WO2023272739A1 - Procédé de décodage de canal, appareil, procédé de formation de modèle de réseau neuronal utilisé pour le décodage de canal, et appareil - Google Patents

Procédé de décodage de canal, appareil, procédé de formation de modèle de réseau neuronal utilisé pour le décodage de canal, et appareil Download PDF

Info

Publication number
WO2023272739A1
WO2023272739A1 PCT/CN2021/104367 CN2021104367W WO2023272739A1 WO 2023272739 A1 WO2023272739 A1 WO 2023272739A1 CN 2021104367 W CN2021104367 W CN 2021104367W WO 2023272739 A1 WO2023272739 A1 WO 2023272739A1
Authority
WO
WIPO (PCT)
Prior art keywords
decoding
samples
information sequence
codeword
discriminator
Prior art date
Application number
PCT/CN2021/104367
Other languages
English (en)
Chinese (zh)
Inventor
郑凤
庞博文
池连刚
Original Assignee
北京小米移动软件有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 北京小米移动软件有限公司 filed Critical 北京小米移动软件有限公司
Priority to PCT/CN2021/104367 priority Critical patent/WO2023272739A1/fr
Priority to CN202180002056.9A priority patent/CN115804067A/zh
Publication of WO2023272739A1 publication Critical patent/WO2023272739A1/fr

Links

Images

Classifications

    • H—ELECTRICITY
    • H03—ELECTRONIC CIRCUITRY
    • H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
    • H03M13/00—Coding, decoding or code conversion, for error detection or error correction; Coding theory basic assumptions; Coding bounds; Error probability evaluation methods; Channel models; Simulation or testing of codes
    • H03M13/37—Decoding methods or techniques, not specific to the particular type of coding provided for in groups H03M13/03 - H03M13/35
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L25/00—Baseband systems
    • H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines

Definitions

  • the present disclosure relates to the technical field of mobile communication, and in particular to a channel decoding method and device, and a training method and device for a neural network model used for channel decoding.
  • the present disclosure proposes a channel decoding method and device, which realizes decoding of codewords based on a pre-trained neural network model, thereby providing a channel decoding method with low bit error rate, low decoding time and low decoding complexity. Degree channel decoding scheme.
  • the present disclosure also proposes a training method and device for a neural network model for channel decoding, so as to obtain a neural network model that can be used for channel decoding through iterative training.
  • the embodiment of the first aspect of the present disclosure provides a training method for a neural network model for channel decoding, the neural network model includes a decoding generator and a decoding discriminator, and the method includes: obtaining information sequences including The information sequence training set of samples, and obtain the received codeword samples to be decoded based on the information sequence samples; use the received codeword samples as the input features of the decoding generator, and use the information sequence samples and the The decoded codeword sample output by the decoding generator is used as the input feature of the decoded discriminator, and whether the information sequence sample and the decoded codeword sample can be distinguished is used as the output of the decoded discriminator feature, perform iterative training to obtain the pre-trained neural network model.
  • the obtaining a received codeword sample to be decoded based on the information sequence sample includes: obtaining an encoded codeword sample by an encoder based on the information sequence sample; modulating the encoded codeword sample to obtain modulating codeword samples; and inputting the modulated codeword samples into a noise-added channel to obtain received codeword samples transmitted through the channel.
  • each round of training in iterative training includes: based on the received codeword samples obtained from the information sequence samples used for this round of training, the decoding code corresponding to the information sequence samples is obtained by the decoding generator word sample; based on the decoded codeword sample and the information sequence sample, determine whether the decoded codeword sample and the information sequence sample can be distinguished through the decoding discriminator; if it is determined that the decoded codeword sample and the information sequence sample can be distinguished codeword samples and the information sequence samples, update the decoding generator and the decoding discriminator through the backpropagation method, and repeat the above steps until all the information sequence samples used for this round of training have passed The next round of training starts after the decoding generator and the decoding discriminator process; and if it is determined that the decoding codeword sample and the information sequence sample cannot be distinguished, the iterative training is ended and the pre-training neural network model.
  • the training target of the decoding discriminator and the decoding generator is expressed as:
  • G represents the decoding generator
  • D represents the decoding discriminator
  • V(D, G) represents the difference between the information sequence samples and the decoding codeword samples
  • the training goal of the decoding discriminator is to maximize the difference between the information sequence samples and the decoding codeword samples
  • x represents the information sequence samples input to the decoding discriminator
  • x ⁇ p data (x) means that x obeys the data distribution input to the decoding generator
  • z means input noise
  • z ⁇ p z (z) means that z obeys the noise variable distribution input to the decoding generator, represents a probability distribution.
  • the decoding discriminator is updated using a gradient ascent method based on the following formula:
  • D represents the decoding discriminator
  • m represents the number of information sequence samples participating in the current round of training
  • x i represents the information sequence samples input to the decoding discriminator for the ith time
  • the decoding generator is updated using a gradient descent method based on the following formula:
  • G represents the decoding generator
  • D represents the decoding discriminator
  • m represents the number of information sequence samples participating in the current round of training
  • y i represents the received codeword samples input to the decoding generator for the ith time.
  • the encoder is a low density parity check LDPC code encoder.
  • the modulator is binary phase shift keying BPSK modulation.
  • the noise adding channel is one of the following: additive white noise AWGN channel; and Rayleigh channel.
  • the neural network model is a GAN model for generating an adversarial neural network.
  • the embodiment of the second aspect of the present disclosure provides a channel decoding method, including: based on the pre-trained neural network model, decoding the received codeword to be decoded after the information sequence is transmitted through the channel, so as to obtain the A decoding codeword corresponding to the received codeword; wherein, the pre-trained neural network model includes a decoding generator, and the decoding generator decodes the receiving codeword to output the decoding code Character.
  • the neural network model further includes a decoding discriminator
  • the pre-trained neural network model is obtained through the following process: obtaining an information sequence training set including information sequence samples, and obtaining the information sequence to be translated based on the information sequence samples.
  • the received codeword sample of the code take the received codeword sample as the input feature of the decoding generator, and use the information sequence sample and the decoded codeword sample output by the decoding generator as the decoding.
  • the input feature of the discriminator, and whether the information sequence sample can be distinguished from the decoded codeword sample is used as the output feature of the decoding discriminator, and iterative training is performed to obtain the pre-trained neural network model.
  • the obtaining a received codeword sample to be decoded based on the information sequence sample includes: obtaining an encoded codeword sample by an encoder based on the information sequence sample; modulating the encoded codeword sample to obtain modulating codeword samples; and inputting the modulated codeword samples into a noise-added channel to obtain received codeword samples transmitted through the channel.
  • each round of training in iterative training includes: based on the received codeword samples obtained from the information sequence samples used for this round of training, the decoding code corresponding to the information sequence samples is obtained by the decoding generator word sample; based on the decoded codeword sample and the information sequence sample, determine whether the decoded codeword sample and the information sequence sample can be distinguished through the decoding discriminator; if it is determined that the decoded codeword sample and the information sequence sample can be distinguished codeword samples and the information sequence samples, update the decoding generator and the decoding discriminator through the backpropagation method, and repeat the above steps until all the information sequence samples used for this round of training have passed The next round of training starts after the decoding generator and the decoding discriminator process; and if it is determined that the decoding codeword sample and the information sequence sample cannot be distinguished, the iterative training is ended and the pre-training neural network model.
  • the training target of the decoding discriminator and the decoding generator is expressed as:
  • G represents the decoding generator
  • D represents the decoding discriminator
  • V(D, G) represents the difference between the information sequence samples and the decoding codeword samples
  • the training goal of the decoding discriminator is to maximize the difference between the information sequence samples and the decoding codeword samples
  • x represents the information sequence samples input to the decoding discriminator
  • x ⁇ p data (x) means that x obeys the data distribution input to the decoding generator
  • z means input noise
  • z ⁇ p z (z) means that z obeys the noise variable distribution input to the decoding generator, represents a probability distribution.
  • the decoding discriminator is updated using a gradient ascent method based on the following formula:
  • D represents the decoding discriminator
  • m represents the number of information sequence samples participating in the current round of training
  • x i represents the information sequence samples input to the decoding discriminator for the ith time
  • the decoding generator is updated using a gradient descent method based on the following formula:
  • G represents the decoding generator
  • D represents the decoding discriminator
  • m represents the number of information sequence samples participating in the current round of training
  • y i represents the received codeword samples input to the decoding generator for the ith time.
