CN115426225A - A signal processing method and receiver of a super-Nyquist direct detection system - Google Patents

Abstract

The invention discloses a signal processing method and a receiver of a direct detection system exceeding Nyquist, wherein the method comprises the following steps: after the signal passes through the post filter, acquiring the output of the post filter and a training sequence; based on the training sequence and the output of the post filter, performing channel estimation again by using a least square method LS to obtain a tap coefficient of a maximum likelihood sequence estimation MLSE; after obtaining the MLSE tap coefficient, using the MLSE based on the Viterbi algorithm to find the best path, and after obtaining the final survival path, the signal on the path is the judged signal. The invention realizes the optimization of MLSE tap coefficients in the PAM4 signal direct detection technology and can greatly improve the system performance. And in the aspect of computational complexity, compared with the original system, the method does not increase a lot.

Description

A signal processing method and receiver of a super-Nyquist direct detection system

technical field

The present invention relates to the technical field of DDFTN (Direct Detection Fasterthan Nyquist, Direct Detection Faster than Nyquist) signal processing technology for short-distance high-speed optical communication, in particular to a method for realizing direct detection of PAM (Pulse Amplitude Modulation, pulse amplitude modulation) signal MLSE (Maximum Likelihood Sequence Estimation, Maximum Likelihood Sequence Estimation) tap coefficient optimized super-Nyquist direct detection system signal processing method and receiver in the technology.

Background technique

The massive growth of intensive applications such as data centers, media services, and cloud storage has led to a significant increase in data traffic in fiber-optic edge networks. Due to the large market size of such networks, the future development trend of optical transceivers must be cost-effective. The optical transmission system of intensity modulation (IM)/direct detection (DD) has significant advantages in terms of cost and reliability, and can meet the power consumption and cost-effectiveness required by these cost-sensitive applications, so in short-distance optical communication Usually consider using DD technology. In order to improve the transmission capacity in the IM-DD system, high-order modulation technology is usually used. The hotspots of modulation technology research in the IM-DD system mainly include DMT (Discrete Multi-Tone, discrete multi-tone) and PAM4 modulation based on multi-carrier modulation. Solution, in which DMT technology is highly complex to implement, while PAM4 has the advantages of simple implementation.

However, there are still some problems to be solved in PAM4 modulation in IM-DD systems, such as demodulation of PAM4 signals in band-limited systems, low-complexity equalization technology, and so on. Now the main application is a signal processing technology based on "linear equalizer + DDFTN" to deal with noise and intersymbol interference. Among them, DDFTN is composed of Postfilter plus MLSE. Postfilter is mainly used to combat the enhanced noise after linear equalizer processing, and MLSE deals with the remaining intersymbol interference.

In prior art one, a two-tap post-filter is added with MLSE. The post-filter has a duobinary mode, which can be easily realized by a two-tap FIR structure, where one tap coefficient is fixed at 1, and the other tap coefficient is not fixed, represented by α, and its transfer function is H(z) =1+αz ^-1 , different α corresponds to different post-filter shapes, different system rates correspond to different optimal parameters, α does not need to be calculated, and the system performance can be optimized by adjusting the value of α good. Because the post-filter has a fixed partial response mode, the memory length of MLSE corresponds to Postfilter, so the memory length of MLSE is one at this time, and the tap coefficient is the tap coefficient of the post-filter, because the memory length is short and no channel Estimation to calculate the tap coefficient of MLSE has a great advantage in computational complexity.

In the second prior art, a multi-tap post-filter is added with MLSE. In order to further improve the performance of the system, a multi-tap post-filter is proposed. The tap coefficient of the post-filter is no longer fixed. The tap coefficient of the post-filter is calculated by the autocorrelation function of the noise. By adding the post-filter The number of taps is used to improve the performance of the system. The tap coefficient of MLSE corresponds to the previous post-filter, and there is no need to perform channel estimation again to calculate the tap coefficient of MLSE. Compared with the two-tap Postfilter plus MLSE, more The performance of the tapped post-filter plus MLSE has a certain improvement, but the tap coefficient of the post-filter needs to be recalculated, and the computational complexity of MLSE increases with the increase of the tap coefficient. There is a tradeoff between complexity.

In summary, in the prior art, the tap coefficients of MLSE correspond to the tap coefficients of Postfilter, although this saves the recalculation of MLSE tap coefficients, but there is a great loss in performance, although by adding postfilter The number of taps can improve performance, but the effect is not significant. In addition, the channel response is truncated according to the memory length set by Postfilter and the transmission channel response beyond the memory length is discarded, which makes the MLSE algorithm lose part of the channel response in the process of compensating for signal damage, thus affecting the overall performance of the algorithm. It is necessary to further improve the prior art.

