Nonlinear damage signal processing method and device of space division multiplexing system

Abstract

The invention discloses a non-linear damage signal processing method and device for a space division multiplexing system, and relates to the technical field of network communication, and the method comprises the steps: inputting a non-linear damage signal in the space division multiplexing system into a feedforward neural network equalizer based on a cascade structure, in the feedforward neural network equalizer based on the cascade structure, an input layer is connected with a middle layer through a first weight, the middle layer is connected with an output layer through a second weight, and the input layer is connected with the output layer through a third weight; and carrying out equalization processing on the nonlinear damage signal through three weights in the feedforward neural network equalizer based on the cascade structure. According to the feed-forward neural network equalizer, the linear path which is directly connected with the input layer and the output layer is introduced into the feed-forward neural network equalizer, the effect of united linear and nonlinear optimization on the received signals can be achieved, nonlinear damage of the signals is effectively restrained, only the low number of real multiplication times is additionally introduced, and the extra overhead of complexity is extremely low.

Description

Technical Field

[0001] The present application relates to the field of network communication technology, and in particular to a method and device for processing nonlinear impairment signals of a space division multiplexing system. Background Art

[0002] With the rapid development of various services such as 5G, cloud computing and artificial intelligence, the data traffic in short-distance optical communication applications is growing rapidly. In order to meet these needs, intensity modulation direct detection (IM-DD) transceivers based on single-mode fiber or multimode fiber are widely adopted. Such systems have the advantages of simple structure, low cost and small size. At present, the typical 400G transmission scheme has been standardized based on the combination of pulse amplitude modulation (PAM) and wavelength division multiplexing. However, with the emergence of next-generation switching chips that support 800G, 1.6T optical interfaces or higher bandwidths, the cost, size and power consumption of transceivers may also increase dramatically with the increase in the number of parallel lasers.

[0003] In response to the exponential growth of network traffic, space division multiplexing (SDM) optical interconnect technology has been widely proposed, and it coincides with the high-density co-packaged optical interface that has been hotly discussed in recent years. Integrated lasers, modulators, photodetectors (PDs), and SDM multiplexers / demultiplexers (such as distributed feedback lasers, thin-film lithium niobate modulators, femtosecond laser direct writing core fan-in / fan-out devices, and mode multiplexing chips) are expected to be co-packaged with high-speed chips to achieve high-density optical interfaces. In addition, multiple parallel transmitters can share a laser, which greatly reduces the space, cost, and power consumption of the transceiver. However, in single-channel and space-division multiplexing optical interconnect systems, there are serious system nonlinear impairments. The existing solutions have weak nonlinear processing capabilities and are not suitable for the high-throughput parallel computing mode of actual receivers. The computational complexity is very high and it is difficult to be truly implemented in real receivers. Summary of the invention

[0004] The main purpose of the present application is to provide a method and device for processing nonlinear damage signals in a space-division multiplexing system, aiming to solve the technical problems of weak processing capability and high computational complexity of existing solutions for nonlinear damage.

[0005] To achieve the above object, the present application proposes a method for processing nonlinear impairment signals in a space division multiplexing system, the method comprising: Inputting a nonlinear impairment signal in a space division multiplexing system into a feedforward neural network equalizer based on a cascade structure, wherein the feedforward neural network equalizer based on the cascade structure comprises an input layer, an intermediate layer and an output layer, wherein the input layer is connected to the intermediate layer via a first weight, the intermediate layer is connected to the output layer via a second weight, and the input layer is connected to the output layer via a third weight; The nonlinear damage signal is equalized by using the first weight, the second weight and the third weight in the feedforward neural network equalizer based on the cascade structure to obtain an equalized signal.

[0006] In one embodiment, the step of performing equalization processing on the nonlinear damage signal by using the first weight, the second weight, and the third weight in the feedforward neural network equalizer based on the cascade structure to obtain an equalized signal includes: Determine the output value of the middle layer according to the first weight, the threshold of the middle layer, the activation function of the middle layer and the nonlinear damage signal in the feedforward neural network equalizer based on the cascade structure; The nonlinear damage signal is equalized according to the output value of the intermediate layer, the second weight, the third weight, the threshold of the output layer and the activation function of the output layer to obtain an equalized signal.

[0007] In one embodiment, the step of inputting the nonlinear impairment signal in the space division multiplexing system into a feedforward neural network equalizer based on a cascade structure comprises: receiving a nonlinear impairment signal of each branch in the space division multiplexing system, and setting a sliding window for the nonlinear impairment signal of each branch; The nonlinear damage signal within the sliding window range of each branch is input into the feedforward neural network equalizer based on the cascade structure in turn.

