which is the object of the first embodiment. In FIGS. 6 and 7, the four values of QPSK (+ 1 / 4π, -1 / 4π, +3 / 4π, -3 / 4π) are shown using (0,1,2,3).

6 (a) is an example of adding a redundant code sequence 71a and the redundant code sequence 72a before and after the frequency offset estimation training signal 70a of QPSK having a sequence length _{L f.} The redundant code sequence 71a has the same code sequence as the code sequence 70a-e _{L f ahead in terms of the number of symbols.} Further, the redundant code sequences 72a also has a L _f-number after the code sequence 70a-s of the same code sequence in number of symbols. Therefore, the redundant code sequence 71a and the redundant code sequence 72a are code sequences satisfying the equation (31).

Further, instead of adding the redundant code sequences 71a and 72a before and after the training signal 70a for frequency offset estimation as shown in FIG. 6A, the redundant code sequences 72a (or 71a) are added to either the front or the back. It may be added. For example, FIG. 6B is an example in which a redundant code sequence 72a is added after the training signal 70a for frequency offset estimation. When a redundant code sequence 72a (or 71a) is added before or after the training signal 70a for frequency offset estimation, the position at the beginning of the section for calculating the vector average is determined by the error that occurs and the redundant code sequence 72a (or the redundant code sequence 72a). Or it is necessary to determine according to the length of 71a).

For ease of understanding, an example modulated to a 10-value symbol value will be described. For example, suppose that the original symbol is a series of "0,1,2,3,4,5,6,7,8,9" with a series length of 10. If the redundant code sequence "0,1,2,3", which is the same sequence as the four symbols from the beginning of the original symbol, is added after the original symbol, the sequence length 14 is "0,1,2,3". , 4, 5, 6, 7, 8, 9, 0, 1, 2, 3 "is obtained. In this case, to extract the 10 symbols as the first is a section for calculating the vector average after two symbols from the beginning of the redundant code sequence has been added the training code sequence L _f, "2,3,4,5,6, It becomes a series of "7,8,9,0,1". If the vector average is calculated based on the sequence of the extracted 10 symbols, the vector average can be calculated while satisfying the relationship of the equation (31) when an error within 2 symbols before and after the timing detection occurs. _{Note that FIG. 6B is an example in which the head of L f} , which is a section for calculating the vector average, is set after one symbol, and the length of the redundant code sequence to be added is determined according to the error that occurs. The start position for calculating the vector average is determined within the range of the length of the redundant code series.

FIG. 7 is a diagram showing another example of generating a redundant training code sequence satisfying the equation (31). For example, the training code sequence 70b is generated by giving π and π / 2 phase rotations for each symbol to the four-valued symbols [0, 1, 2, 3]. As shown in FIG. 7, the generated training code sequence 70b is [0,2,3,1,2,0,1,3,0,2,3,1,2,0,1,3, ..., 0,2,3,1,2,0,1,3]. This is a pattern in which a series consisting of eight symbols [0,2,3,1,2,0,1,3] indicated by reference numeral 70bs in FIG. 7 is repeated.

_{In this case, if L f,} which is an interval for calculating the vector average, is a multiple of 8 symbols, a part of the generated training code sequence 70b without adding a new redundant code sequence as shown in FIG. Can be used as a redundant training code sequence. _{For example, when L f} , which is a section for calculating the vector average, is a section indicated by reference numeral 73b, the four symbols before and after the section of reference numeral 73b are designated as redundant code sequences 71b and 72b, and the section indicated by reference numeral 74b including the redundant code sequence is used. It can be used as a redundant training code sequence. Note that FIG. 7 shows an example in which the four symbols before and after the section of the reference numeral 73b are the redundant code sequences 71b and 72b, but the redundant code sequences 71b and 72b are not limited to the four symbols before and after, and the four symbols before and after are not limited. All peripheral symbols (eg, sections followed by a repeating pattern of 8 symbols) may be used. In this case, the section including all the sections in which the repetition pattern of 8 symbols continues may be used as a redundant training code sequence. In the following, a case will be described in which the four symbols before and after are designated as redundant code sequences 71b and 72b, and the section indicated by the reference numeral 74b including the redundant code sequence is used as a redundant training code sequence.

