Signal compensation processing method and device

Abstract

An embodiment of the present invention provides a signal compensation processing method and device, wherein the method includes: obtaining an I-channel received data signal obtained by receiving an I-channel transmitted data signal, and a Q-channel received data signal obtained by receiving a Q-channel transmitted data signal; performing cascade transformation on the I-channel received data signal and the Q-channel received data signal to obtain a data characteristic signal; compensating the data characteristic signal according to a predetermined target model parameter to obtain an IQ equalized signal, which can solve the problem of high bit error rate caused by mutual influence between IQ signals in optical communication in the related art, and compensates the data characteristic signal obtained by cascade transformation of the IQ received data signal based on the predetermined target model parameter to obtain an IQ equalized signal, thereby reducing the bit error rate of the IQ signal.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of communications, and in particular, to a signal compensation processing method and device. Background Art

[0002] IQ optical signal is a commonly used optical modulation method in communication. The two orthogonal signals are up-converted and superimposed, and then the two signals are down-converted at the receiving end to separate them. However, for access network systems, when the I and Q signals are from different sources, the superposition of the two signals will greatly deteriorate the signal-to-noise ratio due to power mismatch, clock synchronization, nonlinear effects and other problems, thereby reducing the capacity of the system.

[0003] At this time, it is necessary to use the equalization algorithms commonly used in optical communications, such as the least mean square error (SISO-LMS) [1], recursive least squares (RLS) and other algorithms to compensate for signal distortion. However, even if the nonlinear term based on the Volterra series is used, the compensation effect of LMS (Least Mean Square) and RLS is quite limited for strong nonlinear effects, especially when the I and Q paths affect each other. Therefore, deep neural networks (SISO-DNN) [2] with complex nonlinear mapping capabilities have been introduced into communication systems as post-equalizers. Existing studies have shown that DNN (Deep Neural Network) has good performance in applications such as post-equalization of signals. However, when the system nonlinearity is strong, the use of traditional deep neural networks requires a significant increase in the complexity of the system to provide better equalization performance. Therefore, this is a very unfavorable factor for communication networks with extremely high real-time requirements. Moreover, most of the neural network algorithms published so far are based on single input single output (SISO) systems, and there are few studies on equalization networks for two-way signals, such as IQ signals, that interact with each other. Therefore, developing a neural network for IQ signals in optical communications has become an urgent problem to be solved. The traditional deep network has a high bit error rate and high algorithm complexity. Summary of the invention

[0004] The embodiments of the present invention provide a signal compensation processing method and device to at least solve the problem of high bit error rate caused by mutual influence between IQ signals in optical communication in the related art.

[0005] According to an embodiment of the present invention, a signal compensation processing method is provided, comprising:

[0006] Acquire an I-channel received data signal obtained by receiving an I-channel transmitted data signal, and a Q-channel received data signal obtained by receiving a Q-channel transmitted data signal;

[0007] Performing cascade transformation on the I-channel received data signal and the Q-channel received data signal to obtain a data characteristic signal;

[0008] The data characteristic signal is compensated according to a predetermined target model parameter to obtain an IQ equalized signal.

[0009] In one embodiment, compensating the data characteristic signal according to a predetermined target model parameter to obtain the IQ equalized signal includes:

[0010] Mapping the data characteristic signal according to the target model parameter to obtain an I-way linear data signal, a Q-way linear data signal, an I-way nonlinear data signal, and a Q-way nonlinear data signal;

[0011] The sum of the I-way linear data signal and the I-way nonlinear data signal is determined as an I-way balanced signal, and the sum of the Q-way linear data signal and the Q-way nonlinear data signal is determined as a Q-way balanced signal, wherein the IQ balanced signal includes the I-way balanced signal and the Q-way balanced signal.

[0012] In one embodiment, mapping the data feature signal according to the target model parameter to obtain an I-way linear data signal, a Q-way linear data signal, an I-way nonlinear data signal, and a Q-way nonlinear data signal includes:

[0013] Processing the linear crosstalk of I-channel optical signal of the data characteristic signal according to the target model parameters to obtain the I-channel linear data signal;

[0014] Processing the linear crosstalk of the Q-path optical signal of the data characteristic signal according to the target model parameters to obtain the Q-path linear data signal;

[0015] The crosstalk between the IQ optical signals of the data characteristic signal is processed according to the target model parameters to obtain the I-channel nonlinear data signal and the Q-channel nonlinear data signal.

[0016] In one embodiment, processing the linear crosstalk of the I-channel optical signal of the data characteristic signal according to the target model parameter to obtain the I-channel linear data signal includes:

[0017] The I-way linear data signal is obtained by linearly mapping the data feature signal according to the target model parameters in the following manner: Wherein, 1≤t≤N, N is the length of the data characteristic signal; and / or

[0018] Processing the linear crosstalk of the Q-path optical signal of the data characteristic signal according to the target model parameter to obtain the Q-path linear data signal comprises:

[0019] The Q-line linear data signal is obtained by linearly mapping the data feature signal according to the target model parameters in the following manner: and / or

[0020] Processing the crosstalk between the IQ optical signals of the data characteristic signal according to the target model parameters to obtain the I-path nonlinear data signal and the Q-path nonlinear data signal comprises:

[0021] The I-channel nonlinear data signal and the Q-channel nonlinear data signal are obtained by performing nonlinear mapping on the data feature signal according to the target model parameters in the following manner:

[0022] in, is the target model parameter, and f(x) is the activation function.

[0023] In one embodiment, performing cascade transformation on the I-channel received data signal and the Q-channel received data signal to obtain the data characteristic signal includes:

[0024] The I-channel received data signal and the Q-channel received data signal are sampled respectively according to a window length of taps to obtain N groups of I-channel sampled signals and N groups of Q-channel sampled signals, wherein the center signal of each group of sampled signals in the N groups of I-channel sampled signals and the N groups of Q-channel sampled signals is the current sampling point, the (taps-1) / 2 points on the left side of the center signal are the (taps-1) / 2 sampling points intercepted forward from the current sampling point, and the (taps-1) / 2 points on the right side of the center signal are the (taps-1) / 2 sampling points intercepted backward from the current sampling point, if the forward or backward sampling points are less than (taps-1) / 2, 0 is added, the taps is an odd number, and N is the length of the I-channel received data signal and the Q-channel received data signal;

[0025] The N groups of I-channel sampling signals with a length of the taps are cascaded with the N groups of Q-channel sampling signals respectively to obtain the data characteristic signal including N signals with a length of 2 taps.

[0026] In another embodiment, before performing cascade transformation on an IQ received data signal obtained by receiving the IQ transmitted data signal to obtain a data characteristic signal, the method further includes:

[0027] Acquire an I-channel receiving modeling signal obtained by receiving an I-channel transmitting modeling signal, and a Q-channel receiving modeling signal obtained by receiving a Q-channel transmitting modeling signal;

[0028] Performing cascade transformation on the I-channel received modeling signal and the Q-channel received modeling signal to obtain a modeling characteristic signal;

[0029] The model parameters are updated according to the modeling characteristic signal to obtain the target model parameters.

[0030] In another embodiment, performing a cascade transformation on the I-channel receiving modeling signal and the Q-channel receiving modeling signal to obtain the modeling characteristic signal includes:

[0031] The I-channel receiving modeling signal and the Q-channel receiving modeling signal are sampled respectively according to the window length taps to obtain N groups of I-channel sampling signals and N groups of Q-channel sampling signals, wherein the center signal of each group of sampling signals in the N groups of I-channel sampling signals and the N groups of Q-channel sampling signals is the current sampling point, the (taps-1) / 2 points on the left side of the center signal are the (taps-1) / 2 sampling points intercepted forward from the current sampling point, and the (taps-1) / 2 points on the right side of the center signal are the (taps-1) / 2 sampling points intercepted backward from the current sampling point, if the forward or backward sampling points are less than (taps-1) / 2, 0 is added, the taps is an odd number, and N is the length of the I-channel receiving modeling signal and the Q-channel receiving modeling signal;

[0032] The N groups of I-channel sampling signals with a length of the taps are cascaded with the N groups of Q-channel sampling signals respectively to obtain the modeling characteristic signal including N signals with a length of 2 taps.

[0033] In another embodiment, updating the model parameters according to the modeling characteristic signal to obtain the target model parameters includes:

[0034] Repeat the following steps until the number of iterations t meets the preset iteration condition, and determine that the updated model parameters are the target model parameters:

[0035] When t=0, the model parameters are initialized, and the modeling characteristic signals are mapped according to the model parameters to obtain an I-path linear modeling signal, a Q-path linear modeling signal, an I-path nonlinear modeling signal, and a Q-path nonlinear modeling signal;

[0036] Performing sum processing on the I-way linear modeling signal and the I-way nonlinear modeling signal to obtain an I-way modeling output signal, and performing sum processing on the Q-way linear modeling signal and the Q-way nonlinear modeling signal to obtain a Q-way modeling output signal;

[0037] The model parameters are updated according to the I-channel transmission modeling signal, the Q-channel transmission modeling signal, the I-channel modeling output signal, and the Q-channel modeling output signal to obtain updated model parameters, wherein the IQ transmission modeling signal includes the I-channel transmission modeling signal and the Q-channel transmission modeling signal;

[0038] t=t+1.