Contents of the invention

The present invention provides a kind of super-Nyquist direct detection system signal processing method and receiver, to solve the problem in the prior art, because the tap coefficient of MLSE corresponds to the tap coefficient of postfilter Postfilter, although save the need for MLSE The recalculation of the tap coefficients is a technical problem with a large loss in performance.

In order to solve the problems of the technologies described above, the present invention provides the following technical solutions:

In one aspect, the present invention provides a signal processing method for a super-Nyquist direct detection system, which is used for a receiving end in a super-Nyquist direct detection system, and the signal processing method includes:

After the signal passes through the post-filter, obtain the output of the post-filter and the training sequence;

Based on the output of the training sequence and the post-filter, the least squares method LS is used to re-estimate the channel to obtain the tap coefficient of the maximum likelihood sequence estimation MLSE;

After obtaining the tap coefficients of MLSE, use MLSE based on Viterbi algorithm to find the best path. After obtaining the final surviving path, the signal on the path is the signal of decision.

Further, based on the output of the training sequence and the post-filter, the least square method LS is used to perform channel estimation again, and the tap coefficients of the maximum likelihood sequence estimation MLSE are obtained, including:

When performing least squares estimation, first take a part of the training sequence as the training sequence of the least squares method; then according to the extracted training sequence of the least squares method and the output of the post filter, use the least squares method LS to perform channel estimation again, Get the tap coefficients of the maximum likelihood sequence estimation MLSE.

Further, based on the output of the training sequence and the post-filter, the least square method LS is used to perform channel estimation again, and the formula for obtaining the tap coefficient of the maximum likelihood sequence estimation MLSE is:

w=(X ^T X) ^-1 X ^T Y

Among them, w represents the tap coefficient of MLSE calculated by the least square method LS, Y represents the output of the post-filter, and X represents the training sequence of the least square method.

Further, the number of MLSE points does not exceed the original channel response.

On the other hand, the present invention also provides a super Nyquist direct detection receiver, comprising:

The MLSE tap coefficient calculation module is used to obtain the output of the post-filter and the training sequence after the signal passes through the post-filter; based on the output of the training sequence and the post-filter, the least square method LS is used to re-estimate the channel to obtain the maximum likelihood The sequence estimates the tap coefficients of MLSE;

The MLSE module is used to find the best path by using the MLSE based on the Viterbi algorithm after obtaining the tap coefficients of the MLSE. After obtaining the final surviving path, the signal on the path is the decision signal.

Further, the MLSE tap coefficient calculation module is specifically used for:

When performing least squares estimation, first take a part of the training sequence as the training sequence of the least squares method; then according to the extracted training sequence of the least squares method and the output of the post filter, use the least squares method LS to perform channel estimation again, Get the tap coefficients of the maximum likelihood sequence estimation MLSE.

Further, based on the output of the training sequence and the post-filter, the least square method LS is used to perform channel estimation again, and the formula for obtaining the tap coefficient of the maximum likelihood sequence estimation MLSE is:

w=(X ^T X) ^-1 X ^T Y

Among them, w represents the tap coefficient of MLSE calculated by the least square method LS, Y represents the output of the post-filter, and X represents the training sequence of the least square method.

Further, when using MLSE based on Viterbi algorithm to find the optimal path, the number of MLSE points does not exceed the original channel response.

In another aspect, the present invention also provides an electronic device, which includes a processor and a memory; at least one instruction is stored in the memory, and the instruction is loaded and executed by the processor to implement the above method.

In yet another aspect, the present invention also provides a computer-readable storage medium, wherein at least one instruction is stored in the storage medium, and the instruction is loaded and executed by a processor to implement the above method.

The beneficial effects brought by the technical solution provided by the present invention at least include:

The present invention proposes a new type of improved DDFTN. After Postfilter, channel estimation is carried out again through the least square method to obtain the tap coefficient of MLSE, so that the tap coefficient of MLSE is closer to the original channel response. With the increase of MLSE points, the performance of the system It will get better and better, but the number of points cannot exceed the original channel response. The technical solution of the present invention realizes the optimization of the MLSE tap coefficient in the PAM4 signal direct detection technology. Compared with the previous direct use of the Postfilter tap coefficient as the MLSE tap coefficient, the performance of the system is greatly improved after recalculating the MLSE tap coefficient. At the same time, since the LS algorithm is the basis of the channel estimation algorithm, and has low complexity, simple structure, stable performance, and easy implementation, the technical solution of the present invention does not increase much compared with the original system in terms of computational complexity.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings that need to be used in the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those skilled in the art, other drawings can also be obtained based on these drawings without creative effort.