In this way, when a redundant training code sequence satisfying the equation (31) is generated by repeating a specific pattern as shown by the reference numeral 70b-s, the length Lf of the section for calculating the vector average is set to a specific _{length L f.} A redundant training code sequence 74b can be generated by simply including the symbols before and after the pattern as long as it is a multiple of the pattern. Therefore, the method shown in FIG. 7 does not require a procedure of newly adding a redundant code sequence as in the method shown in FIG. 6, so that the training code sequence is more easily made redundant than the method shown in FIG. 74b can be obtained. Further, also in the example of FIG. 7, as in FIG. 6B, only one of the redundant code sequences 71b and 72b is added, and the position at the beginning of the section for calculating the vector average is added. It may be determined within the range of the length of one of the redundant code sequences 71b and 72b.

By the way, in general, in an optical transmission system, the optical transmission device and the optical reception device have a limitation in the frequency band in which they can operate, and the same code is continuously used in the transmitted / received signals (for example, π, π, π, π, 0). , 0, 0, 0), it contains a low frequency component, and waveform deterioration due to the low frequency component may become a problem. Therefore, in the configuration of the first embodiment, it is desirable to use a code sequence such that the number of continuations of the same code is as small as possible as the training code sequence 70 for frequency offset estimation.

In the configuration of the first embodiment, as described above, when calculating the vector average of the one-symbol delay differential component, the results of the vector average of the one-symbol delay differential component are strengthened in order to improve the noise immunity. Such a training code sequence 70 for frequency offset estimation is required. Therefore, in the configuration of the first embodiment, it is sufficient that one symbol delay differential component of the training signal generated from the training code sequence 70 for frequency offset estimation is required, and the code sequence itself can be freely selected. become. Therefore, it is possible to select a code sequence that does not have a continuation of the same code, and by making such a selection, it is possible to suppress waveform deterioration due to low frequency components.

According to the configuration of the first embodiment described above, the optical receiver 2 receives a coherent optical signal including a training signal generated based on a predetermined training code sequence. The timing detection unit 320 included in the frequency offset compensation unit 32 of the optical receiver 2 is a section including a part or all of the training code sequence 70 for frequency offset estimation included in the training code sequence from the received signal, and is for frequency offset estimation. A received signal having a length corresponding to the sequence length of the training code sequence 70 is detected as a received training signal for frequency offset estimation. The frequency offset estimation unit 322 is based on the one-symbol delay differential component of the received training signal detected by the timing detection unit 320 and the one-symbol delay differential component of the training signal generated from the training code sequence 70 for frequency offset estimation. Then, the frequency offset amount used for the frequency offset compensation is calculated. The frequency offset compensation processing unit 323 compensates for the frequency offset of the received signal using the frequency offset amount calculated by the frequency offset estimation unit 322. The training code sequence 70 for frequency offset estimation is, for example, a code sequence consisting of BPSK symbols, and the deviation angle of the vector average of one symbol delay differential component of the training signal generated from the training code sequence 70 for frequency offset estimation is It is a code sequence having a predetermined angle. This makes it possible to estimate the frequency offset amount with a wide code sequence and a small code sequence in the digital coherent method. As a result, with a relatively simple configuration, it is possible to reduce the signal-to-noise ratio and suppress wavelength deterioration, improve the code error rate, extend the maximum transmission distance of the optical transmission line, and extend the optical transmission line. It is possible to increase the number of branches in. Further, in the configuration of the first embodiment, the applicable frequency range is −B / 2 ≦ f _IF <B / 2, regardless of the number of multiple values of phase modulation, which is a wide frequency range. Further, since high accuracy is not required for detecting the position of the training signal in the received signal, the number of the training code series 60 for timing detection can be reduced, so that the communication efficiency can be improved.

(Second embodiment)
As a second embodiment, the training code sequence 70 for frequency offset estimation used when QPSK is applied as a modulation method to the configuration of the first embodiment will be described. FIG. 8A is an example of the training signal t (n) generated from the training code sequence 70 for frequency offset estimation in the case of QPSK. 8 (b) shows a received signal r (n) output by the timing detection unit 320 when the optical transmitting device 1 transmits the training signal of FIG. 8 (a) and the optical receiving device 2 receives the training signal. ), That is, it is a figure which showed the received training signal.

As shown in FIG. 8A, the training code sequence 70 for frequency offset estimation is a code sequence that gives π and π / 2 phase rotation for each symbol, that is, (π / 4,5π / 4,7π /). The sequence of 4,3π / 4,5π / 4) is given.

At n = 0 in FIG. 8A, the position of the solid circle of reference numeral 81 indicates “π / 4” of the training signal t (n). Note that FIG. 8 (a) shows a case where there is no influence of the frequency offset when n = 0 for the sake of simplification of the explanation. Therefore, it is shown by a solid circle of reference numeral 81r in FIG. 8 (b). The position of the received signal r (n) is the same as that of the reference numeral 81 in the IQ plane, which is the position of "π / 4".