[0039] In another embodiment, mapping the modeling characteristic signal according to the model parameter to obtain the I-path linear modeling signal, the Q-path linear modeling signal, the I-path nonlinear modeling signal, and the Q-path nonlinear modeling signal comprises:

[0040] Processing the linear crosstalk of the I-channel optical signal of the modeling characteristic signal according to the model parameters to obtain the I-channel linear modeling signal;

[0041] Processing the linear crosstalk of the Q-path optical signal of the modeling characteristic signal according to the model parameters to obtain the Q-path linear modeling signal;

[0042] The crosstalk between the I and Q optical signals of the modeling characteristic signal is processed according to the model parameters to obtain the I-path nonlinear modeling signal and the Q-path nonlinear modeling signal.

[0043] In another embodiment, processing the linear crosstalk of the I-channel optical signal of the modeling characteristic signal according to the model parameters to obtain the I-channel linear modeling signal includes:

[0044] The I-line linear modeling signal is obtained by linearly mapping the modeling characteristic signal according to the model parameters in the following manner: Wherein, 1≤t≤N, N is the length of the modeling feature; and / or

[0045] Processing the linear crosstalk of the Q-path optical signal of the modeling characteristic signal according to the model parameters to obtain the Q-path linear modeling signal includes:

[0046] The Q-line linear modeling signal is obtained by linearly mapping the modeling characteristic signal according to the model parameters in the following manner: and / or

[0047] Processing the crosstalk between the I and Q optical signals of the modeling characteristic signal according to the model parameters to obtain the I nonlinear modeling signal and the Q nonlinear modeling signal comprises:

[0048] The modeling characteristic signal is nonlinearly mapped according to the model parameters to obtain an I-channel nonlinear modeling signal and a Q-channel nonlinear modeling signal in the following manner: in, are the model parameters, and f(x) is the activation function.

[0049] In another embodiment, the model parameters are updated according to the I-channel transmission modeling signal, the Q-channel transmission modeling signal, the I-channel modeling output signal, and the Q-channel modeling output signal to obtain the updated model parameters, which includes:

[0050] Determine a signal output error according to the I-channel transmission modeling signal, the Q-channel transmission modeling signal, the I-channel modeling output signal, and the Q-channel modeling output signal;

[0051] The model parameters are updated according to the signal output error to obtain the updated model parameters.

[0052] In another embodiment, updating the model parameters according to the signal output error to obtain updated model parameters includes:

[0053] Determine the partial derivatives of the signal output error with respect to the model parameters respectively to obtain a target matrix with the same dimension as the model parameters;

[0054] The model parameters are updated according to the target matrix to obtain the updated model parameters.

[0055] In another embodiment, updating the model parameters according to the target matrix to obtain the updated model parameters includes:

[0056] The update parameters are determined by:

[0057]

[0058]

[0059]

[0060] M(t)=β1M(t-1)+(1-β1)G(t),

[0061] V(t)=β2V(t-1)+(1-β2)G 2 (t), where H(t) is the update parameter, M(t) is the momentum matrix, V(t) is the velocity matrix, G(t) is the target matrix, α=0.001, β1=0.9, β2=0.999, ε=10 -8 , M(-1)=0, V(-1)=0, 0≤t≤t max , t max is the preset iteration number threshold;

[0062] The difference between the model parameter and the update parameter is determined as the updated model parameter.

[0063] In another embodiment, the method further comprises:

[0064] The signal output error is determined according to the I-channel transmission modeling signal, the Q-channel transmission modeling signal, the I-channel modeling output signal, and the Q-channel modeling output signal in the following manner:

[0065] Where r(t) is the signal output error, p I (a) is the output signal of the I-channel modeling, p Q (a) Output signal for Q-path modeling, s I (a) is the modeling signal of the I-path transmission, s Q (a) is the Q-path transmission modeling signal, and N is the length of the modeling characteristic signal.

[0066] According to another embodiment of the present invention, there is further provided a signal compensation processing device, comprising:

[0067] A first acquisition module is used to acquire an I-channel received data signal obtained by receiving an I-channel transmitted data signal, and a Q-channel received data signal obtained by receiving a Q-channel transmitted data signal;

[0068] A first cascade module, used for performing cascade transformation on the I-channel received data signal and the Q-channel received data signal to obtain a data characteristic signal;

[0069] The compensation module is used to compensate the characteristic signal according to a predetermined target model parameter to obtain an IQ equalized signal.

[0070] In one embodiment, the compensation module includes:

[0071] A mapping submodule, used for mapping the data characteristic signal according to the target model parameter to obtain an I-way linear data signal, a Q-way linear data signal, an I-way nonlinear data signal and a Q-way nonlinear data signal;

[0072] A determination submodule is used to determine the sum of the I-way linear data signal and the I-way nonlinear data signal as an I-way balanced signal, and to determine the sum of the Q-way linear data signal and the Q-way nonlinear data signal as a Q-way balanced signal, wherein the IQ balanced signal includes the I-way balanced signal and the Q-way balanced signal.

[0073] In one embodiment, the mapping submodule includes:

[0074] A first processing unit, configured to process the linear crosstalk of the I-channel optical signal of the data characteristic signal according to the target model parameters to obtain the I-channel linear data signal;

[0075] A second processing unit, configured to process the linear crosstalk of the Q-path optical signal of the data characteristic signal according to the target model parameters to obtain the Q-path linear data signal;

[0076] The third processing unit is used to process the crosstalk between the IQ two-path optical signals of the data characteristic signal according to the target model parameters to obtain the I-path nonlinear data signal and the Q-path nonlinear data signal.

[0077] In one embodiment, the first processing unit is further configured to perform linear mapping on the data feature signal according to the target model parameter to obtain the I-line linear data signal in the following manner: Wherein, 1≤t≤N, N is the length of the data characteristic signal; and / or

[0078] The second processing unit is further configured to perform linear mapping on the data feature signal according to the target model parameter in the following manner to obtain the Q-line linear data signal: and / or

[0079] The third processing unit is further configured to perform nonlinear mapping on the data feature signal according to the target model parameter in the following manner to obtain the I-channel nonlinear data signal and the Q-channel nonlinear data signal:

[0080] in, is the target model parameter, and f(x) is the activation function.

[0081] In one embodiment, the first cascading module includes:

[0082] A first sampling submodule is used to sample the I-channel received data signal and the Q-channel received data signal according to a window length of taps, respectively, to obtain N groups of I-channel sampled signals and N groups of Q-channel sampled signals, wherein the center signal of each group of sampled signals in the N groups of I-channel sampled signals and the N groups of Q-channel sampled signals is a current sampling point, (taps-1) / 2 points on the left side of the center signal are (taps-1) / 2 sampling points intercepted forward from the current sampling point, and (taps-1) / 2 points on the right side of the center signal are (taps-1) / 2 sampling points intercepted backward from the current sampling point, if the number of forward or backward sampling points is less than (taps-1) / 2, 0 is added, taps is an odd number, and N is the length of the I-channel received data signal and the Q-channel received data signal;

[0083] The first cascade submodule is used to cascade the N groups of I-channel sampling signals with a length of the taps and the N groups of Q-channel sampling signals respectively to obtain the data characteristic signal including N signals with a length of 2 taps.

[0084] In another embodiment, the apparatus further comprises:

[0085] A second acquisition module is used to acquire an I-channel receiving modeling signal obtained by receiving an I-channel transmitting modeling signal, and a Q-channel receiving modeling signal obtained by receiving a Q-channel transmitting modeling signal;

[0086] A second cascade module is used to perform cascade transformation on the I-channel received modeling signal and the Q-channel received modeling signal to obtain a modeling characteristic signal;

[0087] An updating module is used to update the model parameters according to the modeling characteristic signal to obtain the target model parameters.

