US-20260128801-A1 - SIGNAL PROCESSING METHOD AND SIGNAL PROCESSING DEVICE

US20260128801A1US 20260128801 A1US20260128801 A1US 20260128801A1US-20260128801-A1

Abstract

A signal processing method for converting an input signal sampled at a first rate into an output signal sampled at a second rate in a communication system, includes input/output interval calculation processing of calculating input/output interval data related to a temporal interval between an input sample and an output sample adjacent in the input signal and the output signal, the input/output interval data being able to be shared in calculation of each output sample, based on the first rate and the second rate, coefficient calculation processing of calculating a coefficient sequence used for calculation of an output sample to be calculated based on an input sample around the output sample, and output signal calculation processing of calculating the output sample to be calculated using the input/output interval data and the coefficient sequence.

Inventors

Takahiro ODAGAWA

Assignees

NEC CORPORATION

Dates

Publication Date: 20260507
Application Date: 20251014
Priority Date: 20241101

Claims (18)

1 . A signal processing method for converting an input signal sampled at a first rate into an output signal sampled at a second rate in a communication system, the signal processing method comprising: input/output interval calculation processing of calculating input/output interval data related to a temporal interval between an input sample and an output sample adjacent in the input signal and the output signal, the input/output interval data being able to be shared in calculation of each output sample, based on the first rate and the second rate; coefficient calculation processing of calculating a coefficient sequence used for calculation of an output sample to be calculated based on an input sample around the output sample; and output signal calculation processing of calculating the output sample to be calculated using the input/output interval data and the coefficient sequence.
2 . The signal processing method according to claim 1 , wherein the output signal calculation processing uses a high-order polynomial that interpolates at least two input samples with a curve, and the input/output interval calculation processing calculates an interval value indicating the interval and a power of the interval value included in the high-order polynomial as the input/output interval data.
3 . The signal processing method according to claim 1 , wherein the input/output interval calculation processing applies, as an interval between the input sample and the output sample adjacent to each other, an interval between an output sample and an input sample immediately before the output sample, or an interval between an output sample and an input sample immediately after the output sample.
4 . The signal processing method according to claim 1 , wherein the input/output interval calculation processing calculates data including a constant multiple of a value based on the interval as the input/output interval data, and the constant is a number for reducing the number of times of multiplication other than an integer power of 2 in the coefficient calculation processing.
5 . The signal processing method according to claim 1 , wherein the input/output interval calculation processing calculates the input/output interval data for each of the intervals of a plurality of patterns, and the output signal calculation processing selects input/output interval data used to calculate the output sample to be calculated from the input/output interval data of the plurality of patterns.
6 . The signal processing method according to claim 1 , wherein the input/output interval calculation processing calculates the input/output interval data for each of the intervals of N patterns (N is an integer of 2 or more), and the output signal calculation processing includes branch processing of inputting a coefficient sequence associated with the output sample to be calculated to any one of the calculation processing associated with the N patterns, and calculation processing of calculating the output sample to be calculated using the input/output interval data of the associated pattern and the coefficient sequence input by the branch processing in each of the N patterns of calculation processing.
7 . The signal processing method according to claim 1 , for converting a sampling rate of each of the plurality of input signals input in parallel from the first rate to the second rate and outputting the plurality of output signals in parallel, wherein the input/output interval calculation processing is executed in common for a plurality of combinations of the input signals and the output signals; and the coefficient calculation processing and the output signal calculation processing are executed in parallel for each of the plurality of combinations.
8 . The signal processing method according to claim 1 , further comprising parallelization processing of parallelizing the input signal into a plurality of parallel channels, the parallelization processing performing parallelization in such a way that at least one first parallel channel among the plurality of parallel channels is configured by an input sample having the interval of zero, wherein an input sample included in the first parallel channel is output as an output sample, the coefficient calculation processing further calculates the coefficient sequence based on the surrounding input samples included in at least a part of the plurality of parallel channels, with the output samples having the interval other than zero as a calculation target, and the output signal calculation processing further calculates the output sample to be calculated using the input/output interval data and the coefficient sequence.
9 . The signal processing method according to claim 1 , wherein the signal processing method is used for sampling rate conversion of a reception signal in an optical signal communication system using an optical fiber, and the reception signal is used as the input signal.
10 . A signal processing device for converting an input signal sampled at a first rate into an output signal sampled at a second rate in a communication system, the signal processing device comprising: an input/output interval calculation circuit configured to calculate input/output interval data related to a temporal interval between an input sample and an output sample adjacent in the input signal and the output signal, the input/output interval data being able to be shared in calculation of each output sample, based on the first rate and the second rate; a coefficient calculation circuit configured to calculate a coefficient sequence used for calculation of an output sample to be calculated based on an input sample around the output sample; and an output signal calculation circuit configured to calculate the output sample to be calculated using the input/output interval data and the coefficient sequence.
11 . The signal processing device according to claim 10 , wherein the output signal calculation circuit uses a high-order polynomial that interpolates at least two input samples with a curve, and the input/output interval calculation circuit calculates an interval value indicating the interval and a power of the interval value included in the high-order polynomial as the input/output interval data.
12 . The signal processing device according to claim 10 , wherein the input/output interval calculation circuit applies, as an interval between the input sample and the output sample adjacent to each other, an interval between an output sample and an input sample immediately before the output sample, or an interval between an output sample and an input sample immediately after the output sample.
13 . The signal processing device according to claim 10 , wherein the input/output interval calculation circuit calculates data including a constant multiple of a value based on the interval as the input/output interval data, and the constant is a number for reducing the number of times of multiplication other than an integer power of 2 in the coefficient calculation circuit.
14 . The signal processing device according to claim 10 , wherein the input/output interval calculation circuit calculates the input/output interval data for each of the intervals of a plurality of patterns, and the output signal calculation circuit selects input/output interval data used to calculate the output sample to be calculated from the input/output interval data of the plurality of patterns.
15 . The signal processing device according to claim 10 , wherein the input/output interval calculation circuit calculates the input/output interval data for each of the intervals of N patterns (N is an integer of 2 or more), and the output signal calculation circuit includes a branch circuit configured to input a coefficient sequence associated with the output sample to be calculated to any one of the calculation circuit associated with the N patterns, and a calculation circuit configured to calculate the output sample to be calculated using the input/output interval data of the associated pattern and the coefficient sequence input by the branch circuit in each of the N patterns of the calculation circuit.
16 . The signal processing device according to claim 10 , for converting a sampling rate of each of the plurality of input signals input in parallel from the first rate to the second rate and outputting the plurality of output signals in parallel, wherein the input/output interval calculation circuit is provided in common for a plurality of combinations of the input signals and the output signals, and the coefficient calculation circuit and the output signal calculation circuit are provided in parallel for each of the plurality of combinations.
17 . The signal processing device according to claim 10 , further comprising a parallelization circuit configured to parallelize the input signal into a plurality of parallel channels, the parallelization circuit performing parallelization in such a way that at least one first parallel channel among the plurality of parallel channels is configured by an input sample having the interval of zero, wherein an input sample included in the first parallel channel is output as an output sample, the coefficient calculation circuit calculates the coefficient sequence based on the surrounding input samples included in at least a part of the plurality of parallel channels, with the output samples having the interval other than zero as a calculation target, and the output signal calculation circuit calculates the output sample to be calculated using the input/output interval data and the coefficient sequence.
18 . The signal processing device according to claim 10 , wherein the signal processing device is used for sampling rate conversion of a reception signal in an optical signal communication system using an optical fiber, and the reception signal is used as the input signal.