[0008] In one embodiment, the step of inputting the nonlinear impairment signal in the space division multiplexing system into a feedforward neural network equalizer based on a cascade structure comprises: Receive the nonlinear impairment signal of each branch in the space division multiplexing system, and aggregate the nonlinear impairment signals of all branches to obtain an aggregated single channel signal; A sliding window is set for the aggregated single-channel signal, and the aggregated single-channel signal within the sliding window is input into a feedforward neural network equalizer based on a cascade structure.

[0009] In one embodiment, after the step of performing equalization processing on the nonlinear damage signal by using the first weight, the second weight, and the third weight in the feedforward neural network equalizer based on the cascade structure to obtain an equalized signal, the step further includes: Generate a time delay feedback according to the equalized signal, and add the time delay feedback to the input layer to obtain a new input layer; Connecting the time-delay feedback in the new input layer to the intermediate layer via a fourth weight; Connecting the new input layer and the output layer through the third weight to obtain a new feedforward neural network equalizer; The nonlinear damage signal is subjected to secondary equalization processing by using the first weight, the second weight, the third weight and the fourth weight in the new feedforward neural network equalizer to obtain a signal after secondary equalization.

[0010] In one embodiment, the step of performing secondary equalization processing on the nonlinear damage signal by using the first weight, the second weight, the third weight and the fourth weight in the new feedforward neural network equalizer to obtain a secondary equalized signal includes: The formula for performing secondary equalization processing on the nonlinear damage signal by the new feedforward neural network equalizer is:

[0011] Where y represents the signal after secondary equalization, x represents the nonlinear damage signal to be processed in the sliding window, and f [1] represents the activation function of the middle layer, f [2] represents the activation function of the output layer, W [1] represents the first weight, W [2] represents the second weight, W c represents the third weight, W d represents the fourth weight, y d represents the time delay feedback generated according to the equalized signal, b [1] represents the threshold of the middle layer, b [2] represents the threshold of the output layer, and the superscript T represents the transposed symbol.

[0012] In addition, to achieve the above-mentioned purpose, the present application also proposes a nonlinear impairment signal processing device for a space division multiplexing system, the nonlinear impairment signal processing device for a space division multiplexing system comprising: A signal input module, used for inputting a nonlinear impairment signal in a space division multiplexing system into a feedforward neural network equalizer based on a cascade structure, wherein the feedforward neural network equalizer based on the cascade structure comprises an input layer, an intermediate layer and an output layer, wherein the input layer is connected to the intermediate layer via a first weight, the intermediate layer is connected to the output layer via a second weight, and the input layer is connected to the output layer via a third weight; The equalization processing module is used to perform equalization processing on the nonlinear damage signal through the first weight, the second weight and the third weight in the feedforward neural network equalizer based on the cascade structure to obtain an equalized signal.

[0013] In addition, to achieve the above-mentioned purpose, the present application also proposes a nonlinear damage signal processing device for a space-division multiplexing system, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the nonlinear damage signal processing method for a space-division multiplexing system as described above.

[0014] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the nonlinear damage signal processing method of the space division multiplexing system as described above are implemented.

[0015] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, the steps of the nonlinear impairment signal processing method of the space division multiplexing system as described above are implemented.

[0016] The present application provides a method for processing nonlinear impaired signals in a space-division multiplexing system, by inputting the nonlinear impaired signals in the space-division multiplexing system into a feedforward neural network equalizer based on a cascade structure, the feedforward neural network equalizer based on the cascade structure includes an input layer, an intermediate layer and an output layer, the input layer and the intermediate layer are connected by a first weight, the intermediate layer and the output layer are connected by a second weight, and the input layer and the output layer are connected by a third weight; the nonlinear impaired signals are equalized by the first weight, the second weight and the third weight in the feedforward neural network equalizer based on the cascade structure to obtain the equalized signal. The present application can achieve the effect of joint linear and nonlinear optimization of the received signal by introducing a linear path directly connecting the input and output layers in the feedforward neural network equalizer, better suppressing the nonlinear damage caused by signal transmission, thereby significantly reducing the system bit error rate and improving the system performance, and only additionally introducing a lower number of real number multiplications, the additional complexity overhead is extremely low, and does not affect the real-time implementation of the receiver. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0018] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0019] Figure 1 A flowchart of a method for processing a nonlinear impairment signal in a space division multiplexing system according to an embodiment of the present invention is provided; Figure 2 This is a schematic diagram of the structure of a feedforward neural network equalizer based on a cascade structure in this application; Figure 3 It is a flow chart of a space division multiplexing optical fiber communication system based on single-channel cascade feedforward neural network reception in this application; Figure 4 It is a flow chart of a space-division multiplexing optical fiber communication system based on extended full-channel cascade feedforward neural network reception in this application; Figure 5 A flow chart of the second embodiment of the method for processing nonlinear impairment signals in a space division multiplexing system of the present application; Figure 6 A flowchart of the third embodiment of the method for processing nonlinear impairment signals in a space division multiplexing system of the present application is provided; Figure 7 It is a schematic diagram of the structure of the new feedforward neural network equalizer in this application; Figure 8 This is a schematic diagram of the module structure of a nonlinear impairment signal processing device for a space division multiplexing system according to an embodiment of the present application; Fig. 9 Schematic diagram of the device structure of the hardware operating environment involved in the nonlinear impairment signal processing method of the space division multiplexing system in the embodiment of the present application.