Fig. 1 is a block diagram of a super-Nyquist direct detection system provided by an embodiment of the present invention;

Fig. 2 is the simplified model figure of the Postfilter that the embodiment of the present invention provides;

Fig. 3 is the least squares fitting image provided by the embodiment of the present invention;

FIG. 4 is a schematic diagram of an MLSE decision path based on a Viterbi algorithm provided by an embodiment of the present invention;

Fig. 5 is a flowchart of digital signal processing at the receiving end provided by an embodiment of the present invention.

Detailed ways

In order to make the object, technical solution and advantages of the present invention clearer, the implementation manner of the present invention will be further described in detail below in conjunction with the accompanying drawings.

first embodiment

This embodiment provides a signal processing method for a super-Nyquist direct detection system. As shown in FIG. 1 , in the super-Nyquist direct detection system, at the sending end, a pseudo-random binary sequence is mapped into a PAM4 format. After pre-emphasis, the PAM4 signal is uploaded to an arbitrary waveform generator (AWG). A power amplifier is then used to boost the electrical signal, driving a Mach-Zehnder modulator (MZM) with an ECL laser, and the output optical signal is launched into a standard single-mode fiber (SSMF).

At the receiving end, a variable optical attenuator (VOA) adjusts the received optical power (ROP) after the SSMF. The signal is then detected by a photodetector (PD) and captured by a real-time sampling oscilloscope (RTO). After being sampled by an analog-to-digital converter, the PAM4 signal is restored using digital signal processing (DSP). The signal passes through the postfilter Postfilter, and the simplified structure of the postfilter is shown in Figure 2. In the case that the training sequence and the output of the post-filter are known, this embodiment uses the least squares method LS to perform channel estimation again, and the fitting image of the least squares is shown in FIG. 3 . After obtaining the tap coefficients of MLSE, the Viterbi algorithm is used to find the optimal path. The MLSE based on the Viterbi algorithm is shown in Figure 4. The flow chart of DSP processing at the receiving end is shown in Figure 5.

Specifically, the DSP processing flow at the receiving end includes the following steps:

After the linear equalizer at the receiving end, the least mean square error algorithm LMS has eliminated most of the intersymbol interference, but at the same time the noise is also enhanced. In order to further suppress the enhanced noise, the signal passes through a two-tap post-filter. The transfer function of the post filter is:

H(z)＝1+αz ^-1 (1)

Among them, α represents the second tap coefficient of Postfilter, and the output of the postfilter can be expressed as:

Y(i)=r(i)+α*r(i+1) (2)

Among them, r(i) represents the signal received by the post-filter;

After obtaining the output of the post-filter, the next step is to use the simple least square method LS to perform channel estimation. Suppose D={(x ₁ ,y ₁ ),(x ₂ ,y ₂ ),...(x _N ,y _N )} is a data set, x _i ∈ R ^p , y _i ∈ R,

Among them, Y represents the output of the post-filter, and the linear equation of least squares can be expressed as:

f(w)=w ^T x (5)

Here, f(w) refers to the value of the training sequence, w refers to the calculated tap coefficients, and the least squares method finds the best function match for the data by minimizing the sum of squares of the errors. Unknown data can be easily obtained by using the least square estimation method, and the sum of the squares of the errors between the obtained data and the actual data is the smallest. The tap coefficient obtained by the least square method can be expressed as:

w＝(X ^T X) ^-1 X ^T Y (6)

When performing least squares estimation, first take a part of the training sequence as the training sequence of the least squares method, and the sequence length taken out also has the best corresponding value, which cannot be too long or too short, and the tap of MLSE is obtained in the least squares estimation After the coefficients, the Viterbi algorithm starts to find the most probable path from the extracted sequence. The number of MLSE taps based on the Viterbi algorithm is an odd number, and the number of taps is related to the memory length of the signal. The expression is as follows:

n=2*k+1 (7)

Among them, n represents the number of MLSE points, k represents the memory length of the signal, because the signal estimation is related to the symmetrical signal before and after, so the number of taps of MLSE is an odd number, when k=1, that is, the number of MLSE points is three. As the number of points increases, the system performance will become better and better. The expression of MLSE estimation is:

y _i =w ₁ x _i-1 +w ₂ x _i +w ₃ x _i+1 (8)