At n = 1 in FIG. 8 (a), the position of the solid line circle of reference numeral 82 is "+ π after phase rotation from the position of the broken line circle of reference numeral 81p indicating the position of the previous" π / 4 ". 5π / 4 ”is shown. At n = 1 in FIG. 8 (b), the broken line circle of reference numeral 82t indicates the position of “5π / 4” indicated by reference numeral 82 of n = 1 in FIG. 8 (a).

Similarly, the solid circles of reference numerals 83 to 85 in FIG. 8A indicate the positions corresponding to each of the code series (7π / 4), (3π / 4), and (5π / 4). The broken line circles of reference numerals 82p, 83p, 84p in FIG. 8 (a) indicate the positions of the previous reference numerals 82, 83, 84, and the broken lines of reference numerals 83t, 84t, 85t in FIG. 8 (b). The circles indicate the positions of reference numerals 83, 84, 85 in FIG. 8 (a).

8 (b), the case n = 1, the the phase rotation of 2 [pi] f _IF T min is applied as a frequency offset with respect to the training signal t (n), the position of the received signal r (n), solid line code 82r It becomes the position of the circle. Similarly, when n = 2, the order 4πf _IF T in the frequency offset is applied, the position of the received signal r (n), the position of the solid circle symbols 83r. in the case of n = 3, since the 6πf _IF T in the frequency offset is applied, the position of the received signal r (n), the position of the solid circle symbols 84r. Also, when n = 4, for 8πf _IF T in the frequency offset is applied, the position of the received signal r (n), the position of the solid circle symbols 85r.

9 (a) is a diagram showing the vector average of the one-symbol delay differential component Δt (n) of the training signal of FIG. 8 (a). In the case of FIG. 8A, since the appearance frequencies of π and π / 2 are equal, the declination of the vector average of the one-symbol delay differential component Δt (n) of the training signal t (n) is shown in FIG. 9A. As shown in, it becomes "3π / 4". _{This is θ t} in Eq. (27).

9 (b) is a diagram showing the vector average of the one-symbol delay differential component Δr (n) of the received signal r (n) of FIG. 8 (b) and the declination θ _{r of the vector average.} Therefore, the frequency offset amount can be calculated by subtracting _{θ t from θ r,} and the _{f IF} from the equation (27) becomes the value represented by the following equation (32).

The above code series (π / 4,5π / 4,7π / 4,3π / 4,5π / 4) does not have the same code, and there is no difference in the code even when the long-term average is obtained. It is a series. It can be said that such a code sequence is generally a code sequence in which the adverse effect due to the transient response on the receiving side in optical communication is small. In the code sequence, the one-symbol delay differential component Δt (n) shown as an example of BPSK described above is concentrated at one point on the IQ plane like (+ π, + π, + π, + π, ...), And is a vector average. It is not a code sequence that always strengthens each other when the above is obtained. However, the code sequence is a code sequence in which one symbol delay differential component Δt (n) is concentrated at two adjacent points on the IQ plane. Therefore, even in such a code sequence, the vector average of the one-symbol delay differential component Δt (n) is larger than the noise component due to the error, and therefore has the relationship shown in the equation (30). Therefore, the code sequence (π / 4,5π / 4,7π / 4,3π / 4,5π / 4) is an appropriate code sequence as the training code sequence 70 for frequency offset estimation.

(Third embodiment)
In the configuration of the first embodiment described above, the method of detecting the head position of the training signal by the timing detection unit 500 is applied to the timing detection unit 320. As described above, in the configuration of the first embodiment, since the timing detection unit 320 is not required to perform the position detection with high accuracy as required by the timing detection unit 500, another detection method is applied. You can also do it. As a third embodiment, the other detection method applicable to the timing detection unit 320 will be described.

As described above, in the optical transmission device 1, the frame generation unit 10 generates and transmits the burst frame 10out, so that the reception signal received by the optical reception device 2 is a burst signal. As shown in FIG. 10A, for example, the waveform of the light intensity of the burst signal is not a square wave, but a waveform that gradually increases linearly and then becomes a constant light intensity.

The position where the training code sequence 70 for frequency offset estimation is written in the preamble of the burst frame 10out is predetermined. _{Therefore, as shown in FIG. 10A, the time n'from the} time nth when the light intensity of the burst signal exceeds the threshold value to the head position of the training code sequence 70 for frequency offset estimation is almost constant. Therefore, the timing detection unit 320 determines whether or not the light intensity exceeds the threshold value based on the light intensity of the received signal and a predetermined threshold value, and the time n when the light intensity exceeds the threshold value. For example, a timer provided inside is activated in _th.