[0020] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0021] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0022] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0023] The main solution of the embodiment of the present application is: inputting the nonlinear damage signal in the space division multiplexing system into a feedforward neural network equalizer based on a cascade structure, the feedforward neural network equalizer based on the cascade structure includes an input layer, an intermediate layer and an output layer, wherein the input layer is connected to the intermediate layer through a first weight, the intermediate layer is connected to the output layer through a second weight, and the input layer is connected to the output layer through a third weight; the nonlinear damage signal is equalized by the first weight, the second weight and the third weight in the feedforward neural network equalizer based on the cascade structure to obtain an equalized signal.

[0024] Due to the exponential growth of network traffic in the existing technology, space-division multiplexing optical interconnection technology has been widely proposed, and it coincides with the high-density co-packaged optical interface that has been hotly discussed in recent years. Integrated lasers, modulators, photodetectors, and SDM multiplexers / demultiplexers are expected to be co-packaged with high-speed chips to achieve high-density optical interfaces. In addition, multiple parallel transmitters can share a laser, which greatly reduces the space, cost, and power consumption of the transceiver. However, in single-channel and space-division multiplexing optical interconnection systems, there are serious system nonlinear damages. The existing solutions have weak nonlinear processing capabilities and are not suitable for the high-throughput parallel computing mode of actual receivers. The computational complexity is very high and it is difficult to be truly implemented in actual receivers.

[0025] The present application provides a solution, which inputs the nonlinear damage signal in the space division multiplexing system into a feedforward neural network equalizer based on a cascade structure, wherein the feedforward neural network equalizer based on the cascade structure includes an input layer, an intermediate layer and an output layer, wherein the input layer is connected to the intermediate layer through a first weight, the intermediate layer is connected to the output layer through a second weight, and the input layer is connected to the output layer through a third weight; the nonlinear damage signal is equalized by the first weight, the second weight and the third weight in the feedforward neural network equalizer based on the cascade structure to obtain an equalized signal. The present application can achieve the effect of joint linear and nonlinear optimization of the received signal by introducing a linear path directly connecting the input and output layers in the feedforward neural network equalizer, and better suppress the nonlinear damage caused by signal transmission, thereby significantly reducing the system bit error rate and improving the system performance, and only additionally introduces a lower number of real number multiplications, and the additional complexity overhead is extremely low, which does not affect the real-time implementation of the receiver.

[0026] It should be noted that the execution subject of the method of this embodiment can be an optical receiver equipped with a feedforward neural network equalizer based on a cascade structure; it can also be a nonlinear impairment signal processing device of a space division multiplexing system with the same or similar functions. This embodiment and the following embodiments will be described by taking the nonlinear impairment signal processing device of the space division multiplexing system as an example.

[0027] Based on this, the embodiment of the present application provides a method for processing nonlinear impairment signals in a space division multiplexing system, referring to Figure 1 , Figure 1 This is a flow chart of the first embodiment of the method for processing nonlinear impairment signals in a space division multiplexing system of the present application.

[0028] In this embodiment, the nonlinear impairment signal processing method of the space division multiplexing system includes steps S10 to S20: Step S10, inputting the nonlinear damage signal in the space division multiplexing system into a feedforward neural network equalizer based on a cascade structure, wherein the feedforward neural network equalizer based on the cascade structure includes an input layer, an intermediate layer and an output layer, wherein the input layer is connected to the intermediate layer through a first weight, the intermediate layer is connected to the output layer through a second weight, and the input layer is connected to the output layer through a third weight.

[0029] It is worth noting that this embodiment proposes a nonlinear damage signal processing method for a space-division multiplexing system, namely, a cascade-structured feedforward neural network-based equalization scheme (Cascade Feedforward Neural Network-Based Equalization, CFNNE), which solves the problem that the prior art has weak processing capability for nonlinear effects or high computational complexity, overcomes the signal damage caused by nonlinear effects, and improves the overall performance of the space-division multiplexing optical communication system.