  • the encoder is a low density parity check LDPC code encoder.
  • the modulator is binary phase shift keying BPSK modulation.
  • the noise adding channel is one of the following: additive white noise AWGN channel; and Rayleigh channel.
  • the neural network model is a GAN model for generating an adversarial neural network.
  • the embodiment of the third aspect of the present disclosure provides a neural network model training device for channel decoding, the neural network model includes a decoding generator and a decoding discriminator, the device includes: an acquisition module, used Obtaining an information sequence training set including information sequence samples, and obtaining received codeword samples to be decoded based on the information sequence samples; and a training module, configured to use the received codeword samples as the decoding generator
  • the input feature is to use the information sequence sample and the decoded code word sample output by the decoding generator as the input feature of the decoding discriminator, and to distinguish the information sequence sample from the decoded code Word samples are used as the output features of the decoding discriminator, and iterative training is performed to obtain the pre-trained neural network model.
  • the embodiment of the fourth aspect of the present disclosure provides a channel decoding device, including: a decoding module, used for receiving codewords to be decoded after the information sequence is transmitted through the channel based on the pre-trained neural network model Decoding is performed to obtain a decoding codeword corresponding to the received codeword; wherein, the pre-trained neural network model includes a decoding generator, and the decoding generator decodes the receiving codeword to output the decoded codeword.
  • the embodiment of the fifth aspect of the present disclosure provides an electronic device, including: a memory; a processor connected to the memory and configured to implement the above-mentioned embodiment of the first aspect by executing computer-executable instructions on the memory
  • a neural network model training method for channel decoding or the channel decoding method described in the embodiment of the second aspect is a neural network model training method for channel decoding or the channel decoding method described in the embodiment of the second aspect.
  • the embodiment of the sixth aspect of the present disclosure provides an electronic device, including: a model trainer, which is used to obtain a pre-trained neural network based on information sequence samples and received codeword samples corresponding to the information sequence samples to be decoded.
  • a network model wherein the neural network model includes a decoding generator and a decoding discriminator, the model trainer is used to use the received codeword sample as the input feature of the decoding generator, and the information sequence
  • the sample and the decoded codeword sample output by the decoding generator are used as the input features of the decoding discriminator, and whether the information sequence sample and the decoded codeword sample can be distinguished is used as the decoding discrimination
  • the output features of the device perform iterative training to obtain the pre-trained neural network model; the decoder, the decoder is used to obtain the information sequence to be obtained after the information sequence is transmitted through the channel based on the pre-trained neural network model
  • the decoded received codeword is decoded to obtain a decoded codeword corresponding to
  • the embodiment of the seventh aspect of the present disclosure provides a computer storage medium, wherein the computer storage medium stores computer-executable instructions; after the computer-executable instructions are executed by a processor, the above-mentioned embodiment of the first aspect can be implemented.
  • An embodiment of the present disclosure also provides a training method and device for a neural network model for channel decoding, by obtaining an information sequence training set including information sequence samples and obtaining the received code to be decoded based on the information sequence samples This, and take the received codeword sample as the input feature of the decoding generator, take the information sequence sample and the decoded codeword sample output by the decoding generator as the input feature of the decoding discriminator, and use whether the information sequence sample can be distinguished And the decoded codeword samples are used as the output features of the decoded discriminator, and iterative training is performed to obtain the trained neural network model.
  • the trained neural network model thus obtained can decode the received codeword to be decoded after the information sequence is transmitted through the channel to obtain the original information sequence.
  • the embodiment of the present disclosure provides a channel decoding method and device, which decodes the received codewords to be decoded after the information sequence is transmitted through the channel based on the pre-trained neural network model, so as to obtain the received codewords A corresponding decoding codeword, wherein the pre-trained neural network model includes a decoding generator for decoding the received codeword to output the decoding codeword. Therefore, the decoding of the codeword to be decoded can be realized through the pre-trained neural network model, thereby realizing a channel decoding scheme with low bit error rate, low decoding duration and low decoding complexity.
  • FIG. 1 is a schematic flowchart of a training method for a neural network model for channel decoding according to an embodiment of the present disclosure
  • FIG. 2 is a schematic flowchart of a training method for a neural network model for channel decoding according to an embodiment of the present disclosure
  • FIG. 3 is a schematic flowchart of a training method for a neural network model for channel decoding according to an embodiment of the present disclosure
  • FIG. 4 is a schematic flowchart of a channel decoding method according to an embodiment of the present disclosure
  • FIG. 5 is a schematic structural diagram of a training method and device for a neural network model for channel decoding provided by an embodiment of the present disclosure
  • FIG. 6 is a schematic structural diagram of a channel decoding device provided by an embodiment of the present disclosure.
  • FIG. 7 is a schematic structural diagram of an electronic device provided by an embodiment of the present disclosure.
  • FIG. 8 is a schematic structural diagram of an electronic device provided by an embodiment of the present disclosure.
  • the existing channel decoding algorithm based on the belief propagation algorithm after receiving the information sequence, all variable nodes receive the corresponding received value, and each variable node will transmit a reliability message to all the check nodes adjacent to it , each check node will process after receiving the reliability message, and will pass a new reliability message to all the variable nodes adjacent to it, this process can be regarded as an iteration, after an iteration, the Judgment, if the verification equation is satisfied, the decoding ends and the judgment result is output, otherwise iterates again, and so on until the verification equation is satisfied or the maximum number of iterations is reached. Due to the need for continuous iterations to achieve decoding, the existing channel decoding algorithm based on the belief propagation algorithm has high complexity, long decoding time, and complex implementation of decoding devices.
  • channel decoding technology is in the evolution stage of a new generation technology.
  • An important feature of the new generation of channel decoding technology is to maintain massive data transmission and maintain a low bit error rate while increasing the transmission rate.
  • different service types have different requirements for channel decoding technology according to performance requirements. For example, in eMBB service data, LDPC codes are used for long codes, and Polar codes are used for short codes. Therefore, it is necessary to provide a channel decoding scheme with lower bit error rate, faster decoding speed and lower decoding model complexity to support the transmission requirements of various service data.
  • the present disclosure proposes a channel decoding method and device, which realizes decoding of codewords based on a pre-trained neural network model, thereby providing a channel decoding method with low bit error rate, low decoding time and low decoding complexity. Degree channel decoding scheme.
  • the present disclosure also proposes a training method and device for a neural network model for channel decoding, so as to obtain a neural network model that can be used for channel decoding through a training process.
  • Fig. 1 shows a schematic flowchart of a training method for a neural network model for channel decoding according to an embodiment of the present disclosure.
  • the neural network model includes a decoding generator and a decoding discriminator. As shown in Figure 1, the method includes the following steps.
  • Step S101 obtaining an information sequence training set including information sequence samples, and obtaining received codeword samples to be decoded based on the information sequence samples.
  • the neural network model may be a Generative Adversarial Networks (GAN, Generative Adversarial Networks) model.