The timing detection unit 320 determines whether or not the time indicated by the timer matches n', and if the time indicated by the timer matches n', the position of the received signal at the time of matching is set as the training code for frequency offset estimation. It is detected as the head position corresponding to the section of the series 70. The timing detector 320, the head position, i.e. time _{n th} + n 'from _{n th} + _n' + _L received signals to _f corresponding to the section of the frequency offset estimation training code sequence 70 the received signal r (n), that is, the reception It is output to the frequency offset estimation unit 322 as a training signal 90. As a result, the frequency offset estimation unit 322 can obtain the frequency offset amount based on the received training signal 90.

Further, there is another method using a code sequence in which a redundant code sequence is added before and after the training code sequence 70 for frequency offset estimation described in the first embodiment. For example, all of the preambles are code sequences in which the same sequence as the redundant code sequence shown in FIG. 7 is repeated. In this case, how the signal length sections of which corresponds to a to sequence length L _f in arbitrary section in the training signal. Thus, taken interval, since the section of redundant training code sequence 74b length were extracted arbitrarily from L _f, or frequency offset estimation training code sequence 70b itself, or, redundant code sequence 71b, 72b of the Since all or part of the training code sequence for frequency offset estimation is part of the 70b, the frequency offset amount can be obtained. The procedure of the process, timing detection unit 320, receives a received signal r (n) from the time _{n th} light intensity exceeds a predetermined is the threshold of _n th + _{L f} section as shown in FIG. 10 (b) It is output to the frequency offset estimation unit 322 as a training signal 91. The frequency offset estimation unit 322 is a procedure of calculating the frequency offset amount based on the received training signal 91.

In the case of the burst signal shown in FIG. 10B, it is not necessary to include the training code sequence 60 for timing detection in the preamble. Further, also in the case of FIG. 10A, it is not always necessary to include the training code sequence 60 for timing detection.

In the third embodiment described above, the determination process of "whether or not the light intensity exceeds the threshold value" is only an example, and "whether or not the light intensity is equal to or higher than the threshold value" depending on how the threshold value is determined. May be determined. Further, the determination process of "whether the time indicated by the timer matches n'" is only an example, and "whether the time indicated by the timer exceeds n'" and "the time indicated by the timer is n'. Whether or not it becomes the above may be determined.

In the first to third embodiments described above, the training signal analysis unit 321 of the frequency offset compensation unit 32 shown in FIG. 3 calculates _{θ t from the training code sequence 70 for frequency offset estimation.} The configuration of the present invention is not limited to the embodiment. For example, the training signal analysis unit 321 may be provided in another external device, and θ _t may be calculated in the other device. For example, when the other device is the optical transmission device 1, the training signal analysis unit 321 in the optical transmission device 1 _{calculates θ t} from the training code sequence for frequency offset estimation, and the calculated θ _t has good communication quality. It is transmitted to the optical receiver 2 via another line. The optical receiving device 2 may have a configuration in which the received θ _{t is taken in and the θ t} taken in by the frequency offset compensating unit 32 is stored in an internal storage area. In the case of this configuration, when the subtractor 56 subtracts θ _{t from θ r} , it reads _{θ t} from the storage area.

In the first to third embodiments described above, the optical coherent receiving unit 20 of the optical receiving device 2 shown in FIG. 1 is supposed to perform coherent reception of the phase diversity method, but is to perform coherent reception of the polarization diversity method. You may do it.

Although BPSK has been described in the first embodiment and QPSK has been described in the second embodiment, the configuration of the present invention is not limited to the embodiment. The present invention may calculate the frequency offset amount for a more multi-valued modulation method such as M-PSK (Phase Shift Keying) or QAM (Quadrature Amplitude Modulation) in which the value of M exceeds 4. The present invention has the same effect on such a multi-level modulation method.

In the third embodiment described above, a burst signal is assumed as the signal to be transmitted / received, but in the first and second embodiments, the signal may not necessarily be a burst signal, and the light intensity of the transmitted / received signal is high. It may be a signal that is constant in time.