[0030] refer to Figure 2 The structure of the feedforward neural network equalizer based on the cascade structure in this embodiment is explained. The existing neural network equalizer has a certain nonlinear signal processing capability through a unique layer design and a nonlinear activation function design for each layer. For example, an ordinary two-layer neural network has an input layer, an intermediate layer and an output layer, and the layers are connected by weights, that is, the input layer and the intermediate layer are connected by a first weight, and the intermediate layer and the output layer are connected by a second weight. This embodiment adds a cascade structure on the basis of the existing neural network equalizer, connects each layer to the subsequent layer, and for a two-layer neural network, directly connects its input layer and output layer through a third weight, thereby increasing a direct linear path from input to output.

[0031] It can be understood that the feedforward neural network equalizer based on the cascade structure in the present application is located at the optical receiver position of the space-division multiplexing optical communication system, and performs a new type of nonlinear processing on the downsampled signal. Figure 3The structure of the space-division multiplexing optical fiber communication system based on single-channel cascade feedforward neural network reception in this application is described. At the transmitting end, multiple optical transmitters send modulated signals and couple them into the space-division multiplexing optical fiber through fan-in devices. The space-division multiplexing optical fiber can be any type of multi-core optical fiber, few-mode optical fiber, or multi-core few-mode optical fiber. After transmission through the space-division multiplexing optical fiber, it is decoupled at the receiving end through the fan-out device to form a multi-channel signal corresponding to the transmitting end. Each received signal is pre-processed by a series of digital signal processing (DSP) such as resampling, clock synchronization, matched filtering, and downsampling, and then input into an optical receiver based on cascade structure feedforward or recurrent neural network equalization for final signal processing, and the bit error rate (BER) of the communication system is calculated and the system performance is evaluated. By introducing a linear path that directly connects the input and output on the existing nonlinear equalization scheme, the effect of joint linear and nonlinear optimization of the received signal can be achieved, the nonlinear damage caused by signal transmission can be better suppressed, the system bit error rate can be greatly reduced, and the system performance can be improved.

[0032] In a feasible implementation, step S10 may include steps S101-S102: Step S101, receiving a nonlinear impairment signal of each branch in a space division multiplexing system, and setting a sliding window for the nonlinear impairment signal of each branch.

[0033] Step S102, sequentially inputting the nonlinear damage signal within the sliding window range of each branch into a feedforward neural network equalizer based on a cascade structure.

[0034] It is understandable that the overall received signal can be processed in the form of a sliding window, that is, the nonlinear damage signal of each branch of the space division multiplexing system is collected, and a sliding window is set for each signal, and the received signal within the sliding window range of each branch is calculated in turn, and the received signal within the sliding window range can be any sampling rate. A similar signal equalization processing method can be used for each signal in the space division multiplexing system, and any signal can be nonlinearly equalized using the method proposed in this application.

[0035] In a feasible implementation manner, step S10 may further include steps S103-S104: Step S103, receiving the nonlinear impairment signal of each branch in the space division multiplexing system, and summing up the nonlinear impairment signals of all branches to obtain a summed single channel signal.

[0036] Step S104, setting a sliding window for the aggregated single-channel signal, and inputting the aggregated single-channel signal within the sliding window range into a feedforward neural network equalizer based on a cascade structure.

[0037] It should be noted that for the space-division multiplexing optical communication system, there will be multiple channels of original unprocessed received signals at the receiving end, which can also be processed by Figure 4 The extended full-channel cascade feedforward neural network receives the signal. Figure 3 The difference is the receiving part. In this embodiment, the multi-channel receiving signals that have been pre-processed with digital signals are aggregated to form a single-channel unprocessed receiving signal, and then input into a feedforward neural network equalizer based on a cascade structure to perform equalization processing on all the data, and all transmission signals are restored at the same time. This kind of data processing for the integrated data set usually requires larger system parameters than a single channel, that is, the sliding window width of the data set and the middle layer of the neural network, but on average for each single channel, the integrated signal processing does not require parameter increases that are so many times the number of space-division multiplexing channels. In addition, the number of equalization modules of the system receiver is reduced from the same as the space-division multiplexing channel to 1, further reducing the receiver cost of the space-division multiplexing optical communication system.

[0038] In this embodiment, by integrating the signals of multiple channels at the receiving end to form a single channel unprocessed signal, all data are processed at one time and a unified equalization operation is performed. Although a larger sliding window and intermediate layer parameters are required, the overall system complexity is significantly reduced because the number of equalization modules is reduced from being equal to the number of space division multiplexing channels to 1, which not only reduces the hardware cost, but also simplifies the receiving end architecture of the space division multiplexing optical communication system.

[0039] Step S20, performing equalization processing on the nonlinear damage signal by using the first weight, the second weight and the third weight in the cascade-based feedforward neural network equalizer to obtain an equalized signal.