[0088] In another embodiment, the second cascade module includes:

[0089] A second sampling submodule is used to sample the I-channel receiving modeling signal and the Q-channel receiving modeling signal according to a window length of taps, respectively, to obtain N groups of I-channel sampling signals and N groups of Q-channel sampling signals, wherein the center signal of each group of sampling signals in the N groups of I-channel sampling signals and the N groups of Q-channel sampling signals is the current sampling point, the (taps-1) / 2 points on the left side of the center signal are the (taps-1) / 2 sampling points intercepted forward from the current sampling point, and the (taps-1) / 2 points on the right side of the center signal are the (taps-1) / 2 sampling points intercepted backward from the current sampling point, if the forward or backward sampling points are less than (taps-1) / 2, 0 is added, the taps is an odd number, and N is the length of the I-channel receiving modeling signal and the Q-channel receiving modeling signal;

[0090] The second cascade submodule is used to cascade the N groups of I-channel sampling signals with a length of the taps and the N groups of Q-channel sampling signals respectively to obtain the modeling characteristic signal including N signals with a length of 2 taps.

[0091] In another embodiment, the update module includes:

[0092] The execution submodule is used to repeatedly execute the following steps until the number of iterations t meets the preset iteration condition, and determine that the updated model parameters are the target model parameters:

[0093] A mapping unit, used for initializing the model parameters when t=0, and mapping the modeling characteristic signal according to the model parameters to obtain an I-path linear modeling signal, a Q-path linear modeling signal, an I-path nonlinear modeling signal, and a Q-path nonlinear modeling signal;

[0094] A summing unit, configured to perform summing processing on the I-way linear modeling signal and the I-way nonlinear modeling signal to obtain an I-way modeling output signal, and perform summing processing on the Q-way linear modeling signal and the Q-way nonlinear modeling signal to obtain a Q-way modeling output signal;

[0095] An updating unit is used to update the model parameters according to the I-channel transmission modeling signal, the Q-channel transmission modeling signal, the I-channel modeling output signal and the Q-channel modeling output signal to obtain updated model parameters, wherein the IQ transmission modeling signal includes the I-channel transmission modeling signal and the Q-channel transmission modeling signal; t=t+1.

[0096] In another embodiment, the mapping unit comprises:

[0097] A first processing subunit, configured to process the linear crosstalk of the I-channel optical signal of the modeling characteristic signal according to the model parameters to obtain the I-channel linear modeling signal;

[0098] A second processing subunit is used to process the linear crosstalk of the Q-path optical signal of the modeling characteristic signal according to the model parameters to obtain the Q-path linear modeling signal;

[0099] The third processing subunit is used to process the crosstalk between the IQ two-path optical signals of the modeling characteristic signal according to the model parameters to obtain the I-path nonlinear modeling signal and the Q-path nonlinear modeling signal.

[0100] In another embodiment, the first processing subunit is further configured to obtain the I-line linear modeling signal by linearly mapping the modeling feature signal according to the model parameters in the following manner: Wherein, 1≤t≤N, N is the length of the modeling feature; and / or

[0101] The second processing subunit is further configured to perform linear mapping on the modeling characteristic signal according to the model parameters in the following manner to obtain the Q-line linear modeling signal: and / or

[0102] The third processing subunit is further configured to perform nonlinear mapping on the modeling characteristic signal according to the model parameters in the following manner to obtain an I-channel nonlinear modeling signal and a Q-channel nonlinear modeling signal: in, are the model parameters, and f(x) is the activation function.

[0103] In another embodiment, the updating unit comprises:

[0104] a determination subunit, configured to determine a signal output error according to the I-channel transmission modeling signal, the Q-channel transmission modeling signal, the I-channel modeling output signal, and the Q-channel modeling output signal;

[0105] An updating subunit is used to update the model parameters according to the signal output error to obtain the updated model parameters.

[0106] In another embodiment, the updating subunit is further used to

[0107] Determine the partial derivatives of the signal output error with respect to the model parameters respectively to obtain a target matrix with the same dimension as the model parameters;

[0108] The model parameters are updated according to the target matrix to obtain the updated model parameters.

[0109] In another embodiment, the updating subunit is further used to

[0110] The update parameters are determined by:

[0111]

[0112]

[0113]

[0114] M(t)=β1M(t-1)+(1-β1)G(t),

[0115] V(t)=β2V(t-1)+(1-β2)G 2 (t), where H(t) is the update parameter, M(t) is the momentum matrix, V(t) is the velocity matrix, G(t) is the target matrix, α=0.001, β1=0.9, β2=0.999, ε=10 -8 , M(-1)=0, V(-1)=0, 0≤t≤t max , t max is the preset iteration number threshold;

[0116] The difference between the model parameter and the update parameter is determined as the updated model parameter.

[0117] In another embodiment, the determining subunit is further configured to determine the signal output error according to the I-channel transmission modeling signal, the Q-channel transmission modeling signal, the I-channel modeling output signal, and the Q-channel modeling output signal in the following manner:

[0118] Where r(t) is the signal output error, p I (a) is the output signal of the I-channel modeling, p Q (a) Output signal for Q-path modeling, s I (a) is the modeling signal of the I-path transmission, s Q (a) is the Q-path transmission modeling signal, and N is the length of the modeling characteristic signal.

[0119] According to yet another embodiment of the present invention, a computer-readable storage medium is provided, in which a computer program is stored, wherein the computer program is configured to execute the steps of any one of the above method embodiments when run.

[0120] According to yet another embodiment of the present invention, there is provided an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0121] Through the present invention, an I-channel received data signal obtained by receiving an I-channel transmitted data signal and a Q-channel received data signal obtained by receiving a Q-channel transmitted data signal are obtained; the I-channel received data signal and the Q-channel received data signal are cascade-transformed to obtain a characteristic signal; the data characteristic signal is compensated according to a predetermined target model parameter to obtain an IQ balanced signal, which can solve the problem of high bit error rate caused by mutual influence between IQ signals in optical communication in the related art, and the data characteristic signal obtained by cascade transformation of the IQ received data signal is compensated based on the predetermined target model parameter to obtain an IQ balanced signal, thereby reducing the bit error rate of the IQ signal. BRIEF DESCRIPTION OF THE DRAWINGS

[0122] Figure 1 is a hardware structure block diagram of a mobile terminal of the signal compensation processing method according to an embodiment of the present invention;

[0123] Figure 2 is a flow chart of a signal compensation processing method according to an embodiment of the present invention;

[0124] Figure 3 is a flow chart of a signal compensation processing method according to an embodiment of the present invention;

[0125] Figure 4 Flow chart of a method for receiving and compensating an IQ signal according to an embodiment of the present invention;

[0126] Figure 5 is a schematic diagram of a multi-input multi-output multi-branch neural network equalizer for a heterogeneous deep neural network according to an embodiment of the present invention;

[0127] Figure 6 is a schematic diagram of a verification experiment platform according to an embodiment of the present invention;

[0128] Figure 7 is a structural block diagram of a signal compensation processing device according to an embodiment of the present invention;

[0129] Figure 8 is a structural block diagram of a signal compensation processing device according to a preferred embodiment of the present invention. DETAILED DESCRIPTION

[0130] Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings and in combination with the embodiments.

[0131] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0132] The method embodiments provided in the embodiments of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Taking running on a mobile terminal as an example, Figure 1 is a hardware structure block diagram of a mobile terminal of a signal compensation processing method according to an embodiment of the present invention, such as Figure 1 As shown, the mobile terminal may include one or more ( Figure 1 Only one is shown in the figure) a processor 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data, wherein the mobile terminal may also include a transmission device 106 and an input / output device 108 for communication functions. It can be understood by those skilled in the art that Figure 1 The structure shown is for illustration only and does not limit the structure of the mobile terminal. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0133] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the data processing method in the embodiment of the present invention. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, to implement the above method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the mobile terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0134] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the mobile terminal. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0135] In this embodiment, a signal compensation processing method running on the above mobile terminal or network architecture is provided. Figure 2 is a flow chart of a signal compensation processing method according to an embodiment of the present invention. Figure 2 As shown, the process includes the following steps:

[0136] Step S202, obtaining an I-channel received data signal obtained by receiving an I-channel transmitted data signal, and a Q-channel received data signal obtained by receiving a Q-channel transmitted data signal;

[0137] Step S204, performing cascade transformation on the I-channel received data signal and the Q-channel received data signal to obtain a data characteristic signal;

[0138] In one embodiment, performing cascade transformation on the I-channel received data signal and the Q-channel received data signal to obtain the data characteristic signal includes:

[0139] In this embodiment, the above step S204 may specifically include: sampling the I-channel received data signal and the Q-channel received data signal according to a window length of taps, respectively, to obtain N groups of I-channel sampled signals and N groups of Q-channel sampled signals, wherein the center signal of each group of sampled signals in the N groups of I-channel sampled signals and the N groups of Q-channel sampled signals is the current sampling point, (taps-1) / 2 points on the left side of the center signal are (taps-1) / 2 sampling points intercepted forward from the current sampling point, and (taps-1) / 2 points on the right side of the center signal are (taps-1) / 2 sampling points intercepted backward from the current sampling point, if the number of forward or backward sampling points is less than (taps-1) / 2, 0 is added, taps is an odd number, and N is the length of the I-channel received data signal and the Q-channel received data signal; and cascading the N groups of I-channel sampled signals and the N groups of Q-channel sampled signals of length taps, respectively, to obtain the data characteristic signal including N signals of length 2taps.