The digital signal processing unit 23 in the above-described embodiment may be realized by a computer. In that case, a program for realizing this function may be recorded on a computer-readable recording medium, and the program recorded on the recording medium may be read by a computer system and executed. The term "computer system" as used herein includes hardware such as an OS and peripheral devices. Further, the "computer-readable recording medium" refers to a portable medium such as a flexible disk, a magneto-optical disk, a ROM, or a CD-ROM, and a storage device such as a hard disk built in a computer system. Further, a "computer-readable recording medium" is a communication line for transmitting a program via a network such as the Internet or a communication line such as a telephone line, and dynamically holds the program for a short period of time. It may also include a program that holds a program for a certain period of time, such as a volatile memory inside a computer system that is a server or a client in that case. Further, the above program may be for realizing a part of the above-mentioned functions, and may be further realized for realizing the above-mentioned functions in combination with a program already recorded in the computer system. It may be realized by using a programmable logic device such as FPGA (Field Programmable Gate Array).

Although the embodiments of the present invention have been described in detail with reference to the drawings, the specific configuration is not limited to this embodiment, and includes designs and the like within a range that does not deviate from the gist of the present invention.

1 ... Optical transmitter, 2 ... Optical receiver, 3 ... Optical transmission path, 10 ... Frame generator, 11 ... IQ signal generator, 12 ... Optical coherent transmitter, 20 ... Optical coherent receiver, 21 ... Local oscillator, 22 ... ADC, 23 ... Digital signal processing unit, 24 ... Output unit, 31 ... Equalizer, 32 ... Frequency offset compensation unit, 33 ... Phase offset compensation unit

Claims (5)

A coherent optical signal including a first training signal generated based on a predetermined training code sequence including a training code sequence for frequency offset estimation is received, and the received signal included in the received signal obtained from the received coherent optical signal. An optical receiver that compensates for the frequency offset of the received signal using the first training signal.
A section including a part or all of the frequency offset estimation training code sequence included from the first training signal included in the received signal in the training code sequence, sequence of the frequency offset estimation training code sequence A timing detector for detecting the first training signal having a length corresponding to the length as a received training signal for frequency offset estimation, and a timing detection unit.
The detected 1-symbol delay differential component of the received training signal and the 1-symbol delay differential component of the second training signal generated based on the frequency offset estimation training code sequence stored in advance by the own device. A frequency offset estimation unit that calculates the frequency offset amount based on,
The training code sequence for frequency offset estimation is
A code sequence consisting of symbols of the multilevel phase modulation code argument of the vector average of one symbol delay differential component of the signal that will be generated based on the frequency offset estimate for training code sequence has a predetermined angle A series of optical receivers. The frequency offset estimation training code sequence,
When calculating the vector average of those that frequency offset estimation training code sequence to one symbol delay differential component of each of the first and second training signal is generated based, constructive rather not cancel each other, and wherein The optical receiver according to claim 1, wherein the declination of the vector average is a code sequence having a predetermined angle. The training code sequence is
A redundant code sequence is added before, after, or either before or after the training code sequence for frequency offset estimation.
When the redundant code sequence is added before the training code sequence for frequency offset estimation, the redundant code sequence is a partial sequence including the end of the training code sequence for frequency offset estimation and is the same as the redundant code sequence. It is a code sequence that is the same as the partial sequence of the sequence length,
When the redundant code sequence is added after the training code sequence for frequency offset estimation, the redundant code sequence is a partial sequence including the beginning of the training code sequence for frequency offset estimation and is the same as the redundant code sequence. The optical receiver according to claim 1 or 2, which has the same code sequence as the partial sequence of the sequence length. The timing detection unit
The optical receiving device according to any one of claims 1 to 3, wherein when the received signal is a burst signal, the received training signal is detected based on the light intensity of the burst signal and a predetermined threshold value. .. It is a frequency offset estimation method in an optical receiver that compensates for the frequency offset of the received signal obtained from the received coherent optical signal.
Receives a coherent optical signal containing a first training signal generated based on a predetermined training code sequence including a training code sequence for frequency offset estimation.
A section including a part or all of the frequency offset estimation training code sequence included in the training code sequence from the first training signal included in the received signal obtained from the received coherent optical signal, and the frequency. The first training signal having a length corresponding to the sequence length of the training code sequence for offset estimation is detected as a received training signal for frequency offset estimation.
One symbol delay and one-symbol delay differential component of the detected said received training signal, a second training signal the light receiving device is generated based on distichum wavenumber offset estimation training code sequence before stored in advance Calculate the frequency offset amount based on the differential component and
The training code sequence for frequency offset estimation is
A code sequence consisting of multi-valued phase modulation symbols, in which the declination of the vector average of the one-symbol delay differential component of the signal generated based on the training code sequence for frequency offset estimation is a predetermined angle. Is a frequency offset estimation method.

Images