[0040] It should be understood that the sliding window width for the received data set is , the middle layer of the neural network is ,The output layer is designed with only one neuron, which represents the calculated equalized received symbol. First, the output value of the intermediate layer is calculated according to the first weight, and then the equalized signal is calculated based on the intermediate output value according to the second weight and the third weight, forming a many-to-one correspondence with the received symbol to be processed in the current sliding window.

[0041] It should be noted that compared with the existing solutions, the additional complexity overhead introduced in this embodiment is very low. Here, the number of real number multiplications required to equalize a single symbol receiver is considered as a measure of complexity. The complexity of the cascade structure feedforward neural network equalization is expressed as , can be calculated using the following formula:

[0042] It can be seen from the above formula that, compared with the existing solution, the solution in this embodiment only introduces The number of real number multiplications is very low, and the additional complexity overhead is extremely low, which hardly affects the real-time implementation of the receiver. At the same time, due to the introduction of the additional linear path, the training process of the equalizer can also be simplified, and the number of iterations required for training can be greatly reduced.

[0043] The present embodiment provides a method for processing nonlinear impaired signals of a space-division multiplexing system, by inputting the nonlinear impaired signals in the space-division multiplexing system into a feedforward neural network equalizer based on a cascade structure, the feedforward neural network equalizer based on the cascade structure includes an input layer, an intermediate layer and an output layer, the input layer and the intermediate layer are connected by a first weight, the intermediate layer and the output layer are connected by a second weight, and the input layer and the output layer are connected by a third weight; the nonlinear impaired signals are equalized by the first weight, the second weight and the third weight in the feedforward neural network equalizer based on the cascade structure to obtain an equalized signal. The present application can achieve the effect of joint linear and nonlinear optimization of the received signal by introducing a linear path directly connecting the input and output layers in the feedforward neural network equalizer, better suppressing the nonlinear damage caused by signal transmission, thereby significantly reducing the system bit error rate and improving the system performance, and only additionally introducing a lower number of real number multiplications, the additional complexity overhead is extremely low, and does not affect the real-time implementation of the receiver.

[0044] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above-mentioned embodiment 1 can refer to the above introduction, and will not be repeated later. Figure 5 , step S20, the nonlinear impairment signal processing method of the space division multiplexing system further includes steps S201-S202: Step S201, determining the output value of the middle layer according to the first weight, the threshold of the middle layer, the activation function of the middle layer and the nonlinear damage signal in the feedforward neural network equalizer based on the cascade structure.

[0045] It can be understood that the sliding window width for the received data set is , the middle layer of the neural network is , the output layer is designed with only one neuron, representing the calculated equalized received symbol. Assume that the first weight corresponding to the input-intermediate layer is expressed as R , the second weight corresponding to the intermediate-output layer is expressed as R , the third weight corresponding to the input-output layer is expressed as R , the threshold of the middle layer is represented by b[1], and the threshold of the output layer is represented by b[2]. The formula for the output value of the middle layer is as follows:

[0046] in R represents the nonlinear damage signal to be processed within the sliding window, R represents the calculated output value of the intermediate layer, Represents the activation function of the intermediate layer of the neural network.

[0047] Step S202, performing equalization processing on the nonlinear damage signal according to the output value of the intermediate layer, the second weight, the third weight, the threshold of the output layer and the activation function of the output layer to obtain an equalized signal.

[0048] It can be understood that the final equalized symbol can be calculated next, as shown in the following formula:

[0049] in Represents the activation function of the output layer of the neural network, and the superscript T represents the matrix transpose. y represents the signal processed by the equalizer, and forms a many-to-one correspondence with the received symbol to be processed in the current sliding window. By sliding the sliding window with a step size of 1 on the overall received signal and calculating the data in each sliding window in turn according to the above method, the data after each calculation can be collected in turn to obtain the final equalized symbol.

[0050] In this embodiment, by calculating the output value of the middle layer according to the first weight, and then calculating the signal processed by the equalizer according to the second weight and the third weight combined with the output value of the middle layer, a many-to-one correspondence is formed with the received symbol to be processed in the current sliding window, and the data after each calculation can be collected in sequence to obtain the final equalized symbol. By introducing a linear path directly connecting the input and output in the traditional nonlinear equalization scheme, the effect of joint linear and nonlinear optimization of the received signal can be achieved, the nonlinear damage caused by signal transmission can be better suppressed, the system bit error rate can be greatly reduced, and the system performance can be improved.