Among them, y _i represents the value estimated at time i, w represents the tap coefficient obtained by least squares, and x represents the possible value of the PAM4 signal. After obtaining the estimated value at the i-th moment, the minimum error corresponding to this moment can be calculated, that is, the minimum Euclidean distance, and its expression formula is:

d＝(y _i '-y _i ) ² (9)

Among them, d represents the Euclidean distance, y _i ' represents the estimated value, and y _i is the output of the post-filter, that is, the input of MLSE. The number of distances to be calculated for different signals is different when performing the Viterbi algorithm. When the signal is PAM4 and the MLSE memory length is 1, the corresponding possible path is 4 ² =16, because the possible state corresponding to each moment of PAM4 is There are 4 types (-3, -1, 1, 3), there is a set of corresponding distances between each two states, and the final reserved path distance is expressed as:

d _i ＝min{d _i-1 +(w ₁ x _i-1 +w ₂ x _i +w ₃ x _i+1 -y _i ) ² } (10)

After the final surviving path is obtained, the signal on the path is the decision signal.

To sum up, the technical solution of this embodiment further optimizes the tap coefficient of MLSE on the basis of the existing DDFTN, so that the tap coefficient of MLSE is closer to the original channel response, reduces the bit error rate, and improves the system performance. And this embodiment uses LS (Least Squares Method, Least Squares Method) to perform channel estimation again after the two-tap Postfilter. Compared with other channel estimation algorithms, the least squares algorithm is extremely simple, but it is easily affected by noise, because the previous The Postfilter has already filtered out most of the noise, so there is no need to worry about this problem. Although a certain amount of computational complexity is added, the performance of the system can be greatly improved. In the case of a significant improvement in system performance, the corresponding increase in computational complexity is acceptable, and it can effectively solve the problem that the MLSE decision signal is not accurate enough. The problem.

second embodiment

This embodiment provides a super-Nyquist direct detection receiver, which includes:

The MLSE tap coefficient calculation module is used to obtain the output of the post-filter and the training sequence after the signal passes through the post-filter; based on the output of the training sequence and the post-filter, the least square method LS is used to re-estimate the channel to obtain the maximum likelihood The sequence estimates the tap coefficients of MLSE;

The MLSE module is used to find the best path by using the MLSE based on the Viterbi algorithm after obtaining the tap coefficients of the MLSE. After obtaining the final surviving path, the signal on the path is the decision signal.

The super-Nyquist direct detection receiver of this embodiment corresponds to the signal processing method of the super-Nyquist direct detection system of the first embodiment above; wherein, in the super-Nyquist direct detection receiver of this embodiment The functions realized by each functional module correspond one-to-one to each process step in the signal processing method of the super-Nyquist direct detection system of the first embodiment; therefore, details are not repeated here.

third embodiment

This embodiment provides an electronic device, which includes a processor and a memory; at least one instruction is stored in the memory, and the instruction is loaded and executed by the processor, so as to implement the method of the first embodiment.

The electronic device may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) and one or more memories, wherein at least one instruction is stored in the memory, so The above instruction is loaded by the processor and executes the above method.

Fourth embodiment

This embodiment provides a computer-readable storage medium, where at least one instruction is stored, and the instruction is loaded and executed by a processor, so as to implement the method of the above-mentioned first embodiment. Wherein, the computer-readable storage medium may be ROM, random access memory, CD-ROM, magnetic tape, floppy disk, optical data storage device and the like. The instructions stored therein can be loaded by the processor in the terminal to execute the above method.

In addition, it should be noted that the present invention may be provided as a method, device or computer program product. Accordingly, embodiments of the invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, embodiments of the invention may take the form of a computer program product embodied on one or more computer-usable storage media having computer-usable program code embodied therein.

Embodiments of the present invention are described with reference to flowcharts and/or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each procedure and/or block in the flowchart and/or block diagram, and a combination of procedures and/or blocks in the flowchart and/or block diagram can be realized by computer program instructions. These computer program instructions may be provided to a general purpose computer, embedded processor, or other programmable data processing terminal processor to produce a machine such that instructions executed by the computer or other programmable data processing terminal processor produce instructions for A device for realizing the functions specified in one or more procedures of a flowchart and/or one or more blocks of a block diagram.