  • GAN Generative Adversarial Networks
  • GAN is a feature learning method of artificial intelligence network. It converts the original data into a higher-level and more abstract expression through some simple but nonlinear models. As long as there are enough conversion combinations, very complex features can also be be learned. At present, GAN has been applied in computer vision, image processing, speech recognition, natural language processing and other fields and has achieved superior performance.
  • the information sequence training set is a data set used for training the neural network model, which may include multiple information sequences.
  • the information sequence may be, for example, a binary information sequence.
  • the received codeword samples are obtained based on the information sequence samples, and the received codeword samples are obtained after encoding. Therefore, the received codeword samples are codeword samples that need to be decoded.
  • Step S102 take the received codeword sample as the input feature of the decoding generator, use the information sequence sample and the decoded codeword sample output by the decoding generator as the input feature of the decoding discriminator, and determine whether the information sequence sample can be distinguished and the decoded codeword samples are used as the output features of the decoded discriminator, and iterative training is performed to obtain a pre-trained neural network model.
  • the decoding generator outputs decoded codeword samples based on the received codeword samples.
  • the decoding generator in the neural network model is used to decode the received codeword samples to be decoded and output the corresponding decoded codeword samples, while the decoding discriminator is used to compare the information sequence samples and the decoded codeword samples obtained from the decoding generator. codeword samples to distinguish.
  • the decoding discriminator is used to compare the information sequence samples and the decoded codeword samples obtained from the decoding generator. codeword samples to distinguish.
  • an information sequence training set including information sequence samples and obtaining received codeword samples to be decoded based on the information sequence samples, and using the received codeword samples as input features of the decoding generator, to
  • the information sequence samples and the decoded codeword samples output by the decoding generator are used as the input features of the decoding discriminator, and the ability to distinguish the information sequence samples and the decoded codeword samples as the output features of the decoding discriminator is used for iterative training to obtain a trained neural network model.
  • the trained neural network model thus obtained can decode the received codeword to be decoded after the information sequence is transmitted through the channel to obtain the original information sequence.
  • FIG. 2 shows a schematic flowchart of a training method for a neural network model for channel decoding according to an embodiment of the present disclosure. Based on the embodiment shown in FIG. 1 , in this example embodiment, an information sequence based The samples obtain a specific implementation of the received codeword samples to be decoded.
  • step S101 shown in FIG. 1 may specifically include the following steps.
  • step S201 based on the information sequence samples, codeword samples are obtained through an encoder.
  • An encoder may be used to encode information sequence samples to obtain encoded codeword samples.
  • the encoder can be an encoder used in a channel coding scheme in a mobile communication system, such as a low density parity check code (LDPC, Low Density Parity Check) encoder, a polar (Polar) code encoder, a Turbo code encoder device.
  • LDPC low density parity check code
  • polar (Polar) code encoder a Turbo code encoder device.
  • Turbo code encoder device since LDPC code is a linear block code with a sparse check matrix, its characteristics are completely determined by the parity check matrix, that is, it has a structured structure, which is more suitable for deep learning network learning. In other words, it is easier to obtain a neural network model suitable for decoding LDPC codes.
  • Step S202 Modulate the encoded codeword samples to obtain modulated codeword samples.
  • modulated codeword samples can be obtained by modulating the encoded codeword samples.
  • modulated codeword samples for example, binary phase shift keying (BPSK, Binary Phase Shift Keying) modulation, frequency shift keying (FSK, Frequency Shift Keying) modulation and other modulation schemes can be used to modulate the encoded codeword samples.
  • BPSK Binary Phase Shift Keying
  • FSK Frequency Shift Keying
  • Step S203 input the modulated codeword samples into the noise adding channel to obtain the received codeword samples transmitted through the channel.
  • the modulated codeword samples obtained through modulation can be transmitted through the noise-added channel to obtain received codeword samples, so that the received codeword samples are codewords with noise.
  • the noise-added channel may be an Additive White Gaussian Noise (AWGN) channel or a Rayleigh channel.
  • AWGN Additive White Gaussian Noise
  • the received codeword samples to be decoded can be obtained by encoding, modulating, and adding noise to the channel of the information sequence samples.
  • FIG. 3 shows a schematic flowchart of a training method for a neural network model for channel decoding according to an embodiment of the present disclosure. Based on the embodiment shown in FIG. 1, in this example embodiment, it describes the iterative training The specific implementation of each round of training.
  • step S102 shown in FIG. 1 may specifically include the following steps.
  • Step S301 based on the received codeword samples obtained from the information sequence samples used for the current round of training, the decoded codeword samples corresponding to the information sequence samples are obtained through the decoding generator.
  • one or more information sequence samples can be used.
  • received codeword samples obtained from one or more information sequence samples may be input into the decoding generator to obtain one or more decoded codeword samples respectively corresponding to the one or more information sequence samples.
  • Step S302 based on the decoded codeword samples and the information sequence samples, determine whether the decoded codeword samples and the information sequence samples can be distinguished through the decoding discriminator.
  • One or more decoded codeword samples and corresponding one or more information sequence samples output from the decode generator are input into a decode discriminator, which determines whether the decoded codeword samples and the information sequence samples can be distinguished .
  • the training objectives of the decode discriminator and the decode generator can be expressed as:
  • G represents the decoding generator
  • D represents the decoding discriminator
  • V(D, G) represents the difference between the information sequence samples and the decoded codeword samples
  • the difference between the information sequence sample and the decoded codeword sample is distinguished in the z region
  • x represents the input data of the decoding discriminator
  • x ⁇ p data (x) represents that x obeys the data distribution input to the decoding generator
  • z Indicates the input noise
  • z ⁇ p z (z) indicates that z obeys the noise variable distribution input to the decoding generator
  • p z (z) may obey a normal distribution.
  • the training goal of the decoding discriminator is to update the decoding discriminator while keeping the decoding generator unchanged, and the training target of the decoding generator is to analyze the information sequence samples based on the updated decoding discriminator. This is accomplished by updating the decoding generator to differentiate between the differences from the decoded codeword samples.
  • Step S303 if it is determined that the decoded codeword sample and the information sequence sample can be distinguished, update the decoding generator and the decoding discriminator through the back propagation method, and repeat the above steps until all the information sequence samples used for this round of training After being processed by the decoding generator and the decoding discriminator, the next round of training starts.
  • the decoding discriminator can distinguish the decoded codeword samples from the information sequence samples, it means that the difference between the decoded codeword samples generated by the decoding generator from the received codeword samples and the information sequence samples can be distinguished by the decoding discriminator , that is, the decoding generator does not completely restore the information sequence to the decoding of the received codeword samples. Therefore, the current neural network model is not suitable for decoding the received codewords, and the decoding generator needs to be updated so that it can Learn a more suitable decoding algorithm, and at the same time need to update the decoding discriminator to enhance the discriminative ability of the decoding discriminator, and continuously update the decoding generator and the decoding discriminator through the training process, so as to finally reach the Nash equilibrium state.