[0051] Based on the first embodiment of the present application, in the third embodiment of the present application, the same or similar contents as those in the first embodiment can be referred to the above description, and will not be described in detail later. Figure 6 After step S20, the method for processing nonlinear impairment signals of the space division multiplexing system further includes steps S30 to S60: Step S30, generating a time delay feedback according to the equalized signal, and adding the time delay feedback to the input layer to obtain a new input layer.

[0052] It is worth noting that in this embodiment, the above feedforward neural network equalizer based on the cascade structure is further improved by adding the delay part of the equalized signal, and connecting its delay feedback to the input layer of the neural network. Delay feedback is generated according to the equalized signal y, and the delay feedback is added to the input layer to obtain a new input layer, which can be referred to Figure 7 , Figure 7 The first layer in the input layer is the new input layer, and the d1, d2…d k The part represents the time delay feedback part.

[0053] Step S40: connecting the time delay feedback in the new input layer and the intermediate layer through a fourth weight.

[0054] At the same time, the delay feedback in the new input layer is connected to the middle layer through the fourth weight, thereby increasing the communication path between the delay feedback part and the middle layer.

[0055] Step S50, connecting the new input layer and the output layer through the third weight to obtain a new feedforward neural network equalizer.

[0056] It should be noted that after obtaining the new input layer, the new input layer is also connected to the output layer through the third weight, which also increases the linear path during signal equalization and achieves a better joint linear and nonlinear optimization effect. However, unlike the third weight mentioned above, the third weight here is expressed as R , where k represents the number of equalization signal delay feedbacks. Thus, a new feedforward neural network equalizer further improved on the basis of the feedforward neural network equalizer based on the cascade structure can be obtained. This structure similar to decision feedback equalization provides multiple past equalized symbols as references for current signal equalization. These highly accurate past signal information are very helpful for estimating the inter-symbol crosstalk related to the current signal, and thus can significantly improve system performance.

[0057] Step S60, performing secondary equalization processing on the nonlinear damage signal by using the first weight, the second weight, the third weight and the fourth weight in the new feedforward neural network equalizer to obtain a secondary equalized signal.

[0058] In a feasible implementation manner, step S60 may further include step S601: Step S601, the formula for performing secondary equalization processing on the nonlinear damage signal by the new feedforward neural network equalizer is:

[0059] Where y represents the signal after secondary equalization, x represents the nonlinear damage signal to be processed in the sliding window, and f [1] represents the activation function of the middle layer, f [2] represents the activation function of the output layer, W [1] represents the first weight, W [2] represents the second weight, W c represents the third weight, W d represents the fourth weight, y d represents the time delay feedback generated according to the equalized signal, b [1] represents the threshold of the middle layer, b [2] represents the threshold of the output layer, and the superscript T represents the transposed symbol.

[0060] It can be understood that, assuming that the equalization signal delay feedback of this variant is k, the additional input brought about by this is expressed as y d ∈R k , then according to the above formula, the nonlinear damage signal can be subjected to secondary equalization to obtain the signal after secondary equalization.

[0061] It should be noted that the receiver complexity involved in the equalization processing scheme based on the new feedforward neural network equalizer, that is, the number of real multiplications required to equalize a single signal, is expressed as , can be calculated using the following formula:

[0062] In this embodiment, it can be seen from the above formula that, compared with the solution of the feedforward neural network equalizer based on the cascade structure in this application, the new feedforward neural network equalizer only introduces Compared with the traditional solution, this solution only introduces times the number of real number multiplications. Considering that the value of k in actual operation is much smaller than and The additional complexity overhead is also very low, and hardly affects the real-time implementation of the receiver. The above-mentioned variant can also be flexibly applied to the architecture of single-channel discrete equalization or multi-channel combined equalization space-division multiplexing optical communication system.

[0063] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the nonlinear damage signal processing method of the space division multiplexing system of the present application. More forms of simple transformations based on this technical concept are all within the protection scope of the present application.

[0064] The present application also provides a nonlinear damage signal processing device for a space division multiplexing system, please refer to Figure 8 , the nonlinear impairment signal processing device of the space division multiplexing system comprises: A signal input module 10, used to input a nonlinear impairment signal in a space division multiplexing system into a feedforward neural network equalizer based on a cascade structure, wherein the feedforward neural network equalizer based on the cascade structure includes an input layer, an intermediate layer and an output layer, wherein the input layer is connected to the intermediate layer through a first weight, the intermediate layer is connected to the output layer through a second weight, and the input layer is connected to the output layer through a third weight; The equalization processing module 20 is used to perform equalization processing on the nonlinear damage signal through the first weight, the second weight and the third weight in the feedforward neural network equalizer based on the cascade structure to obtain an equalized signal.