[0140] Step S206: Compensate the data characteristic signal according to a predetermined target model parameter to obtain an IQ equalized signal.

[0141] Through the above steps S202 to S206, an I-channel received data signal obtained by receiving an I-channel transmitted data signal and a Q-channel received data signal obtained by receiving a Q-channel transmitted data signal are obtained; the I-channel received data signal and the Q-channel received data signal are cascaded transformed to obtain a characteristic signal; the data characteristic signal is compensated according to a predetermined target model parameter to obtain an IQ balanced signal, which can solve the problem of high bit error rate caused by mutual influence between IQ signals in optical communication in the related art. The data characteristic signal obtained by cascade transformation of the IQ received data signal is compensated based on the predetermined target model parameter to obtain an IQ balanced signal, thereby reducing the bit error rate of the IQ signal.

[0142] In one embodiment, the above step S206 may specifically include:

[0143] S2061, mapping the data feature signal according to the target model parameter to obtain an I-way linear data signal, a Q-way linear data signal, an I-way nonlinear data signal, and a Q-way nonlinear data signal;

[0144] In one embodiment, the above step S2061 may specifically include:

[0145] The linear crosstalk of the I-channel optical signal of the data characteristic signal is processed according to the target model parameters to obtain the I-channel linear data signal, and the I-channel linear data signal is obtained by linearly mapping the data characteristic signal according to the target model parameters in the following manner: Wherein, 1≤t≤N, N is the length of the data characteristic signal;

[0146] The linear crosstalk of the Q-path optical signal of the data characteristic signal is processed according to the target model parameters to obtain the Q-path linear data signal, and the Q-path linear data signal is obtained by linearly mapping the data characteristic signal according to the target model parameters in the following manner:

[0147] The crosstalk between the IQ optical signals of the data characteristic signal is processed according to the target model parameters to obtain the I nonlinear data signal and the Q nonlinear data signal, specifically by performing nonlinear mapping on the data characteristic signal according to the target model parameters to obtain the I nonlinear data signal and the Q nonlinear data signal in the following manner: in, is the target model parameter, f(x) is the activation function;

[0148] S2062, determining the sum of the I-way linear data signal and the I-way nonlinear data signal as an I-way balanced signal, and determining the sum of the Q-way linear data signal and the Q-way nonlinear data signal as a Q-way balanced signal, wherein the IQ balanced signal includes the I-way balanced signal and the Q-way balanced signal.

[0149] Figure 3 is a flow chart of a signal compensation processing method according to this preferred embodiment, such as Figure 3 As shown, before performing cascade transformation on the IQ received data signal obtained by receiving the IQ transmitted data signal to obtain the data characteristic signal, the method further includes:

[0150] Step S302, obtaining an I-channel receiving modeling signal obtained by receiving an I-channel transmitting modeling signal, and a Q-channel receiving modeling signal obtained by receiving a Q-channel transmitting modeling signal;

[0151] Step S304, performing cascade transformation on the I-channel received modeling signal and the Q-channel received modeling signal to obtain a modeling characteristic signal;

[0152] The above step S304 may specifically include: sampling the I-channel receiving modeling signal and the Q-channel receiving modeling signal according to the window length taps, respectively, to obtain N groups of I-channel sampling signals and N groups of Q-channel sampling signals, wherein the center signal of each group of sampling signals in the N groups of I-channel sampling signals and the N groups of Q-channel sampling signals is the current sampling point, (taps-1) / 2 points on the left side of the center signal are (taps-1) / 2 sampling points intercepted forward from the current sampling point, and (taps-1) / 2 points on the right side of the center signal are (taps-1) / 2 sampling points intercepted backward from the current sampling point, if the forward or backward sampling points are less than (taps-1) / 2, 0 is added, the taps is an odd number, and N is the length of the I-channel receiving modeling signal and the Q-channel receiving modeling signal; and the N groups of I-channel sampling signals and the N groups of Q-channel sampling signals of the length of the taps are cascaded respectively to obtain the modeling characteristic signal including N signals of the length of 2taps.

[0153] Step S306: Update the model parameters according to the modeling characteristic signal to obtain the target model parameters.

[0154] In this embodiment, the above step S306 may specifically include:

[0155] Repeat the following steps until the number of iterations t meets the preset iteration condition, and determine that the updated model parameters are the target model parameters:

[0156] When t=0, the model parameters are initialized, and the modeling characteristic signals are mapped according to the model parameters to obtain an I-way linear modeling signal, a Q-way linear modeling signal, an I-way nonlinear modeling signal, and a Q-way nonlinear modeling signal; further, the linear crosstalk of the I-way optical signal of the modeling characteristic signal is processed according to the model parameters to obtain the I-way linear modeling signal. Specifically, the modeling characteristic signal can be linearly mapped according to the model parameters in the following manner to obtain the I-way linear modeling signal: Wherein, 1≤t≤N, N is the length of the modeling feature; the linear crosstalk of the Q-path optical signal of the modeling feature signal is processed according to the model parameters to obtain the Q-path linear modeling signal. Specifically, the Q-path linear modeling signal can be obtained by linearly mapping the modeling feature signal according to the model parameters in the following manner: The crosstalk between the IQ two-path optical signal of the modeling characteristic signal is processed according to the model parameters to obtain the I-path nonlinear modeling signal and the Q-path nonlinear modeling signal. Specifically, the I-path nonlinear modeling signal and the Q-path nonlinear modeling signal can be obtained by performing nonlinear mapping on the modeling characteristic signal according to the model parameters in the following manner: in, are the model parameters, f(x) is the activation function;

[0157] Performing sum processing on the I-way linear modeling signal and the I-way nonlinear modeling signal to obtain an I-way modeling output signal, and performing sum processing on the Q-way linear modeling signal and the Q-way nonlinear modeling signal to obtain a Q-way modeling output signal;

[0158] The model parameters are updated according to the I-channel transmission modeling signal, the Q-channel transmission modeling signal, the I-channel modeling output signal, and the Q-channel modeling output signal to obtain updated model parameters, wherein the IQ transmission modeling signal includes the I-channel transmission modeling signal and the Q-channel transmission modeling signal;

[0159] t=t+1.

[0160] In this embodiment, the processing process is divided into two stages: model building or training and signal reception compensation. The switching between the two stages can be indicated by a flag bit. For example, when the flag bit is 0 or false, model building or training is executed, and when the flag bit is 1 or true, signal reception compensation is executed. The above-mentioned preset iteration conditions can be implemented by the flag bit or by other means, such as a fixed-length training sequence, parameter convergence to a certain error, etc.

[0161] In one embodiment, the updating of the model parameters to obtain the updated model parameters may specifically include:

[0162] The signal output error is determined according to the I-channel transmission modeling signal, the Q-channel transmission modeling signal, the I-channel modeling output signal, and the Q-channel modeling output signal. Specifically, the signal output error can be determined according to the I-channel transmission modeling signal, the Q-channel transmission modeling signal, the I-channel modeling output signal, and the Q-channel modeling output signal in the following manner: Where r(t) is the signal output error, p I (a) is the output signal of the I-channel modeling, p Q (a) Output signal for Q-path modeling, s I (a) is the modeling signal of the I-path transmission, s Q (a) is the Q-path transmission modeling signal, and N is the length of the modeling characteristic signal;

[0163] The model parameters are updated according to the signal output error to obtain the updated model parameters. Further, the partial derivatives of the signal output error with respect to the model parameters are determined respectively to obtain a target matrix with the same dimension as the model parameters. The model parameters are updated according to the target matrix to obtain the updated model parameters. Specifically, the update parameters are determined in the following manner:

[0164]

[0165]

[0166]

[0167] M(t)=β1M(t-1)+(1-β1)G(t),

[0168] V(t)=β2V(t-1)+(1-β2)G 2 (t), where H(t) is the update parameter, M(t) is the momentum matrix, V(t) is the velocity matrix, G(t) is the target matrix, α=0.001, β1=0.9, β2=0.999, ε=10 -8 , M(-1)=0, V(-1)=0, 0≤t≤t max , t max The preset iteration number threshold is used, and the difference between the model parameter and the update parameter is determined as the updated model parameter.