  • the decoding discriminator For a decoded codeword sample and an information sequence sample input to the decoding discriminator, if the decoding discriminator can distinguish the decoded codeword sample from the information sequence sample, then the decoding generator and the decoding The code discriminator is updated. Specifically, the decoding generator is updated by using the gradient descent method so that the updated decoding generator can minimize the difference between the information sequence samples and the decoded codeword samples, and the decoding The code discriminator is updated so that the updated decoding discriminator can maximize the difference between the information sequence samples and the decoded code word samples.
  • the decoding discriminator is updated using gradient ascent based on the following formula:
  • D represents the decoding discriminator
  • m represents the number of information sequence samples participating in the current round of training
  • x i represents the information sequence samples input to the decoding discriminator for the i time
  • the decoding generator is updated using gradient descent based on the following formula:
  • G represents the decoding generator
  • D represents the discriminator
  • m represents the number of information sequence samples participating in the current round of training
  • y i represents the received codeword samples input to the decoding generator for the ith time.
  • the decoding generator and the decoding discriminator are updated, if there are information sequences that have not been processed by the decoding generator and the decoding discriminator in one or more information sequence samples used for this round of training sample, then input the received codeword sample obtained from any of the unprocessed information sequence samples into the decoding generator to obtain the decoding codeword sample corresponding to the information sequence sample, and combine the information sequence sample and the decoding code Word samples are input into the decoding discriminator to determine that the two can be distinguished.
  • the decoding generator and the decoding discriminator are updated again, and so on, until one or more information used for this round of training There are no information sequence samples in the sequence samples that have not been processed by the decoding generator and the decoding discriminator, that is, all the information sequence samples used for this round of training have been used, and the next round of training can start.
  • Step S304 if it is determined that the decoded codeword sample and the information sequence sample cannot be distinguished, the iterative training is ended and a pre-trained neural network model is obtained.
  • the decoding discriminator cannot distinguish the decoded codeword samples from the information sequence samples, it indicates that the decoding generator can completely restore the information sequence samples according to the received codeword samples. Therefore, the current neural network model is suitable for decoding the received codeword samples. Code, the iterative training ends, so as to obtain the pre-trained neural network model.
  • the decoding discriminator is continuously updated by the gradient ascending method and the decoding generator is continuously updated by the gradient descent method, so that it is possible to obtain A trained neural network model.
  • Fig. 4 shows a schematic flowchart of a channel decoding method according to an embodiment of the present disclosure. As shown in Fig. 4 , the method includes the following steps.
  • Step S401 based on the pre-trained neural network model, decode the received codeword to be decoded obtained after the information sequence is transmitted through the channel, so as to obtain the decoded codeword corresponding to the received codeword.
  • the pre-trained neural network model includes a decoding generator, and the decoding generator decodes the received codeword to output the decoded codeword.
  • the received codeword obtained after the information sequence is transmitted through the channel can be input into the pre-trained neural network model, and the pre-trained neural network model decodes the received codeword through its decoding generator to output The decoded codeword corresponding to the received codeword.
  • the neural network model may be a Generative Adversarial Networks (GAN, Generative Adversarial Networks) model.
  • GAN Generative Adversarial Networks
  • Using the pre-trained neural network model, especially the GAN model to realize channel decoding can bring the following advantages:
  • the channel decoding based on the neural network model will shorten the decoding time after the training of the neural network model is completed, which reduces the decoding delay.
  • GAN is based on backpropagation and does not require Markov chain, thus reducing the complexity of decoding algorithm.
  • the decoding of the received codeword obtained after the initial codeword is transmitted through the channel is realized by using a pre-trained neural network model, thereby realizing a low bit error rate, low decoding time and low decoding time. Code complexity channel decoding scheme.
  • the pre-trained neural network model can be obtained based on the training method of the neural network model for channel decoding described above with reference to FIGS.
  • the present disclosure also provides a training device for the neural network model for channel decoding.
  • the training device for the neural network model for channel decoding corresponds to the training method for the neural network model for channel decoding provided by the above-mentioned several embodiments, so in the implementation of the training method for the neural network model for channel decoding
  • the method is also applicable to the training device for the neural network model for channel decoding provided in this embodiment, and will not be described in detail in this embodiment.
  • Fig. 5 is a schematic structural diagram of a training device for a neural network model for channel decoding proposed according to the present disclosure.
  • FIG. 5 is a schematic structural diagram of a neural network model training device 500 for channel decoding provided by an embodiment of the present disclosure.
  • the neural network model includes a decoding generator and a decoding discriminator.
  • the training device 500 for the neural network model of channel decoding comprises:
  • An acquisition module 501 configured to acquire an information sequence training set including information sequence samples, and obtain received codeword samples to be decoded based on the information sequence samples;
  • the training module 502 is configured to use the received codeword samples as the input features of the decoding generator, and use the information sequence samples and the decoded codeword samples output by the decoding generator as the decoding discrimination
  • the input feature of the discriminator, and whether the information sequence sample can be distinguished from the decoded codeword sample is used as the output feature of the decoding discriminator, and iterative training is performed to obtain the pre-trained neural network model.
  • an information sequence training set including information sequence samples and obtaining received codeword samples to be decoded based on the information sequence samples, and using the received codeword samples as input features of the decoding generator, to
  • the information sequence samples and the decoded codeword samples output by the decoding generator are used as the input features of the decoding discriminator, and the ability to distinguish the information sequence samples and the decoded codeword samples as the output features of the decoding discriminator is used for iterative training to obtain a trained neural network model.
  • the trained neural network model thus obtained can decode the received codeword to be decoded after the information sequence is transmitted through the channel to obtain the original information sequence.
  • the obtaining module 501 is configured to: obtain coded codeword samples through an encoder based on information sequence samples; modulate coded codeword samples to obtain modulated codeword samples; and input modulated codeword samples into noise channel to obtain received codeword samples after channel transmission.
  • the training module 502 is configured to: obtain the decoded codeword samples corresponding to the information sequence samples through the decoding generator based on the received codeword samples obtained from the information sequence samples used for the current round of training; Based on the decoded codeword samples and information sequence samples, determine whether the decoded codeword samples and information sequence samples can be distinguished through the decoding discriminator; if it is determined that the decoded codeword samples and information sequence samples can be distinguished, use the back propagation method The decoding generator and the decoding discriminator are updated, and the above steps are repeated until all information sequence samples used for this round of training have been processed by the decoding generator and the decoding discriminator, and then start downloading rounds of training; if it is determined that the decoded codeword samples and information sequence samples cannot be distinguished, the iterative training is ended and a pre-trained neural network model is obtained.
  • the training objectives of the decoding discriminator and the decoding generator are expressed as:
  • G represents the decoding generator