[0065] The nonlinear impairment signal processing device for a space division multiplexing system provided by the present application adopts the nonlinear impairment signal processing method for a space division multiplexing system in the above-mentioned embodiment, and can solve the technical problem. Compared with the prior art, the beneficial effects of the nonlinear impairment signal processing device for a space division multiplexing system provided by the present application are the same as the beneficial effects of the nonlinear impairment signal processing method for a space division multiplexing system provided by the above-mentioned embodiment, and other technical features in the nonlinear impairment signal processing device for a space division multiplexing system are the same as the features disclosed in the above-mentioned embodiment method, and are not described in detail here.

[0066] The present application provides a nonlinear impairment signal processing device for a space-division multiplexing system, the nonlinear impairment signal processing device for a space-division multiplexing system comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the nonlinear impairment signal processing method for the space-division multiplexing system in the above-mentioned embodiment 1.

[0067] Reference below Fig. 9, which shows a schematic diagram of the structure of a nonlinear impairment signal processing device for a space division multiplexing system suitable for implementing an embodiment of the present application. The nonlinear impairment signal processing device for a space division multiplexing system in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Fig. 9 The nonlinear impairment signal processing device for the space division multiplexing system shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0068] like Fig. 9 As shown, the nonlinear impairment signal processing device of the space division multiplexing system may include a processing device 1001 (such as a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. Various programs and data required for the operation of the nonlinear impairment signal processing device of the space division multiplexing system are also stored in the RAM 1004. The processing device 1001, the ROM 1002 and the RAM 1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: input devices 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 1003 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 1009. The communication device 1009 can allow the nonlinear impairment signal processing device of the space division multiplexing system to communicate with other devices wirelessly or by wire to exchange data. Although the figure shows a nonlinear impairment signal processing device of the space division multiplexing system with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems may be implemented or have instead.

[0069] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0070] The nonlinear impairment signal processing device of the space division multiplexing system provided by the present application adopts the nonlinear impairment signal processing method of the space division multiplexing system in the above-mentioned embodiment, which can solve the technical problem of nonlinear impairment signal processing of the space division multiplexing system. Compared with the prior art, the beneficial effects of the nonlinear impairment signal processing device of the space division multiplexing system provided by the present application are the same as the beneficial effects of the nonlinear impairment signal processing method of the space division multiplexing system provided by the above-mentioned embodiment, and the other technical features of the nonlinear impairment signal processing device of the space division multiplexing system are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0071] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0072] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0073] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, and the computer-readable program instructions are used to execute the nonlinear impairment signal processing method for a space division multiplexing system in the above-mentioned embodiment.

[0074] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM: Random Access Memory), a read-only memory (ROM: Read Only Memory), an erasable programmable read-only memory (EPROM: Erasable Programmable Read Only Memory or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM: CD-Read Only Memory), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency: Radio Frequency), etc., or any suitable combination of the above.

[0075] The computer-readable storage medium may be included in the nonlinear impairment signal processing device of the space division multiplexing system; or may exist independently without being assembled into the nonlinear impairment signal processing device of the space division multiplexing system.

[0076] The computer-readable storage medium carries one or more programs. When the one or more programs are executed by a nonlinear impairment signal processing device of a space-division multiplexing system, the nonlinear impairment signal processing device of the space-division multiplexing system: inputs the nonlinear impairment signal in the space-division multiplexing system into a feedforward neural network equalizer based on a cascade structure, wherein the feedforward neural network equalizer based on the cascade structure comprises an input layer, an intermediate layer and an output layer, wherein the input layer is connected to the intermediate layer through a first weight, the intermediate layer is connected to the output layer through a second weight, and the input layer is connected to the output layer through a third weight; and performs equalization processing on the nonlinear impairment signal through the first weight, the second weight and the third weight in the feedforward neural network equalizer based on the cascade structure to obtain an equalized signal.

[0077] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0078] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0079] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0080] The readable storage medium provided by the present application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned method for processing nonlinear impairment signals of a space-division multiplexing system, and can solve the technical problem. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by the present application are the same as the beneficial effects of the method for processing nonlinear impairment signals of a space-division multiplexing system provided by the above-mentioned embodiment, and will not be described in detail here.

[0081] The present application also provides a computer program product, including a computer program, which implements the steps of the above-mentioned method for processing nonlinear impairment signals of a space division multiplexing system when executed by a processor.

[0082] The computer program product provided by the present application can solve the technical problem. Compared with the prior art, the beneficial effects of the computer program product provided by the present application are the same as the beneficial effects of the nonlinear impairment signal processing method of the space division multiplexing system provided by the above embodiment, which will not be described in detail here.