[0169] The input signal of this embodiment has memory, and several symbols before and after are recorded for joint compensation and estimation. Three branches are used, one branch processes the linear crosstalk of the I optical signal, and one branch processes the linear crosstalk of the Q optical signal. The two linear branches can be linear weighted networks of MLP; and there is also a branch to process the crosstalk between the IQ optical signals. Since the mutual crosstalk between the IQ optical signals is not very complicated, a two-layer hidden layer network can be used to train nonlinearity, which greatly reduces the complexity of the system. At the output end, the linear output and nonlinear output of the I route, and the linear output and nonlinear output of the Q route are added respectively. The obtained compensation signal balances both the linear distortion and the nonlinear distortion, achieving a good effect. This method is also applicable to wireless signals with crosstalk in the IQ two-way with the same or similar architecture, and will not be repeated here. By using this embodiment, the performance of the system can be effectively improved. The equalization method proposed in this embodiment can increase the working range by 2.33 times to 8 times or more under the condition of two transmitting end powers. At the same time, a maximum increase of 3.63dB Q value can be achieved.

[0170] The present embodiment is described in detail below by taking a flag bit as an example to indicate switching between model building and signal compensation.

[0171] Figure 4 Flow chart of the IQ signal receiving compensation method according to an embodiment of the present invention, as shown in FIG. Figure 4 As shown, the following steps may be included:

[0172] Step 401, initialize the model parameters and initialize the flag bit to flag=False, specifically:

[0173] Set the number of hidden layers of the model to 2, the number of nodes in the first hidden layer to H1, the number of nodes in the second hidden layer to H2, the activation function f(x), the input layer I and Q signal windows to taps respectively, the number of nodes in each hidden layer, and the weight matrix and the bias vector The initial value of is a random matrix or vector that conforms to the zero-mean Gaussian distribution, and the number of iterations epoch = 0 and the iteration termination value epoch_max are initialized, and the flag flag = False. Among them, the dimensions of the weight matrix and the bias vector are shown in Table 1.

[0174] Table 1

[0175]

[0176] Step 402, determine whether flag = true, if flag = False, go to step 403; if flag = True, go to step 414.

[0177] Step 403, receiving and recording I-channel receiving modeling signals, wherein the transmitting end loads the original transmitting modeling signal s(t), specifically by means of a transmitting modeling signal loading device. Modulating the original transmitting modeling signal and recording it as an I-channel transmitting modeling signal, specifically by means of an I-channel modeling signal transmitting device modulating s(t) and recording the I-channel transmitting modeling signal x I (t). Modulate the original transmission modeling signal and record it as the Q-path transmission modeling signal. Specifically, modulate s(t) through the Q-path modeling signal transmitting device and record the Q-path transmission modeling signal x Q (t).

[0178] Specifically, the I-channel modeling signal receiving device receives and records x I (t), and obtain the I-channel receiving modeling signal y I (t).

[0179] Step 404, receiving and recording the Q-channel receiving modeling signal. Specifically:

[0180] Receive and record x through Q-path modeling signal receiving device I (t), and obtain the Q-path receiving modeling signal y I(t).

[0181] Step 405 , based on the recorded I-channel receiving modeling signal and Q-channel receiving modeling signal, cascade transformation is performed and recorded as a modeling characteristic signal.

[0182] Specifically, by modeling the cascade transformation device, y I (t), y I (t) is cascaded and transformed into the modeling feature signal k(t). The I-channel received data signal and the Q-channel received data signal are sampled respectively according to the window length taps to obtain N groups of I-channel sampled signals and N groups of Q-channel sampled signals, wherein the center signal of each group of sampled signals in the N groups of I-channel sampled signals and the N groups of Q-channel sampled signals is the current sampling point, the (taps-1) / 2 points on the left side of the center signal are the (taps-1) / 2 sampling points intercepted forward from the current sampling point, and the (taps-1) / 2 points on the right side of the center signal are the (taps-1) / 2 sampling points intercepted backward from the current sampling point. If the number of forward or backward sampling points is less than (taps-1) / 2, 0 is added. The taps is an odd number, and N is the length of the I-channel received data signal and the Q-channel received data signal. The N groups of I-channel sampled signals and the N groups of Q-channel sampled signals with a length of the taps are cascaded respectively to obtain the data feature signal including N signals with a length of 2taps. The specific transformation method is:

[0183]

[0184]

[0185]

[0186]

[0187]

[0188]

[0189]

[0190]

[0191]

[0192]

[0193]

[0194] Here, each k(t) is a column vector with a length of 2*taps. After this transformation, the signal has memory, that is, each signal input to the next level includes the previous and next taps / 2-1 signals. The modules that use this signal as output have memory.

[0195] Step 406, linear mapping is performed on the modeling feature signal to obtain an I-line linear modeling signal. Specifically:

[0196] The I-line linear modeling signal obtained by the I-line linear mapping module and k(t) is:

[0197]

[0198] Step 407: linearly map the modeling characteristic signal to obtain a Q-line linear modeling signal. Specifically:

[0199] The Q-line linear modeling signal obtained by the Q-line linear mapping module and k(t) is:

[0200]

[0201] Step 408, nonlinear mapping is performed on the modeling characteristic signal to obtain an I-channel nonlinear modeling signal and a Q-channel nonlinear modeling signal. Specifically:

[0202] Through the nonlinear mapping module and k(t), the I-channel nonlinear modeling signal and the Q-channel nonlinear modeling signal are obtained:

[0203]

[0204] Among them, f(a) is the activation function, which can be expressed as

[0205] Here x can be a scalar, vector, or matrix, and the dimension of f(x) is the same as x. When x is a matrix or vector, f(x) means applying the above formula to each element in the matrix or vector.

[0206] Step 409: add and record the I-channel modeling output signal according to the recorded I-channel linear modeling signal and I-channel nonlinear modeling signal.

[0207] Specifically, I-way linear modeling signal and I-way nonlinear modeling signal are added by I-way summing module to obtain I-way modeling output signal p I (t) = l I (t)+n I (t),1≤t≤N.

[0208] Step 410: Add and record the Q-path modeling output signal based on the recorded Q-path linear modeling signal and Q-path nonlinear modeling signal.

[0209] Specifically, the Q-path linear modeling signal and the Q-path nonlinear modeling signal are added together by the Q-path summing module to obtain the Q-path modeling output signal p Q (t) = l Q (t)+n Q (t),1≤t≤N.

[0210] Step 411, based on the recorded I-channel modeling output signal, Q-channel modeling output signal, I-channel transmission modeling signal, Q-channel transmission modeling signal, calculate the output error and adjust the model parameters to update.

[0211] Specifically, the model parameters are updated through the error calculation and model parameter updating module, and the specific steps are as follows:

[0212] (1) Initialization parameters α = 0.001, β1 = 0.9, β2 = 0.999, ε = 10 -8 , momentum matrix M(-1) = 0, velocity matrix V(-1) = 0, t = 0, set the maximum number of iterations to t max .

[0213] (2) Update all model parameters in step (1) as follows. Take this as an example to illustrate.

[0214] make

[0215] Calculate r(t) for The partial derivative of Matrix G(t) of the same dimension;

[0216] M(t)=β1M(t-1)+(1-β1)G(t);

[0217] V(t)=β2V(t-1)+(1-β2)G 2 (t), where G 2 (t) represents the dot product of the elements of the matrix G(t);

[0218]

[0219]

[0220] renew

[0221] If t = t max , then go to step 411(3), otherwise repeat the above steps at t=t+1.

[0222] (3) If all model parameters have been updated, go to step 412; otherwise, go to step (1) in step 411.

[0223] Step 412, determine whether the iteration condition is terminated. If the determination result is yes, execute step 413, otherwise, increase the number of iterations by 1 and return to steps 406, 407, and 408.

[0224] Step 413, set the flag flag = True, record the model parameters, and go to step 402.

[0225] If the number of iterations reaches the maximum number of iterations epoch_max, the model parameters are recorded through the equalizer They are

[0226] Step 414, receiving and recording I-channel received data signals. The transmitting end loads the original transmitted data signal s'(t), specifically by using a transmitted data signal loading device to load the original transmitted data signal s'(t). The original transmitted data signal is modulated and recorded as I-channel transmitted data signals, specifically by using an I-channel data signal transmitting device to modulate s'(t) and record I-channel transmitted modeling signal x I '(t). Modulate the original transmission data signal and record it as the Q-channel transmission data signal. Specifically, the Q-channel data signal transmitting device modulates s'(t) and records the I-channel transmission modeling signal x Q '(t).

[0227] Specifically, the data signal receiving device receives and records x I '(t), get the I-way receiving modeling signal y I '(t).

[0228] Step 415, receiving and recording the Q-channel received data signal. Specifically:

[0229] Receive and record x through Q-channel data signal receiving device Q '(t), get the I-way receiving modeling signal y Q '(t).

[0230] Step 416, according to the recorded I-channel received data signal and Q-channel received data signal, cascade transform and record as data characteristic signal. Specifically, y I '(t),y Q '(t) is cascade-transformed into a data characteristic signal k'(t). The specific transformation method is similar to step 405 and will not be repeated here.