Description

This application is based upon and claims the benefit of priority from Japanese patent application No. 2024-193241, filed on Nov. 1, 2024, the disclosure of which is incorporated herein in its entirety by reference. TECHNICAL FIELD The present disclosure relates to a signal processing method and a signal processing device. BACKGROUND ART At present, in an optical signal communication system using an optical fiber, a digital coherent technology for compensating and equalizing distortion caused by a transmitter, a receiver and a transmission path by digital signal processing on signal light received by a coherent optical receiver is used. In order to efficiently process a high-speed and large-capacity signal, it is required to reduce the calculation amount in such digital signal processing. One of the signal processing techniques for reducing the calculation amount is signal processing on a signal (hereinafter, referred to as a fractional oversampled signal) input at a non-integral multiple oversampling rate smaller than 2. For example, M. Arikawa and K. Hayashi, “Frequency-domain adaptive MIMO filter with fractional oversampling using stochastic gradient descent for long-haul transmission over coupled 4-core fibers”. Vol. 31, No. 8/10 Apr. 2023/Optics Express 13104-13124 discloses adaptive multi-input multi-output (MIMO) filter processing on a fractional oversampled signal in a frequency domain. SUMMARY Here, in a general optical signal communication system, signal processing by double oversampling is performed based on the Nyquist condition. For example, in M. Arikawa and K. Hayashi, “Frequency-domain adaptive MIMO filter with fractional oversampling using stochastic gradient descent for long-haul transmission over coupled 4-core fibers”. Vol. 31, No. 8/10 Apr. 2023/Optics Express 13104-13124, as pre-processing, wavelength dispersion compensation and frame synchronization processing are performed on a signal (hereinafter, described as a double oversampled signal) subjected to double oversampling. Thereafter, MIMO processing is performed on the signal converted from the double oversampled signal to the fractional oversampled signal. As described above, in a case where the pre-processing by the general-purpose double oversampling and the signal processing on the fractional oversampled signal are continuously performed, it is necessary to convert a sampling rate between the two processing. However, there is a problem that a calculation amount increases in order to accurately perform such sampling rate conversion. Such a problem may occur not only in optical signal communication but also in other communication fields. The present disclosure has been made in view of the above problems, and an exemplary object of the present disclosure is to provide a technique for reducing a calculation amount required for sampling rate conversion in a communication system. A signal processing method according to one exemplary aspect of the present disclosure, for converting an input signal sampled at a first rate into an output signal sampled at a second rate in a communication system, includes input/output interval calculation processing of calculating input/output interval data related to a temporal interval between an input sample and an output sample adjacent in the input signal and the output signal, the input/output interval data being able to be shared in calculation of each output sample, based on the first rate and the second rate, coefficient calculation processing of calculating a coefficient sequence used for calculation of an output sample to be calculated based on an input sample around the output sample, and output signal calculation processing of calculating the output sample to be calculated using the input/output interval data and the coefficient sequence. A signal processing device according to one exemplary aspect of the present disclosure, for converting an input signal sampled at a first rate into an output signal sampled at a second rate in a communication system, the signal processing device includes input/output interval calculation means for calculating input/output interval data related to a temporal interval between an input sample and an output sample adjacent in the input signal and the output signal, the input/output interval data being able to be shared in calculation of each output sample, based on the first rate and the second rate, coefficient calculation means for calculating a coefficient sequence used for calculation of an output sample to be calculated based on an input sample around the output sample, and output signal calculation means for calculating the output sample to be calculated using the input/output interval data and the coefficient sequence. According to an exemplary aspect of the present disclosure, it is possible to provide a technique for reducing a calculation amount required for sampling rate conversion in a communication system. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is