These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing terminal to operate in a specific manner, such that the instructions stored in the computer-readable memory produce an article of manufacture comprising instruction means, the The instruction means implements the functions specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram. These computer program instructions can also be loaded into a computer or other programmable data processing terminal equipment, so that a series of operational steps are performed on the computer or other programmable terminal equipment to produce computer-implemented processing, thereby The instructions executed above provide steps for implementing the functions specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

It should also be noted that in this document, the term "comprises", "comprises" or any other variation thereof is intended to cover a non-exclusive inclusion such that a process, method, article or terminal device comprising a series of elements includes not only those elements, but also other elements not expressly listed, or elements inherent in such process, method, article or terminal equipment. Without further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article or terminal device comprising said element.

Finally, it should be noted that the above description is a preferred embodiment of the present invention, and it should be pointed out that although the preferred embodiment of the present invention has been described, for those skilled in the art, once the basic creative concepts of the present invention are understood , under the premise of not departing from the principle of the present invention, some improvements and modifications can also be made, and these improvements and modifications should also be regarded as the protection scope of the present invention. Therefore, the appended claims are intended to be construed to cover the preferred embodiment and all changes and modifications which fall within the scope of the embodiments of the present invention.

Claims (8)

1. a super-Nyquist direct detection system signal processing method, for the receiver in the super-Nyquist direct detection system, it is characterized in that, the super-Nyquist direct detection system signal processing method comprises: After the signal passes through the post-filter, obtain the output of the post-filter and the training sequence; Based on the output of the training sequence and the post-filter, the least squares method LS is used to re-estimate the channel to obtain the tap coefficient of the maximum likelihood sequence estimation MLSE; After obtaining the tap coefficients of MLSE, use MLSE based on Viterbi algorithm to find the best path. After obtaining the final surviving path, the signal on the path is the signal of decision. 2. the super-Nyquist direct detection system signal processing method as claimed in claim 1, is characterized in that, described based on the output of training sequence and back filter, utilizes least square method LS to carry out channel estimation again, obtains maximum likelihood The sequence estimates the tap coefficients of MLSE, including: When performing least squares estimation, first take a part of the training sequence as the training sequence of the least squares method; then according to the extracted training sequence of the least squares method and the output of the post filter, use the least squares method LS to perform channel estimation again, Get the tap coefficients of the maximum likelihood sequence estimation MLSE. 3. the super-Nyquist direct detection system signal processing method as claimed in claim 2, is characterized in that, described based on the output of training sequence and back filter, utilizes least square method LS to carry out channel estimation again, obtains maximum likelihood However, the formula for the tap coefficient of sequence estimation MLSE is: w=(X ^T X) ^-1 X ^T Y Among them, w represents the tap coefficient of MLSE calculated by the least square method LS, Y represents the output of the post-filter, and X represents the training sequence of the least square method. 4. The super-Nyquist direct detection system signal processing method as claimed in claim 1, wherein the number of MLSE points does not exceed the original channel response. 5. A super-Nyquist direct detection receiver, characterized in that, comprising: The MLSE tap coefficient calculation module is used to obtain the output of the post-filter and the training sequence after the signal passes through the post-filter; based on the output of the training sequence and the post-filter, the least square method LS is used to re-estimate the channel to obtain the maximum likelihood The sequence estimates the tap coefficients of MLSE; The MLSE module is used to find the best path by using the MLSE based on the Viterbi algorithm after obtaining the tap coefficients of the MLSE. After obtaining the final surviving path, the signal on the path is the decision signal. 6. super Nyquist direct detection receiver as claimed in claim 5, is characterized in that, described MLSE tap coefficient calculation module is specifically used for: When performing least squares estimation, first take a part of the training sequence as the training sequence of the least squares method; then according to the extracted training sequence of the least squares method and the output of the post filter, use the least squares method LS to perform channel estimation again, Get the tap coefficients of the maximum likelihood sequence estimation MLSE. 7. super Nyquist direct detection receiver as claimed in claim 6, is characterized in that, described based on the output of training sequence and back filter, utilizes least square method LS to carry out channel estimation again, obtains maximum likelihood sequence The formula for estimating the tap coefficients of MLSE is: w=(X ^T X) ^-1 X ^T Y Among them, w represents the tap coefficient of MLSE calculated by the least square method LS, Y represents the output of the post-filter, and X represents the training sequence of the least square method. 8. The super-Nyquist direct detection receiver as claimed in claim 1, wherein when using MLSE based on Viterbi algorithm to find the optimal path, the number of MLSE points does not exceed the original channel response.

Images