[0231] Step 417: equalize and record the I-channel equalized data signal and the Q-channel equalized data signal q according to the recorded model parameters and data characteristic signal. I '(t) and q Q Specifically, the data characteristic signal is equalized by the equalization device to obtain an I-way equalized data signal and a Q-way equalized data signal q I '(t) and q Q '(t), the specific method is:

[0232]

[0233]

[0234]

[0235]

[0236] Based on the balancing method provided in this embodiment and the multi-input multi-output multi-branch neural network equalizer of the heterogeneous deep neural network, Figure 5 is a schematic diagram of a multi-input multi-output multi-branch neural network equalizer for a heterogeneous deep neural network according to an embodiment of the present invention, such as Figure 5 As shown, the output end adds the linear and nonlinear outputs of the I path and the linear and nonlinear outputs of the Q path respectively through two summing modules (I path summing module 211, Q path summing module 212) and then outputs them. When the modeling process is completed, the storage parameters are no longer updated. At this time, the multi-input multi-output multi-branch neural network is transformed into an equalizing device (220) to equalize the characteristic signal in step 417 to obtain I path and Q path equalized signals. Verification experiments were carried out in a single-receiver IQ optical communication system for a traditional post-equalizer minimum mean square error equalizer (SISO-LMS), a deep neural network (SISO-DNN) equalizer, and a heterogeneous deep neural network equalizer proposed in this embodiment. Figure 6 is a schematic diagram of a verification experiment platform according to an embodiment of the present invention, such as Figure 6As shown, first, the transmission modeling signal loading device (202) loads the original transmission modeling signal, and generates the I-path and Q-path transmission modeling signals through the I-path and Q-path modeling signal transmitting devices (203, 204). After the I-path and Q-path signals are superimposed at the receiving end, they are received by the I-path and Q-path modeling signal receiving devices (205, 206) to obtain the I-path and Q-path receiving modeling signals. Subsequently, the modeling cascade transformation device (207) in the multi-input multi-output multi-branch neural network performs the modeling cascade transformation of step 402 on the I-path and Q-path receiving modeling signals to obtain the modeling characteristic signals. Then, the modeling characteristic signals are transformed by the linear mapping modules (208, 209) and the nonlinear mapping module (210) of the I-path and Q-path to obtain the I-path and Q-path linear modeling signals and the nonlinear modeling signals, and then the I-path and Q-path summing modules are used to obtain the I-path modeling output signal and the Q-path modeling output signal. Then, the parameters in the model are updated by the algorithm in step 412 through the I and Q path transmission modeling signals and the I and Q path modeling output signals until the maximum number of iterations is reached. Then, the model parameters are fixed and the current model parameters are loaded by the equalizer (220). Then, from Figure 6 The above process is repeated until the I-channel and Q-channel data signals are received and recorded by the I-channel and Q-channel data signal devices (217, 218), and then the signals are cascaded and equalized by the data cascade conversion device (219) and the equalization device (220), and finally the I-channel and Q-channel equalized data signals are obtained.

[0237] The initialization parameters of the above equalizer are shown in Table 2.

[0238] Table 2

[0239]

[0240] In the table, '--' indicates that the structure cannot be implemented or is not adopted. It can be seen that compared with the traditional SISO-DNN equalization method, the equalization method proposed in this embodiment has fewer parameters, so the time and space required for training are smaller. Although SISO-LMS has the fewest parameters among the three, its performance is the worst.

[0241] According to the equalizer with parameter settings in Table 2, experiments were carried out on a single-receiver IQ optical communication platform. When the peak-to-peak value of the I-path signal is 350mV, the Q_Factor result graph of the three equalization methods measured by changing the peak-to-peak value of the Q-path signal. The larger the Q_Factor value, the better the system performance. The experimental results show that when the peak-to-peak value of the Q-path signal is 700mV, the working range of 510mV of the MIMO-MBNN implemented based on the equalization method proposed in the present invention is significantly better than the system working range of 170mV using the traditional equalization algorithms SISO-LMS and SISO-DNN. At the same time, the receiving equalization method proposed in this embodiment can bring a Q_Factor value performance gain of up to 3.63dB to the system compared with the traditional SISO-LMS and SISO-DNN. In the embodiment of the present invention, after the model is trained, the transformed received signal only needs to be input into the equalization device to obtain the equalized signal without any other additional operations.

[0242] Through this embodiment, the problem of poor performance and weak anti-power mismatch capability of the existing equalization method in the IQ optical communication system is effectively solved. Compared with the traditional equalization method, the communication system capacity and system performance are effectively improved. This embodiment is also relatively easy to implement. Once the model is trained, as long as the signal strength and system state are not changed, the equalization device can be used to equalize the received signal without additional operation. When the modeling is completed, the I-channel received data and the Q-channel received data are brought into the equalization device after cascade transformation to obtain the equalized signal, and the equalized signal is used for demodulation to improve the performance of the system.

[0243] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for enabling a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0244] In the present embodiment, a signal compensation processing device is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made are omitted. As used below, the term "module" can implement a combination of software and / or hardware of a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0245] Figure 7 is a structural block diagram of a signal compensation processing device according to an embodiment of the present invention. Figure 7 As shown, including:

[0246] A first acquisition module 72 is used to acquire an I-channel received data signal obtained by receiving an I-channel transmitted data signal, and a Q-channel received data signal obtained by receiving a Q-channel transmitted data signal;

[0247] A first cascade module 74, configured to perform cascade transformation on the I-channel received data signal and the Q-channel received data signal to obtain a data characteristic signal;

[0248] The compensation module 76 is used to compensate the characteristic signal according to a predetermined target model parameter to obtain an IQ equalized signal.

[0249] In one embodiment, the compensation module 76 includes:

[0250] A mapping submodule, used for mapping the data characteristic signal according to the target model parameter to obtain an I-way linear data signal, a Q-way linear data signal, an I-way nonlinear data signal and a Q-way nonlinear data signal;

[0251] A determination submodule is used to determine the sum of the I-way linear data signal and the I-way nonlinear data signal as an I-way balanced signal, and to determine the sum of the Q-way linear data signal and the Q-way nonlinear data signal as a Q-way balanced signal, wherein the IQ balanced signal includes the I-way balanced signal and the Q-way balanced signal.

[0252] In one embodiment, the mapping submodule includes:

[0253] A first processing unit, configured to process the linear crosstalk of the I-channel optical signal of the data characteristic signal according to the target model parameters to obtain the I-channel linear data signal;

[0254] A second processing unit, configured to process the linear crosstalk of the Q-path optical signal of the data characteristic signal according to the target model parameters to obtain the Q-path linear data signal;

[0255] The third processing unit is used to process the crosstalk between the IQ two-path optical signals of the data characteristic signal according to the target model parameters to obtain the I-path nonlinear data signal and the Q-path nonlinear data signal.

[0256] In one embodiment, the first processing unit is further configured to perform linear mapping on the data feature signal according to the target model parameter to obtain the I-line linear data signal in the following manner: Wherein, 1≤t≤N, N is the length of the data characteristic signal; and / or

[0257] The second processing unit is further configured to perform linear mapping on the data feature signal according to the target model parameter in the following manner to obtain the Q-line linear data signal: and / or

[0258] The third processing unit is further configured to perform nonlinear mapping on the data feature signal according to the target model parameter in the following manner to obtain the I-channel nonlinear data signal and the Q-channel nonlinear data signal:

[0259] in, is the target model parameter, and f(x) is the activation function.

[0260] In one embodiment, the first cascading module includes:

[0261] A first sampling submodule is used to sample the I-channel received data signal and the Q-channel received data signal according to a window length of taps, respectively, to obtain N groups of I-channel sampled signals and N groups of Q-channel sampled signals, wherein the center signal of each group of sampled signals in the N groups of I-channel sampled signals and the N groups of Q-channel sampled signals is a current sampling point, (taps-1) / 2 points on the left side of the center signal are (taps-1) / 2 sampling points intercepted forward from the current sampling point, and (taps-1) / 2 points on the right side of the center signal are (taps-1) / 2 sampling points intercepted backward from the current sampling point, if the number of forward or backward sampling points is less than (taps-1) / 2, 0 is added, taps is an odd number, and N is the length of the I-channel received data signal and the Q-channel received data signal;

[0262] The first cascade submodule is used to cascade the N groups of I-channel sampling signals with a length of the taps and the N groups of Q-channel sampling signals respectively to obtain the data characteristic signal including N signals with a length of 2 taps.

[0263] Figure 8 is a structural block diagram of a signal compensation processing device according to a preferred embodiment of the present invention. Figure 8 As shown, the device also includes:

[0264] The second acquisition module 82 is used to acquire an I-channel receiving modeling signal obtained by receiving an I-channel transmitting modeling signal, and a Q-channel receiving modeling signal obtained by receiving a Q-channel transmitting modeling signal;

[0265] A second cascade module 84 is used to perform cascade transformation on the I-channel received modeling signal and the Q-channel received modeling signal to obtain a modeling characteristic signal;

[0266] The updating module 86 is used to update the model parameters according to the modeling characteristic signal to obtain the target model parameters.