  • D represents the decoding discriminator
  • V(D, G) represents the difference between the information sequence samples and the decoding codeword samples
  • the training goal of the decoding discriminator is to maximize the difference between the information sequence samples and the decoding codeword samples
  • x represents the information sequence samples input to the decoding discriminator
  • x ⁇ p data (x) means that x obeys the data distribution input to the decoding generator
  • z means input noise
  • z ⁇ p z (z) means that z obeys the noise variable distribution input to the decoding generator, represents a probability distribution.
  • the decoding discriminator is updated using gradient ascent based on the following formula:
  • D represents the decoding discriminator
  • m represents the number of information sequence samples participating in the current round of training
  • x i represents the information sequence samples input to the decoding discriminator for the ith time
  • the decoding generator is updated using a gradient descent method based on the following formula:
  • G represents the decoding generator
  • D represents the decoding discriminator
  • m represents the number of information sequence samples participating in the current round of training
  • y i represents the received codeword samples input to the decoding generator for the ith time.
  • the encoder is a low density parity check LDPC code encoder.
  • the modulator is binary phase shift keyed BPSK modulation.
  • the noise adding channel is one of: an additive white noise AWGN channel; and a Rayleigh channel.
  • the neural network model is a GAN model.
  • FIG. 6 is a schematic structural diagram of a channel decoding apparatus 600 provided by an embodiment of the present disclosure.
  • the device 600 includes: a decoding module 601, configured to decode the received codeword to be decoded after the information sequence is transmitted through the channel based on the pre-trained neural network model, to obtain A decoded codeword corresponding to the received codeword; wherein, the pre-trained neural network model includes a decoded generator, and the decoded generator decodes the received codeword to output the decoded codeword numbers.
  • the received codeword to be decoded after the information sequence is transmitted through the channel is decoded to obtain the decoded codeword corresponding to the received codeword, wherein the pre-trained
  • the neural network model of includes a decoding generator for decoding received codewords to output decoded codewords. Therefore, the decoding of the codeword to be decoded can be realized through the pre-trained neural network model, thereby realizing a channel decoding scheme with low bit error rate, low decoding time and low decoding complexity.
  • the neural network model further includes a decoding discriminator
  • the pre-trained neural network model is obtained through the following process: obtaining an information sequence training set including information sequence samples, and obtaining The received codeword sample to be decoded; the received codeword sample is used as the input feature of the decoding generator, and the information sequence sample and the decoded codeword sample output by the decoding generator are used as the Decoding the input features of the discriminator, and taking whether the information sequence sample and the decoded codeword sample can be distinguished as the output feature of the decoding discriminator, performing iterative training to obtain the pre-trained neural network model .
  • the obtaining the received codeword samples to be decoded based on the information sequence samples includes: obtaining encoded codeword samples by an encoder based on the information sequence samples; and modulating the encoded codeword samples obtaining modulation codeword samples; and inputting the modulation codeword samples into a noise-added channel to obtain received codeword samples transmitted through the channel.
  • each round of training in the iterative training includes: based on the received codeword samples obtained from the information sequence samples used for the current round of training, the decoding generator corresponding to the information sequence samples is obtained by the decoding generator.
  • Codeword samples based on the decoded codeword samples and the information sequence samples, determine whether the decoded codeword samples and the information sequence samples can be distinguished through the decoding discriminator; if it is determined that the information sequence samples can be distinguished
  • the decoding codeword samples and the information sequence samples are updated by the back propagation method to the decoding generator and the decoding discriminator, and repeat the above steps until all the information sequence samples are used for this round of training Start the next round of training after being processed by the decoding generator and the decoding discriminator; and if it is determined that the decoding codeword samples and the information sequence samples cannot be distinguished, end the iterative training and obtain the Pretrained neural network models.
  • the training objectives of the decoding discriminator and the decoding generator are expressed as:
  • G represents the decoding generator
  • D represents the decoding discriminator
  • V(D, G) represents the difference between the information sequence samples and the decoding codeword samples
  • the training goal of the decoding discriminator is to maximize the difference between the information sequence samples and the decoding codeword samples
  • x represents the information sequence samples input to the decoding discriminator
  • x ⁇ p data (x) means that x obeys the data distribution input to the decoding generator
  • z means input noise
  • z ⁇ p z (z) means that z obeys the noise variable distribution input to the decoding generator, represents a probability distribution.
  • the decoding discriminator is updated using gradient ascent based on the following formula:
  • D represents the decoding discriminator
  • m represents the number of information sequence samples participating in the current round of training
  • x i represents the information sequence samples input to the decoding discriminator for the ith time
  • the decoding generator is updated using a gradient descent method based on the following formula:
  • G represents the decoding generator
  • D represents the decoding discriminator
  • m represents the number of information sequence samples participating in the current round of training
  • y i represents the received codeword samples input to the decoding generator for the ith time.
  • the encoder is a low density parity check LDPC code encoder.
  • the modulator is binary phase shift keyed BPSK modulation.
  • the noise adding channel is one of: an additive white noise AWGN channel; and a Rayleigh channel.
  • the neural network model is a GAN model.
  • the present disclosure also provides an electronic device and a computer-readable storage medium.
  • FIG. 7 it is a block diagram of an electronic device according to an embodiment of the present disclosure.
  • Electronic device is intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other suitable computers.
  • Electronic devices may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smart phones, wearable devices, and other similar computing devices.
  • the components shown herein, their connections and relationships, and their functions, are by way of example only, and are not intended to limit implementations of the disclosure described and/or claimed herein.
  • the electronic device includes: one or more processors 710 , a memory 720 , and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces.
  • the various components are interconnected using different buses and can be mounted on a common motherboard or otherwise as desired.
  • the processor may process instructions executed within the electronic device, including instructions stored in or on the memory, to display graphical information of a GUI on an external input/output device such as a display device coupled to an interface.
  • multiple processors and/or multiple buses may be used with multiple memories and multiple memories, if desired.
  • multiple electronic devices may be connected, with each device providing some of the necessary operations (eg, as a server array, a set of blade servers, or a multi-processor system).
  • a processor 710 is taken as an example.
  • the memory 720 is a non-transitory computer-readable storage medium provided in the present disclosure.
  • the memory stores instructions executable by at least one processor, so that the at least one processor executes the data transmission method provided in the present disclosure.
  • the non-transitory computer-readable storage medium of the present disclosure stores computer instructions, and the computer instructions are used to cause a computer to execute the data transmission method provided by the present disclosure.
  • the memory 720 can be used to store non-transitory software programs, non-transitory computer-executable programs and modules, such as program instructions/modules corresponding to the data transmission method in the embodiments of the present disclosure.
  • the processor 710 executes various functional applications and data processing of the server by running the non-transitory software programs, instructions and modules stored in the memory 720, that is, implements the data transmission method in the above method embodiments.