[0083] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A method for processing nonlinear impairment signals in a space division multiplexing system, characterized in that: The method includes: Inputting a nonlinear impairment signal in a space division multiplexing system into a feedforward neural network equalizer based on a cascade structure, wherein the feedforward neural network equalizer based on the cascade structure comprises an input layer, an intermediate layer and an output layer, wherein the input layer is connected to the intermediate layer via a first weight, the intermediate layer is connected to the output layer via a second weight, and the input layer is connected to the output layer via a third weight; The nonlinear damage signal is equalized by using the first weight, the second weight and the third weight in the feedforward neural network equalizer based on the cascade structure to obtain an equalized signal.

2. The method according to claim 1, characterized in that The step of performing equalization processing on the nonlinear damage signal by using the first weight, the second weight and the third weight in the feedforward neural network equalizer based on the cascade structure to obtain an equalized signal comprises: Determine the output value of the middle layer according to the first weight, the threshold of the middle layer, the activation function of the middle layer and the nonlinear damage signal in the feedforward neural network equalizer based on the cascade structure; The nonlinear damage signal is equalized according to the output value of the intermediate layer, the second weight, the third weight, the threshold of the output layer and the activation function of the output layer to obtain an equalized signal.

3. The method according to claim 1, characterized in that The step of inputting the nonlinear impairment signal in the space division multiplexing system into the feedforward neural network equalizer based on the cascade structure comprises: receiving a nonlinear impairment signal of each branch in the space division multiplexing system, and setting a sliding window for the nonlinear impairment signal of each branch; The nonlinear damage signal within the sliding window range of each branch is input into the feedforward neural network equalizer based on the cascade structure in turn.

4. The method according to claim 1, characterized in that The step of inputting the nonlinear impairment signal in the space division multiplexing system into the feedforward neural network equalizer based on the cascade structure comprises: Receive the nonlinear impairment signal of each branch in the space division multiplexing system, and aggregate the nonlinear impairment signals of all branches to obtain an aggregated single channel signal; A sliding window is set for the aggregated single-channel signal, and the aggregated single-channel signal within the sliding window is input into a feedforward neural network equalizer based on a cascade structure.

5. The method according to claim 1, characterized in that After the step of performing equalization processing on the nonlinear damage signal by using the first weight, the second weight and the third weight in the feedforward neural network equalizer based on the cascade structure to obtain an equalized signal, the method further includes: Generate a time delay feedback according to the equalized signal, and add the time delay feedback to the input layer to obtain a new input layer; Connecting the time-delay feedback in the new input layer to the intermediate layer via a fourth weight; Connecting the new input layer and the output layer through the third weight to obtain a new feedforward neural network equalizer; The nonlinear damage signal is subjected to secondary equalization processing by using the first weight, the second weight, the third weight and the fourth weight in the new feedforward neural network equalizer to obtain a signal after secondary equalization.

6. The method according to claim 5, characterized in that The step of performing secondary equalization processing on the nonlinear damage signal by using the first weight, the second weight, the third weight and the fourth weight in the new feedforward neural network equalizer to obtain a secondary equalized signal includes: The formula for performing secondary equalization processing on the nonlinear damage signal by the new feedforward neural network equalizer is: ， Where y represents the signal after secondary equalization, x represents the nonlinear damage signal to be processed in the sliding window, and f [1] represents the activation function of the middle layer, f [2] represents the activation function of the output layer, W [1] represents the first weight, W [2] represents the second weight, W c represents the third weight, W d represents the fourth weight, y d represents the time delay feedback generated according to the equalized signal, b [1] represents the threshold of the middle layer, b [2] represents the threshold of the output layer, and the superscript T represents the transposed symbol.

7. A nonlinear impairment signal processing device for a space division multiplexing system, characterized in that: The nonlinear impairment signal processing device of the space division multiplexing system comprises: A signal input module, used for inputting a nonlinear impairment signal in a space division multiplexing system into a feedforward neural network equalizer based on a cascade structure, wherein the feedforward neural network equalizer based on the cascade structure comprises an input layer, an intermediate layer and an output layer, wherein the input layer is connected to the intermediate layer via a first weight, the intermediate layer is connected to the output layer via a second weight, and the input layer is connected to the output layer via a third weight; The equalization processing module is used to perform equalization processing on the nonlinear damage signal through the first weight, the second weight and the third weight in the feedforward neural network equalizer based on the cascade structure to obtain an equalized signal.

8. A nonlinear impairment signal processing device for a space division multiplexing system, characterized in that: The device comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the method for processing nonlinear impairment signals of a space division multiplexing system according to any one of claims 1 to 6.

9. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the method for processing nonlinear impairment signals of a space division multiplexing system according to any one of claims 1 to 6 are implemented.

10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the steps of the method for processing nonlinear impairment signals of a space division multiplexing system according to any one of claims 1 to 6 are implemented.

Images