[0267] In another embodiment, the second cascade module 84 includes:

[0268] A second sampling submodule is used to sample the I-channel receiving modeling signal and the Q-channel receiving modeling signal according to a window length of taps, respectively, to obtain N groups of I-channel sampling signals and N groups of Q-channel sampling signals, wherein the center signal of each group of sampling signals in the N groups of I-channel sampling signals and the N groups of Q-channel sampling signals is the current sampling point, the (taps-1) / 2 points on the left side of the center signal are the (taps-1) / 2 sampling points intercepted forward from the current sampling point, and the (taps-1) / 2 points on the right side of the center signal are the (taps-1) / 2 sampling points intercepted backward from the current sampling point, if the forward or backward sampling points are less than (taps-1) / 2, 0 is added, the taps is an odd number, and N is the length of the I-channel receiving modeling signal and the Q-channel receiving modeling signal;

[0269] The second cascade submodule is used to cascade the N groups of I-channel sampling signals with a length of the taps and the N groups of Q-channel sampling signals respectively to obtain the modeling characteristic signal including N signals with a length of 2 taps.

[0270] In another embodiment, the update module 86 includes:

[0271] The execution submodule is used to repeatedly execute the following steps until the number of iterations t meets the preset iteration condition, and determine that the updated model parameters are the target model parameters:

[0272] A mapping unit, used for initializing the model parameters when t=0, and mapping the modeling characteristic signal according to the model parameters to obtain an I-path linear modeling signal, a Q-path linear modeling signal, an I-path nonlinear modeling signal, and a Q-path nonlinear modeling signal;

[0273] A summing unit, configured to perform summing processing on the I-way linear modeling signal and the I-way nonlinear modeling signal to obtain an I-way modeling output signal, and perform summing processing on the Q-way linear modeling signal and the Q-way nonlinear modeling signal to obtain a Q-way modeling output signal;

[0274] An updating unit is used to update the model parameters according to the I-channel transmission modeling signal, the Q-channel transmission modeling signal, the I-channel modeling output signal and the Q-channel modeling output signal to obtain updated model parameters, wherein the IQ transmission modeling signal includes the I-channel transmission modeling signal and the Q-channel transmission modeling signal; t=t+1.

[0275] In another embodiment, the mapping unit comprises:

[0276] A first processing subunit, configured to process the linear crosstalk of the I-channel optical signal of the modeling characteristic signal according to the model parameters to obtain the I-channel linear modeling signal;

[0277] A second processing subunit is used to process the linear crosstalk of the Q-path optical signal of the modeling characteristic signal according to the model parameters to obtain the Q-path linear modeling signal;

[0278] The third processing subunit is used to process the crosstalk between the IQ two-path optical signals of the modeling characteristic signal according to the model parameters to obtain the I-path nonlinear modeling signal and the Q-path nonlinear modeling signal.

[0279] In another embodiment, the first processing subunit is further configured to obtain the I-line linear modeling signal by linearly mapping the modeling feature signal according to the model parameters in the following manner: Wherein, 1≤t≤N, N is the length of the modeling feature; and / or

[0280] The second processing subunit is further configured to perform linear mapping on the modeling characteristic signal according to the model parameters in the following manner to obtain the Q-line linear modeling signal: and / or

[0281] The third processing subunit is further configured to perform nonlinear mapping on the modeling characteristic signal according to the model parameters in the following manner to obtain an I-channel nonlinear modeling signal and a Q-channel nonlinear modeling signal: in, are the model parameters, and f(x) is the activation function.

[0282] In another embodiment, the updating unit comprises:

[0283] a determination subunit, configured to determine a signal output error according to the I-channel transmission modeling signal, the Q-channel transmission modeling signal, the I-channel modeling output signal, and the Q-channel modeling output signal;

[0284] An updating subunit is used to update the model parameters according to the signal output error to obtain the updated model parameters.

[0285] In another embodiment, the updating subunit is further used to

[0286] Determine the partial derivatives of the signal output error with respect to the model parameters respectively to obtain a target matrix with the same dimension as the model parameters;

[0287] The model parameters are updated according to the target matrix to obtain the updated model parameters.

[0288] In another embodiment, the updating subunit is further used to

[0289] The update parameters are determined by:

[0290]

[0291]

[0292]

[0293] M(t)=β1M(t-1)+(1-β1)G(t),

[0294] V(t)=β2V(t-1)+(1-β2)G 2 (t), where H(t) is the update parameter, M(t) is the momentum matrix, V(t) is the velocity matrix, G(t) is the target matrix, α=0.001, β1=0.9, β2=0.999, ε=10 -8 , M(-1)=0, V(-1)=0, 0≤t≤t max , t max is the preset iteration number threshold;

[0295] The difference between the model parameter and the update parameter is determined as the updated model parameter.

[0296] In another embodiment, the determining subunit is further configured to determine the signal output error according to the I-channel transmission modeling signal, the Q-channel transmission modeling signal, the I-channel modeling output signal, and the Q-channel modeling output signal in the following manner:

[0297] Where r(t) is the signal output error, p I (a) is the output signal of the I-channel modeling, p Q (a) Output signal for Q-path modeling, s I (a) is the modeling signal of the I-path transmission, s Q (a) is the Q-path transmission modeling signal, and N is the length of the modeling characteristic signal.

[0298] It should be noted that the above modules can be implemented by software or hardware. For the latter, it can be implemented in the following ways, but not limited to: the above modules are all located in the same processor; or the above modules are located in different processors in any combination.

[0299] An embodiment of the present invention further provides a computer-readable storage medium, in which a computer program is stored, wherein the computer program is configured to execute the steps of any of the above method embodiments when running.

[0300] In an exemplary embodiment, the computer-readable storage medium may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk.

[0301] An embodiment of the present invention further provides an electronic device, including a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0302] In an exemplary embodiment, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.

[0303] For specific examples in this embodiment, reference may be made to the examples described in the above embodiments and exemplary implementation modes, and this embodiment will not be described in detail herein.

[0304] Obviously, those skilled in the art should understand that the above modules or steps of the present invention can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order than here, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present invention is not limited to any specific combination of hardware and software.

[0305] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A signal compensation processing method, characterized in that: include: Acquire an I-channel received data signal obtained by receiving an I-channel transmitted data signal, and a Q-channel received data signal obtained by receiving a Q-channel transmitted data signal; Performing cascade transformation on the I-channel received data signal and the Q-channel received data signal to obtain a data characteristic signal; The data characteristic signal is compensated according to a predetermined target model parameter to obtain an IQ equalized signal, including: Mapping the data characteristic signal according to the target model parameter to obtain an I-way linear data signal, a Q-way linear data signal, an I-way nonlinear data signal, and a Q-way nonlinear data signal; The sum of the I-way linear data signal and the I-way nonlinear data signal is determined as an I-way balanced signal, and the sum of the Q-way linear data signal and the Q-way nonlinear data signal is determined as a Q-way balanced signal, wherein the IQ balanced signal includes the I-way balanced signal and the Q-way balanced signal.

2. The method according to claim 1, characterized in that: Mapping the data characteristic signal according to the target model parameter to obtain an I-way linear data signal, a Q-way linear data signal, an I-way nonlinear data signal and a Q-way nonlinear data signal comprises: Processing the linear crosstalk of I-channel optical signal of the data characteristic signal according to the target model parameters to obtain the I-channel linear data signal; Processing the linear crosstalk of the Q-path optical signal of the data characteristic signal according to the target model parameters to obtain the Q-path linear data signal; The crosstalk between the IQ two-path optical signals of the data characteristic signal is processed according to the target model parameters to obtain the I-path nonlinear data signal and the Q-path nonlinear data signal.

3. The method according to claim 2, characterized in that Processing the linear crosstalk of I-channel optical signal of the data characteristic signal according to the target model parameter to obtain the I-channel linear data signal includes: The I-way linear data signal is obtained by linearly mapping the data feature signal according to the target model parameters in the following manner: Wherein, 1≤t≤N, N is the length of the data characteristic signal, k'(t) is the data characteristic signal; and / or Processing the linear crosstalk of the Q-path optical signal of the data characteristic signal according to the target model parameter to obtain the Q-path linear data signal comprises: The Q-line linear data signal is obtained by linearly mapping the data feature signal according to the target model parameters in the following manner: and / or Processing the crosstalk between the IQ optical signals of the data characteristic signal according to the target model parameters to obtain the I-path nonlinear data signal and the Q-path nonlinear data signal comprises: The I-channel nonlinear data signal and the Q-channel nonlinear data signal are obtained by performing nonlinear mapping on the data feature signal according to the target model parameters in the following manner: in, is the target model parameter, and f(x) is the activation function.