  • the memory 720 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and an application program required by at least one function; the data storage area may store data created according to the use of the positioning electronic device, and the like.
  • the memory 720 may include a high-speed random access memory, and may also include a non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state storage devices.
  • the storage 720 may optionally include storages that are set remotely relative to the processor 710, and these remote storages may be connected to the positioning electronic device through a network. Examples of the aforementioned networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.
  • the electronic device may further include: an input device 730 and an output device 740 .
  • the processor 710, the memory 720, the input device 730, and the output device 740 may be connected via a bus or in other ways, and connection via a bus is taken as an example in FIG. 7 .
  • the input device 730 can receive input numbers or character information, and generate key signal input related to user settings and function control of the positioning electronic equipment, such as a touch screen, a keypad, a mouse, a trackpad, a touchpad, a pointing stick, one or more Input devices such as mouse buttons, trackballs, joysticks, etc.
  • the output device 740 may include a display device, an auxiliary lighting device (eg, LED), a tactile feedback device (eg, a vibration motor), and the like.
  • the display device may include, but is not limited to, a liquid crystal display (LCD), a light emitting diode (LED) display, and a plasma display. In some implementations, the display device may be a touch screen.
  • Various implementations of the systems and techniques described herein can be implemented in digital electronic circuitry, integrated circuit systems, application specific ASICs (application specific integrated circuits), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include being implemented in one or more computer programs executable and/or interpreted on a programmable system including at least one programmable processor, the programmable processor Can be special-purpose or general-purpose programmable processor, can receive data and instruction from storage system, at least one input device, and at least one output device, and transmit data and instruction to this storage system, this at least one input device, and this at least one output device an output device.
  • machine-readable medium and “computer-readable medium” refer to any computer program product, apparatus, and/or means for providing machine instructions and/or data to a programmable processor ( For example, magnetic disks, optical disks, memories, programmable logic devices (PLDs), including machine-readable media that receive machine instructions as machine-readable signals.
  • machine-readable signal refers to any signal used to provide machine instructions and/or data to a programmable processor.
  • the systems and techniques described herein can be implemented on a computer having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user. ); and a keyboard and pointing device (eg, a mouse or a trackball) through which a user can provide input to the computer.
  • a display device e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor
  • a keyboard and pointing device eg, a mouse or a trackball
  • Other kinds of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and can be in any form (including Acoustic input, speech input or, tactile input) to receive input from the user.
  • the systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., as a a user computer having a graphical user interface or web browser through which a user can interact with embodiments of the systems and techniques described herein), or including such backend components, middleware components, Or any combination of front-end components in a computing system.
  • the components of the system can be interconnected by any form or medium of digital data communication, eg, a communication network. Examples of communication networks include: Local Area Network (LAN), Wide Area Network (WAN) and the Internet.
  • a computer system may include clients and servers.
  • Clients and servers are generally remote from each other and typically interact through a communication network.
  • the relationship of client and server arises by computer programs running on the respective computers and having a client-server relationship to each other.
  • FIG. 8 shows a block diagram of another electronic device according to an embodiment of the present disclosure.
  • an electronic device 800 includes a model trainer 810 and a decoder 820 .
  • the model trainer 810 is used to obtain a pre-trained neural network model based on information sequence samples and received codeword samples corresponding to the information sequence samples to be decoded, wherein the neural network model includes a decoding generator and decoding discriminator.
  • the model trainer is used to use the received codeword samples as the input features of the decoding generator, and use the information sequence samples and the decoded codeword samples output by the decoding generator as the decoding
  • the input feature of the discriminator, and whether the information sequence sample can be distinguished from the decoded codeword sample is used as the output feature of the decoding discriminator, and iterative training is performed to obtain the pre-trained neural network model.
  • the decoder 820 is configured to decode the received codeword to be decoded after the information sequence is transmitted through the channel based on the pre-trained neural network model, so as to obtain a decoding code corresponding to the received codeword Character.
  • the present application also provides a readable storage medium on which instructions are stored, and when the instructions are executed by a computer, the functions of any one of the above method embodiments are realized.
  • the present application also provides a computer program product, which implements the functions of any one of the above method embodiments when executed by a computer.
  • all or part of them may be implemented by software, hardware, firmware or any combination thereof.
  • software When implemented using software, it may be implemented in whole or in part in the form of a computer program product.
  • the computer program product comprises one or more computer programs. When the computer program is loaded and executed on the computer, all or part of the processes or functions according to the embodiments of the present application will be generated.
  • the computer can be a general purpose computer, a special purpose computer, a computer network, or other programmable devices.
  • the computer program can be stored in or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer program can be downloaded from a website, computer, server or data center Transmission to another website site, computer, server or data center by wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.).
  • the computer-readable storage medium may be any available medium that can be accessed by a computer, or a data storage device such as a server or a data center integrated with one or more available media.
  • the available medium may be a magnetic medium (for example, a floppy disk, a hard disk, a magnetic tape), an optical medium (for example, a high-density digital video disc (digital video disc, DVD)), or a semiconductor medium (for example, a solid state disk (solid state disk, SSD)) etc.
  • a magnetic medium for example, a floppy disk, a hard disk, a magnetic tape
  • an optical medium for example, a high-density digital video disc (digital video disc, DVD)
  • a semiconductor medium for example, a solid state disk (solid state disk, SSD)
  • At least one in this application can also be described as one or more, and multiple can be two, three, four or more, and this application does not make a limitation.
  • the technical feature is distinguished by "first”, “second”, “third”, “A”, “B”, “C” and “D”, etc.
  • the technical features described in the “first”, “second”, “third”, “A”, “B”, “C” and “D” have no sequence or order of magnitude among the technical features described.
  • the corresponding relationships shown in the tables in this application can be configured or predefined.
  • the values of the information in each table are just examples, and may be configured as other values, which are not limited in this application.
  • the corresponding relationship shown in some rows may not be configured.
  • appropriate deformation adjustments can be made based on the above table, for example, splitting, merging, and so on.
  • the names of the parameters shown in the titles of the above tables may also adopt other names understandable by the communication device, and the values or representations of the parameters may also be other values or representations understandable by the communication device.
  • other data structures can also be used, for example, arrays, queues, containers, stacks, linear tables, pointers, linked lists, trees, graphs, structures, classes, heaps, hash tables or hash tables can be used Wait.
  • Predefined in this application can be understood as defining, predefining, storing, prestoring, prenegotiating, preconfiguring, curing, or prefiring.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Theoretical Computer Science (AREA)
  • Power Engineering (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Error Detection And Correction (AREA)