4. The method according to claim 1, characterized in that: Performing cascade transformation on the I-channel received data signal and the Q-channel received data signal to obtain the data characteristic signal includes: The I-channel received data signal and the Q-channel received data signal are sampled respectively according to a window length of taps to obtain N groups of I-channel sampled signals and N groups of Q-channel sampled signals, wherein the center signal of each group of sampled signals in the N groups of I-channel sampled signals and the N groups of Q-channel sampled signals is the current sampling point, the (taps-1) / 2 points on the left side of the center signal are the (taps-1) / 2 sampling points intercepted forward from the current sampling point, and the (taps-1) / 2 points on the right side of the center signal are the (taps-1) / 2 sampling points intercepted backward from the current sampling point, if the forward or backward sampling points are less than (taps-1) / 2, 0 is added, the taps is an odd number, and N is the length of the I-channel received data signal and the Q-channel received data signal; The N groups of I-channel sampling signals with a length of taps are cascaded with the N groups of Q-channel sampling signals respectively to obtain the data characteristic signal including N signals with a length of 2 taps.

5. The method according to any one of claims 1 to 4, characterized in that Before performing cascade transformation on an IQ received data signal obtained by receiving the IQ transmitted data signal to obtain a data characteristic signal, the method further includes: Acquire an I-channel receiving modeling signal obtained by receiving an I-channel transmitting modeling signal, and a Q-channel receiving modeling signal obtained by receiving a Q-channel transmitting modeling signal; Performing cascade transformation on the I-channel received modeling signal and the Q-channel received modeling signal to obtain a modeling characteristic signal; The model parameters are updated according to the modeling characteristic signal to obtain the target model parameters.

6. The method according to claim 5, characterized in that Performing cascade transformation on the I-channel receiving modeling signal and the Q-channel receiving modeling signal to obtain the modeling characteristic signal includes: The I-channel receiving modeling signal and the Q-channel receiving modeling signal are sampled respectively according to the window length taps to obtain N groups of I-channel sampling signals and N groups of Q-channel sampling signals, wherein the center signal of each group of sampling signals in the N groups of I-channel sampling signals and the N groups of Q-channel sampling signals is the current sampling point, the (taps-1) / 2 points on the left side of the center signal are the (taps-1) / 2 sampling points intercepted forward from the current sampling point, and the (taps-1) / 2 points on the right side of the center signal are the (taps-1) / 2 sampling points intercepted backward from the current sampling point, if the forward or backward sampling points are less than (taps-1) / 2, 0 is added, the taps is an odd number, and N is the length of the I-channel receiving modeling signal and the Q-channel receiving modeling signal; The N groups of I-channel sampling signals with a length of the taps are cascaded with the N groups of Q-channel sampling signals respectively to obtain the modeling characteristic signal including N signals with a length of 2 taps.

7. The method according to claim 5, characterized in that The model parameters are updated according to the modeling characteristic signal to obtain the target model parameters, including: Repeat the following steps until the number of iterations t meets the preset iteration condition, and determine that the updated model parameters are the target model parameters: When t=0, the model parameters are initialized, and the modeling characteristic signals are mapped according to the model parameters to obtain an I-path linear modeling signal, a Q-path linear modeling signal, an I-path nonlinear modeling signal, and a Q-path nonlinear modeling signal; Performing sum processing on the I-way linear modeling signal and the I-way nonlinear modeling signal to obtain an I-way modeling output signal, and performing sum processing on the Q-way linear modeling signal and the Q-way nonlinear modeling signal to obtain a Q-way modeling output signal; The model parameters are updated according to the I-channel transmission modeling signal, the Q-channel transmission modeling signal, the I-channel modeling output signal, and the Q-channel modeling output signal to obtain updated model parameters, wherein the IQ transmission modeling signal includes the I-channel transmission modeling signal and the Q-channel transmission modeling signal; t=t+1.

8. The method according to claim 7, characterized in that Mapping the modeling characteristic signal according to the model parameters to obtain the I-path linear modeling signal, the Q-path linear modeling signal, the I-path nonlinear modeling signal, and the Q-path nonlinear modeling signal comprises: Processing the linear crosstalk of the I-channel optical signal of the modeling characteristic signal according to the model parameters to obtain the I-channel linear modeling signal; Processing the linear crosstalk of the Q-path optical signal of the modeling characteristic signal according to the model parameters to obtain the Q-path linear modeling signal; The crosstalk between the I and Q optical signals of the modeling characteristic signal is processed according to the model parameters to obtain the I-path nonlinear modeling signal and the Q-path nonlinear modeling signal.

9. The method according to claim 8, characterized in that Processing the linear crosstalk of the I-channel optical signal of the modeling characteristic signal according to the model parameters to obtain the I-channel linear modeling signal includes: The I-line linear modeling signal is obtained by linearly mapping the modeling characteristic signal according to the model parameters in the following manner: Wherein, 1≤t≤N, N is the length of the modeling feature, k(t) is the modeling feature signal; and / or Processing the linear crosstalk of the Q-path optical signal of the modeling characteristic signal according to the model parameters to obtain the Q-path linear modeling signal includes: The Q-line linear modeling signal is obtained by linearly mapping the modeling characteristic signal according to the model parameters in the following manner: and / or Processing the crosstalk between the I and Q optical signals of the modeling characteristic signal according to the model parameters to obtain the I nonlinear modeling signal and the Q nonlinear modeling signal comprises: The modeling characteristic signal is nonlinearly mapped according to the model parameters to obtain an I-channel nonlinear modeling signal and a Q-channel nonlinear modeling signal in the following manner: in, are the model parameters, and f(x) is the activation function.

10. The method according to claim 7, characterized in that The model parameters are updated according to the I-channel transmission modeling signal, the Q-channel transmission modeling signal, the I-channel modeling output signal, and the Q-channel modeling output signal to obtain the updated model parameters, including: Determine a signal output error according to the I-channel transmission modeling signal, the Q-channel transmission modeling signal, the I-channel modeling output signal, and the Q-channel modeling output signal; The model parameters are updated according to the signal output error to obtain the updated model parameters.

11. The method according to claim 10, characterized in that The model parameters are updated according to the signal output error, and the updated model parameters include: Determine the partial derivatives of the signal output error with respect to the model parameters respectively to obtain a target matrix having the same dimension as the model parameters; The model parameters are updated according to the target matrix to obtain the updated model parameters.

12. The method according to claim 11, characterized in that The model parameters are updated according to the target matrix to obtain the updated model parameters, including: The update parameters are determined by: M(t)=β1M(t-1)+(1-β1)G(t), V(t)=β2V(t-1)+(1-β2)G 2 (t), where H(t) is the update parameter, M(t) is the momentum matrix, V(t) is the velocity matrix, G(t) is the target matrix, α=0.001, β1=0.9, β2=0.999, ε=10 -8 , M(-1)=0, V(-1)=0, 0≤t≤t max , t max is the preset iteration number threshold; The difference between the model parameter and the update parameter is determined as the updated model parameter.

13. The method according to claim 10, characterized in that The method further comprises: The signal output error is determined according to the I-channel transmission modeling signal, the Q-channel transmission modeling signal, the I-channel modeling output signal, and the Q-channel modeling output signal in the following manner: Where r(t) is the signal output error, p I (a) is the output signal of the I-channel modeling, p Q (a) Output signal for Q-path modeling, s I (a) is the modeling signal of the I-path transmission, s Q (a) is the Q-path transmitted modeling signal, N is the length of the modeling characteristic signal, and a is the signal.

14. A signal compensation processing device, characterized in that: include: A first acquisition module is used to acquire an I-channel received data signal obtained by receiving an I-channel transmitted data signal, and a Q-channel received data signal obtained by receiving a Q-channel transmitted data signal; A first cascade module, used for performing cascade transformation on the I-channel received data signal and the Q-channel received data signal to obtain a data characteristic signal; A compensation module is used to compensate the characteristic signal according to a predetermined target model parameter to obtain an IQ equalized signal, including: mapping the data characteristic signal according to the target model parameter to obtain an I-way linear data signal, a Q-way linear data signal, an I-way nonlinear data signal, and a Q-way nonlinear data signal; determining the sum of the I-way linear data signal and the I-way nonlinear data signal as an I-way equalized signal, and determining the sum of the Q-way linear data signal and the Q-way nonlinear data signal as a Q-way equalized signal, wherein the IQ equalized signal includes the I-way equalized signal and the Q-way equalized signal.

15. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein the computer program is configured to execute the method according to any one of claims 1 to 13 when executed.

16. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to run the computer program to perform the method according to any one of claims 1 to 13.

Images