Abstract

La présente divulgation se rapporte au domaine de la communication et concerne un procédé de décodage de canal. Une solution technique de la présente divulgation consiste essentiellement à : décoder, sur la base d'un modèle de réseau neuronal pré-formé, un mot de code reçu devant être décodé qui est obtenu après qu'une séquence d'informations a subi une transmission de canal, de façon à obtenir un mot de code décodé correspondant au mot de code reçu ; le modèle de réseau neuronal pré-formé comprenant un générateur de décodage, et le générateur de décodage décodant le mot de code reçu et délivrant en sortie le mot de code décodé.
PCT/CN2021/104367 2021-07-02 2021-07-02 Procédé de décodage de canal, appareil, procédé de formation de modèle de réseau neuronal utilisé pour le décodage de canal, et appareil WO2023272739A1 (fr)

Priority Applications (2)

Application Number Priority Date Filing Date Title
PCT/CN2021/104367 WO2023272739A1 (fr) 2021-07-02 2021-07-02 Procédé de décodage de canal, appareil, procédé de formation de modèle de réseau neuronal utilisé pour le décodage de canal, et appareil
CN202180002056.9A CN115804067A (zh) 2021-07-02 2021-07-02 信道译码方法及装置、用于信道译码的神经网络模型的训练方法及装置

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2021/104367 WO2023272739A1 (fr) 2021-07-02 2021-07-02 Procédé de décodage de canal, appareil, procédé de formation de modèle de réseau neuronal utilisé pour le décodage de canal, et appareil

Publications (1)

Publication Number Publication Date
WO2023272739A1 true WO2023272739A1 (fr) 2023-01-05

Family

ID=84692207

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/104367 WO2023272739A1 (fr) 2021-07-02 2021-07-02 Procédé de décodage de canal, appareil, procédé de formation de modèle de réseau neuronal utilisé pour le décodage de canal, et appareil

Country Status (2)

Country Link
CN (1) CN115804067A (fr)
WO (1) WO2023272739A1 (fr)

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108650201A (zh) * 2018-05-10 2018-10-12 东南大学 基于神经网络的信道均衡方法、译码方法及对应设备
CN109361404A (zh) * 2018-09-28 2019-02-19 华南理工大学 一种基于半监督深度学习网络的ldpc译码系统及译码方法
CN110351212A (zh) * 2019-08-10 2019-10-18 南京理工大学 快衰落信道下基于卷积神经网络的信道估计方法
CN111106839A (zh) * 2019-12-19 2020-05-05 北京邮电大学 一种基于神经网络的极化码译码方法及装置
CN111901000A (zh) * 2020-08-03 2020-11-06 电子科技大学 基于Inception结构的信道译码方法
US20210201003A1 (en) * 2019-12-30 2021-07-01 Affectiva, Inc. Synthetic data for neural network training using vectors

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108650201A (zh) * 2018-05-10 2018-10-12 东南大学 基于神经网络的信道均衡方法、译码方法及对应设备
CN109361404A (zh) * 2018-09-28 2019-02-19 华南理工大学 一种基于半监督深度学习网络的ldpc译码系统及译码方法
CN110351212A (zh) * 2019-08-10 2019-10-18 南京理工大学 快衰落信道下基于卷积神经网络的信道估计方法
CN111106839A (zh) * 2019-12-19 2020-05-05 北京邮电大学 一种基于神经网络的极化码译码方法及装置
US20210201003A1 (en) * 2019-12-30 2021-07-01 Affectiva, Inc. Synthetic data for neural network training using vectors
CN111901000A (zh) * 2020-08-03 2020-11-06 电子科技大学 基于Inception结构的信道译码方法

Also Published As

Publication number Publication date
CN115804067A (zh) 2023-03-14

Similar Documents

Publication Publication Date Title
CN110226289B (zh) 接收器和用于解码的方法
US11334467B2 (en) Representing source code in vector space to detect errors
KR102617929B1 (ko) 그래프 신경 망을 사전 트레이닝하는 방법 및 장치
US9941900B1 (en) Techniques for general-purpose lossless data compression using a recurrent neural network
US11722148B2 (en) Systems and methods of data compression
CN111539224B (zh) 语义理解模型的剪枝方法、装置、电子设备和存储介质
CN111563593B (zh) 神经网络模型的训练方法和装置
CN110569969A (zh) 网络模型结构的采样方法、装置以及电子设备
Chen et al. Practical polar code construction over parallel channels
Yang et al. Nonlinear programming approaches to decoding low-density parity-check codes
WO2021034387A1 (fr) Approximation d'arbre de huffman flexible en vue d'un codage à faible latence
CN111882035A (zh) 基于卷积核的超网络搜索方法、装置、设备和介质
Cao et al. Minimal sets for capacity-approaching variable-length constrained sequence codes
KR102213345B1 (ko) 패리티 검사 연접 극 부호의 설계 방법 및 그 장치
US9595976B1 (en) Folded integer encoding
US20230069079A1 (en) Statistical K-means Clustering
KR102607536B1 (ko) 옵티마이저 학습 방법, 장치, 전자 기기 및 판독 가능 기록 매체
CN112580723B (zh) 多模型融合方法、装置、电子设备和存储介质
WO2023272739A1 (fr) Procédé de décodage de canal, appareil, procédé de formation de modèle de réseau neuronal utilisé pour le décodage de canal, et appareil
CN111160552B (zh) 新闻信息的推荐处理方法、装置、设备和计算机存储介质
CN112804029A (zh) 基于ldpc码的bats码的传输方法、装置、设备及可读存储介质
Andi et al. Fast decoding of polar codes using tree structure
CN111539225B (zh) 语义理解框架结构的搜索方法和装置
CN113872610A (zh) 一种ldpc码神经网络训练、译码方法及其系统
Sha et al. Improved BP decoder for polar codes based on a modified kernel matrix

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21947684

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE