A channel adaptive hybrid automatic repeat request method and system

Abstract

The application discloses a kind of channel adaptive hybrid automatic repeat request method and system, method includes: using self-supervised mechanism to the recurrent neural network prediction model is trained to obtain channel quality prediction model;Channel quality parameter is predicted based on channel quality prediction model;The prediction result of channel quality parameter is compared with preset threshold value, and according to the comparison result, the quality of the channel is determined, and according to the quality of the channel, the encoding and transmission mode of data are determined to be transmitted;The application can determine the encoding and transmission mode of data according to the prediction result of channel quality, to take into account high reliability, low delay requirement.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of Internet of Things wireless communication, and particularly relates to a channel adaptive hybrid automatic repeat request method and system. BACKGROUND

[0002] With the rapid development of semiconductor technology, computer technology and wireless communication technology, the Internet of Things has been widely used in national defense security, industrial control, agricultural production and many other fields. In the field of wireless communication in the transformer area of smart grid, the Internet of Things technology also plays an important role. In recent years, a large number of distributed power, electric vehicles and energy storage are connected to the grid, and the operation mode of the grid dispatching has begun to change to source-grid-load-storage coordinated control, which has given rise to distributed photovoltaic regulation and precise load control. This type of business needs high-reliable and low-latency communication technology to carry, which puts higher requirements on the wireless communication technology in the transformer area.

[0003] The existing wireless communication technology relies on channel coding and hybrid automatic repeat request mechanism to effectively improve the reliability of transmission. However, due to the limited processing and communication capabilities of Internet of Things terminals, these mechanisms inevitably increase the processing and communication delay. On the one hand, for wireless terminals with high channel quality, non-coding transmission can already obtain good reliability, and coding transmission not only cannot improve the reliability, but also increases the processing delay. On the other hand, for wireless terminals with poor channel quality, the data after channel coding still has a high probability of needing retransmission, which will further increase the delay. SUMMARY

[0004] The purpose of the present application is to provide a channel adaptive hybrid automatic repeat request method and system that can reduce processing and communication delay.

[0005] Technical scheme: The channel adaptive hybrid automatic repeat request method provided by the present application comprises:

[0006] training a recurrent neural network prediction model using a self-supervised mechanism to obtain a channel quality prediction model;

[0007] predicting a channel quality parameter based on the channel quality prediction model;

[0008] comparing the prediction result of the channel quality parameter with a preset threshold, determining the channel quality according to the comparison result, and determining the encoding and transmission mode of the data according to the channel quality for transmission.

[0009] Further, the training of the recurrent neural network prediction model using the self-supervised mechanism to obtain the channel quality prediction model comprises:

[0010] Acquire long-time channel quality history data under a target scene, and use the channel quality history data to construct training samples of a recurrent neural network prediction model;

[0011] Take a sequence of channel quality history data of a previous time in the training samples as input, and take channel quality history data of a next time as output, and use a self-supervision mechanism to train the recurrent neural network prediction model;

[0012] Acquire a cost function of the recurrent neural network prediction model in each iteration process during the training process of the recurrent neural network prediction model, and end the training when the cost function of the recurrent neural network prediction model is less than a preset threshold, to obtain a channel quality prediction model.

[0013] Further, the self-supervision mechanism comprises the following settings:

[0014] Calculate a supervision signal used for supervising the training process of the recurrent neural network prediction model, and the calculation steps of the supervision signal are as follows:

[0015] Calculate according to a current channel quality parameter value:

[0016] E b / N0=f1(CQ(t))

[0017] wherein, E b is energy per bit, N0 is noise power density, f1 is used for mapping the channel quality parameter CQ(t) of the current time t to signal-to-noise ratio, and f1 is determined according to the type of channel quality to be predicted;

[0018] Calculate the bit error rate according to E b / N0:

[0019] BER=f2(E b / N0)

[0020] wherein, BER is the bit error rate, f2 is a bit error rate function, and the bit error rate function is determined according to the modulation / demodulation mode used;

[0021] Calculate the data packet reception rate:

[0022] PRR=(1-BER) Nbit

[0023] wherein, PRR is the data packet reception rate, and Nbit is the number of bits of the data packet;

[0024] Calculate the supervision signal:

[0025] e=PRR

[0026] wherein, e is the supervision signal.

[0027] Further, the training of the recurrent neural network prediction model comprises:

[0028] initializing the recurrent neural network prediction model, in particular:

[0029] setting the number of input layer neurons in the recurrent neural network prediction model as p1, the number of hidden layer neurons as p2, and the number of output layer neurons as p3, the number of input layer neurons and the number of output layer neurons being determined by the number of inputs and the number of outputs, and the number of hidden layer neurons p2 being calculated as follows:

[0030]

[0031] wherein a and b are parameters;

[0032] initializing the weights, in particular comprising setting the weight from the input layer neuron to the hidden layer neuron as v ij , representing the weight from the i-th input layer neuron to the j-th hidden layer neuron; setting the weight from the hidden layer neuron to the output layer neuron as ω jk , representing the weight from the j-th hidden layer neuron to the k-th output layer neuron; and setting the weight from the hidden layer neuron at the previous time to the hidden layer neuron at the current time as u jl , representing the weight from the l-th hidden layer neuron at the previous time to the j-th hidden layer neuron at the current time; and selecting the Sigmoid function as the activation function of the hidden layer neuron;

[0033] initializing the neuron nodes, and setting the j-th hidden layer neuron as:

[0034] h(j) = 0

[0035] setting the k-th output layer neuron as:

[0036] y(k) = 0

[0037] setting the target cost function as follows:

[0038]

[0039] wherein e is a supervision signal, p is a model output, H is a binary cross entropy, a is a penalty coefficient, W is a parameter set of the used neural network, and is an exclusive or logical operation, and the binary cross entropy formula is as follows:

[0040]

[0041] wherein p is a probability value vector, p m is the m-th element of the probability value vector, and the value range is [0, 1]; y is a discrete label vector, y mFor the mth element of the discrete label vector, the value is 0 or 1, and n is the length of the vector;

[0042] The training sample is input into the model, and the output of the output layer neuron is calculated as follows:

[0043] The input of the jth hidden layer neuron is:

[0044]

[0045] The output of the jth hidden layer neuron is:

[0046] h j (t) = f(net j (t))

[0047] The input of the kth output layer neuron is:

[0048]

[0049] The output of the kth output layer neuron is:

[0050] y k (t) = f(net k (t))

[0051] Where f represents the Sigmoid function.

[0052] Set the cost back propagation as the gradient descent algorithm, the learning rate is η, η>0, the momentum factor is λ, 0<λ<1, and the update formula of the connection weight is as follows:

[0053]

[0054] Where:

[0055]

[0056] Further, the channel quality parameter is predicted based on the channel quality prediction model, including:

[0057] An empty channel quality parameter sequence CQ(N) = {} is constructed.

[0058] Based on the time window T W The channel quality parameters are counted to obtain N channel quality parameter sequences [CQ(t-(N-1) x T W ),...,CQ(t-T W ),CQ(t)], t is the current time;

[0059] The channel quality parameter sequence is input into the channel quality prediction model to calculate the predicted value of the channel quality at the next time

[0060] Further, the step of comparing the predicted result of the channel quality parameter with a preset threshold, determining the channel quality according to the comparison result, and determining the encoding and transmission mode of the data according to the channel quality, comprises:

[0061] Setting a preset threshold Th1 and a preset threshold Th2 for determining the corresponding channel quality parameter, wherein Th1 is greater than Th2;

[0062] When the predicted value of the channel quality parameter is greater than Th1 and less than Th2, it is determined that the channel quality is high, and the data is directly transmitted after being check-encoded;

[0063] When the predicted value of the channel quality parameter is greater than Th2, it is determined that the channel quality is general, and the data is transmitted after being error-corrected encoded using a hybrid automatic repeat request transmission method;

[0064] When the predicted value of the channel quality parameter is less than Th2, it is determined that the channel quality is poor, and the data is encoded and transmitted using a hybrid automatic repeat request transmission method with repetition encoding.

[0065] Further, the step of directly transmitting the data after being check-encoded, comprises:

[0066] The sending end sends the data after being check-encoded;

[0067] When the receiving end successfully decodes, the receiving end sends an acknowledgement signal to the sending end to indicate correct reception, and no retransmission is needed; when the receiving end fails to decode, the receiving end sends an error signal to the sending end to indicate reception failure, and the sending end repeatedly sends the corresponding data.

[0068] Further, the step of transmitting the data after being error-corrected encoded using a hybrid automatic repeat request transmission method, comprises:

[0069] The sending end sends the data after being error-corrected encoded;

[0070] The receiving end receives and buffers the encoding, decodes according to the received error-corrected code. After successful decoding, the receiving end sends an acknowledgement signal to the sending end to indicate correct reception, and no retransmission is needed; after decoding failure, the receiving end sends an error signal to the sending end to indicate reception failure, and the sending end repeatedly sends the corresponding data.

[0071] Further, the step of encoding and transmitting the data using a hybrid automatic repeat request transmission method with repetition encoding, comprises:

[0072] The sending end error-corrects the data;

[0073] The sending end copies the same encoding and sends them together.

[0074] The receiving end receives the encoding and buffers; after the first segment is successfully decoded, the receiving end discards the repeatedly received encoding, and sends an acknowledgement signal to the sending end to indicate correct reception and no need for retransmission; after the first segment is unsuccessfully decoded, the receiving end performs Chase combining on the two segments of repeated encoding; after the two segments of repeated encoding are successfully decoded, the receiving end sends an acknowledgement signal to the sending end to indicate correct reception and no need for retransmission; after the two segments of repeated encoding are unsuccessfully decoded, the receiving end sends an error signal to the sending end to indicate reception failure, and the sending end repeatedly sends corresponding data.

[0075] Based on the same inventive concept, the application further provides a channel adaptive hybrid automatic repeat request system, comprising:

[0076] A training unit is configured to train a recurrent neural network prediction model to obtain a channel quality prediction model by using a self-supervised mechanism;

[0077] A prediction unit is configured to predict a channel quality parameter based on the channel quality prediction model;

[0078] A transmission unit is configured to compare a prediction result of the channel quality parameter with a preset threshold, determine a channel quality according to a comparison result, and determine a coding and transmission mode of data according to the channel quality to perform transmission.

[0079] Further, the training of the recurrent neural network prediction model by using the self-supervised mechanism to obtain the channel quality prediction model comprises:

[0080] Long-time channel quality historical data in a target scenario are obtained, and training samples of the recurrent neural network prediction model are constructed by using the channel quality historical data;

[0081] A sequence of channel quality historical data of a previous time in the training samples is taken as input, and channel quality historical data of a next time is taken as output, and the recurrent neural network prediction model is trained by using a self-supervised mechanism;

[0082] A cost function of the recurrent neural network prediction model in each iteration process is obtained during the training process of the recurrent neural network prediction model, and the training is ended when the cost function of the recurrent neural network prediction model is less than a preset threshold, and the channel quality prediction model is obtained.

[0083] Further, the self-supervised mechanism comprises the following settings:

[0084] A supervision signal is calculated, and the supervision signal is used to supervise the training process of the recurrent neural network prediction model, and the calculation steps of the supervision signal are as follows:

[0085] The current channel quality parameter value is calculated according to the following formula:

[0086] E b / N0=f1(CQ(t))

[0087] wherein E b is the energy per bit of the signal, N0 is the noise power density, and f1 is used to map the channel quality parameter CQ(t) at the current time t to the signal-to-noise ratio, and f1 is determined according to the type of channel quality to be predicted;

[0088] The bit error rate is calculated according to E b / N0:

[0089] BER=f2(E b / N0)

[0090] wherein BER is the bit error rate, and f2 is the bit error rate function, and the bit error rate function is determined according to the modulation / demodulation mode used;

[0091] The packet reception rate is calculated:

[0092] PRR=(1-BER) Nbit

[0093] wherein PRR is the packet reception rate, and Nbit is the number of bits of the data packet;

[0094] The supervision signal is calculated:

[0095] e=PRR

[0096] wherein e is the supervision signal.

[0097] Further, the recurrent neural network prediction model is trained, comprising:

[0098] The recurrent neural network prediction model is initialized, specifically:

[0099] The number of input layer neurons in the recurrent neural network prediction model is set to p1, the number of hidden layer neurons is set to p2, and the number of output layer neurons is set to p3, the number of input layer neurons and the number of output layer neurons are determined by the number of inputs and the number of outputs, and the number of hidden layer neurons p2 can be calculated as follows:

[0100]

[0101] wherein a and b are parameters;

[0102] The weights are initialized, specifically including setting the weight from the input layer neuron to the hidden layer neuron to v ij , representing the weight from the i-th input layer neuron to the j-th hidden layer neuron; and the weight from the hidden layer neuron to the output layer neuron is set to ω jk, represents the weight of the jth hidden layer neuron to the kth output layer neuron; the weight of the previous time hidden layer neuron to the current time hidden layer neuron is u jl , represents the weight of the previous time lth hidden layer neuron to the current time jth hidden layer neuron; the Sigmoid function is selected as the activation function of the hidden layer neuron;

[0103] Initialize the neuron node, set the jth hidden layer neuron as:

[0104] h(j) = 0

[0105] Initialize the kth output layer neuron as:

[0106] y(k) = 0

[0107] Set the target cost function as follows:

[0108]

[0109] Where e is the supervision signal, p is the model output, H is the binary cross entropy, a is the penalty coefficient, W is the parameter set of the used neural network, and is the exclusive or logical operation. The binary cross entropy formula is as follows:

[0110]

[0111] Where p is the probability value vector, p m is the mth element of the probability value vector, with a value range of [0, 1]; y is the discrete label vector, y m is the mth element of the discrete label vector, with a value of 0 or 1, and n is the vector length.

[0112] Input the training sample into the model, and calculate the output layer neuron output as follows:

[0113] The input of the jth hidden layer neuron is:

[0114]

[0115] The output of the jth hidden layer neuron is:

[0116] h j (t) = f(net j (t))

[0117] The input of the kth output layer neuron is:

[0118]

[0119] The output of the kth output layer neuron is:

[0120] yk (t) = f(net k (t))

[0121] Where f represents the Sigmoid function;

[0122] Set the cost back propagation to the gradient descent algorithm, the learning rate to η, η>0, the momentum factor to λ, 0<λ<1, and the update formula of the connection weight is as follows:

[0123]

[0124] in:

[0125]

[0126] Furthermore, the channel quality parameter is predicted based on the channel quality prediction model, including:

[0127] Construct an empty channel quality parameter sequence CQ(N)={};

[0128] Based on the time window T W The channel quality parameters are counted to obtain N channel quality parameter sequences [CQ(t-(N-1)×T W ),...,CQ(tT W ), CQ(t)], t is the current moment;

[0129] Input the channel quality parameter sequence into the channel quality prediction model to calculate the predicted value of the channel quality at the next moment

[0130]

[0131] Furthermore, the step of comparing the predicted result of the channel quality parameter with a preset threshold, determining the channel quality based on the comparison result, and determining the encoding and transmission mode of the data for transmission based on the channel quality includes:

[0132] Setting a preset threshold Th1 and a preset threshold Th2 for determining a corresponding channel quality parameter, wherein Th1 is greater than Th2;

[0133] When the channel quality parameter prediction value When the channel quality is determined to be high, the data is directly transmitted after verification encoding;

[0134] When the channel quality parameter prediction value When the channel quality is judged to be average, the hybrid automatic repeat request transmission method is used to perform error correction coding on the data before transmission;

[0135] When the channel quality parameter prediction value When the channel quality is determined to be poor, a hybrid automatic repeat request transmission method using repetition coding is used to encode and transmit the data.

[0136] Further, the encoding and transmitting the data directly includes:

[0137] The data is encoded and transmitted by the sending end.

[0138] When the receiving end successfully decodes, the receiving end sends an acknowledgement signal to the sending end to indicate correct reception, and no retransmission is needed; when the receiving end fails to decode, the receiving end sends an error signal to the sending end to indicate reception failure, and the sending end repeatedly transmits corresponding data.

[0139] Further, the encoding and transmitting the data directly includes:

[0140] The data is encoded and transmitted by the sending end.

[0141] The receiving end receives and buffers the encoding and decodes according to the received error correction code; when the receiving end successfully decodes, the receiving end sends an acknowledgement signal to the sending end to indicate correct reception, and no retransmission is needed; when the receiving end fails to decode, the receiving end sends an error signal to the sending end to indicate reception failure, and the sending end repeatedly transmits corresponding data.

[0142] Further, the encoding and transmitting the data directly includes:

[0143] The data is encoded and transmitted by the sending end.

[0144] The sending end duplicates the same encoding and transmits them together.

[0145] The receiving end receives and buffers the encoding; when the receiving end successfully decodes the first segment of the encoding, the receiving end discards the repeatedly received encoding and sends an acknowledgement signal to the sending end to indicate correct reception, and no retransmission is needed; when the receiving end fails to decode the first segment of the encoding, the receiving end performs Chase combining on the two segments of the repeated encoding; when the two segments of the encoding are successfully decoded, the receiving end sends an acknowledgement signal to the sending end to indicate correct reception, and no retransmission is needed; when the two segments of the encoding fail to be decoded, the receiving end sends an error signal to the sending end to indicate reception failure, and the sending end repeatedly transmits corresponding data.

[0146] Based on the same inventive concept, the application also provides a computer device, comprising: one or more processors; the processor is used to store one or more programs; when the one or more programs are executed by the one or more processors, the channel adaptive hybrid automatic repeat request method according to any one of the above is realized.

[0147] Based on the same inventive concept, the application also provides a computer readable storage medium, which has a computer program stored thereon, the computer program, when executed, implements the channel adaptive hybrid automatic repeat request method according to any one of the above.

[0148] Beneficial effects: compared with the prior art, the application realizes the channel adaptive hybrid automatic repeat request method for high-reliability and low-latency services, schedules the transmission mode according to the prediction result of the channel quality, and maximally reduces the processing and communication delay while meeting the reliability requirement; the method provided by the application firstly predicts the channel quality parameter based on the self-supervised recurrent neural network to obtain the evaluation of the channel quality; secondly, the encoding and transmission mode of data are determined according to the prediction result of the channel quality, and the effect of high reliability and low latency is achieved. BRIEF DESCRIPTION OF DRAWINGS

[0149] Figure 1 The method flowchart of the embodiment of the application is shown in the figure;

[0150] Figure 2 The self-supervised training and prediction schematic diagram of the channel quality prediction model of the embodiment of the application is shown in the figure;

[0151] Figure 3 The transmission mode decision and encoding flowchart of the embodiment of the application is shown in the figure;

[0152] Figure 4 The first transmission success probability schematic diagram of different methods of the embodiment of the application is shown in the figure;

[0153] Figure 5 The average retransmission number schematic diagram of different methods of the embodiment of the application is shown in the figure;

[0154] Figure 6 The frame error rate schematic diagram of different methods of the embodiment of the application is shown in the figure. DETAILED DESCRIPTION

[0155] In order to enable personnel in the art to better understand the scheme of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application.

[0156] As shown in the figure, Figure 1 the channel adaptive hybrid automatic repeat request method of the embodiment includes:

[0157] S1: training a recurrent neural network prediction model by using a self-supervised mechanism to obtain a channel quality prediction model;

[0158] S2: predicting a channel quality parameter based on the channel quality prediction model;

[0159] S3: comparing the prediction result of the channel quality parameter with a preset threshold, determining the channel quality according to the comparison result, and determining the encoding and transmission mode of the data according to the channel quality to perform transmission.

[0160] Specifically, the signal-to-noise ratio (SNR) is selected as the channel quality parameter for prediction.

[0161] The step S1 of training the recurrent neural network prediction model by using the self-supervised mechanism to obtain the channel quality prediction model includes:

[0162] Obtain long-time channel quality measurement data in a target scene to construct a recurrent neural network prediction model training sample;

[0163] Take a continuous N channel quality history sequence as input and the channel quality at the next time as output, and train the recurrent neural network prediction model by using a self-supervised method;

[0164] When the cost function of the prediction model meets the requirements, the training is ended, and the prediction model parameters are obtained.

[0165] The flow of the running phase of the channel quality prediction model in the step S2 includes:

[0166] Construct an empty SNR parameter sequence S(N) = {};

[0167] Every time window T W Statistically obtain a sequence [SNR(t-(N-1)XT W ),...S,NR(t-T W )S,NR(t)] containing N SNR parameters, wherein t is the current time;

[0168] Calculate the SNR prediction value at the next time according to the channel quality measurement model

[0169] The step S1 includes the training of the channel quality prediction model for SNR, and has the following initial settings:

[0170] Calculate the supervision signal: the supervision signal is used to supervise the process of model training, and the calculation steps are as follows: first, calculate E b / N0 according to the SNR prediction value:

[0171]

[0172] Wherein, E b is the energy per bit signal, N0 is the noise power density, R is the symbol rate, and BW is the channel bandwidth;

[0173] Then, the bit error rate is calculated according to E b / N0:

[0174]

[0175] wherein BER is the bit error rate, the signal modulation mode is DSSS-OQPSK, and Q is the complementary cumulative distribution function;

[0176] Next, the data packet reception rate is calculated:

[0177] PRR = (1-BER) Nbit

[0178] wherein PRR is the data packet reception rate, and Nbit is the number of bits;

[0179] Finally, the supervision signal is calculated:

[0180] e = PRR

[0181] wherein e is the supervision signal, and PRR is taken as the supervision signal;

[0182] Parameter setting: let the number of input layer neurons be p1, the number of hidden layer neurons be p2, and the number of output layer neurons be p3. The number of input layer and output layer neurons is determined by the number of inputs and outputs, and the number of hidden layer neurons p2 can be calculated as follows:

[0183]

[0184] The target cost function is set as follows:

[0185]

[0186] wherein e is the supervision signal, p is the model output, H is the binary cross entropy, λ is the penalty coefficient, W is the parameter set of the used neural network, and is the exclusive or logical operation. The binary cross entropy formula is as follows:

[0187]

[0188] wherein p is the probability value vector, p m is the mth element of the probability value vector, and the value range is [0, 1]. y is the discrete label vector, y m is the mth element of the discrete label vector, and the value is 0 or 1. n is the vector length.

[0189] Weight initialization: the weight from the input layer to the hidden layer is set as v ij , which represents the weight from the ith neuron of the input layer to the jth neuron of the hidden layer; and the weight from the hidden layer to the output layer is ω jk, represents the weight of the jth neuron of the hidden layer to the kth neuron of the output layer; the weight of the hidden layer at the last time to the hidden layer at the current time is u jl , represents the weight of the lth neuron of the hidden layer at the last time to the jth neuron of the hidden layer at the current time; the Sigmoid function is selected as the activation function of the hidden layer

[0190] The neuron node is initialized, and the jth hidden layer neuron is initialized as follows:

[0191] h(j) = 0

[0192] The kth output layer neuron is initialized as follows:

[0193] y(k) = 0

[0194] Step S1 includes training of the channel quality prediction model for SNR:

[0195] Set the training threshold;

[0196] Input the training sample into the model, and calculate the output of the output layer neuron as follows:

[0197] The input of the jth hidden layer neuron is:

[0198]

[0199] The output of the jth hidden layer neuron is:

[0200] h j (t) = f(net j (t))

[0201] The input of the kth output layer neuron is:

[0202]

[0203] The output of the kth output layer neuron is:

[0204] y k (t) = f(net k (t))

[0205] Where f represents the Sigmoid function

[0206] Set the cost back propagation as the gradient descent algorithm, the learning rate as η (η>0), and the momentum factor as λ (0<λ<1) to avoid the prediction model from falling into local optimum. The update formula of the connection weight is as follows:

[0207]

[0208] Where:

[0209]

[0210] The self-supervised training and prediction of the channel quality prediction model of this embodiment are as follows Figure 2 shown.

[0211] Determining the data encoding and transmission mode according to the channel quality in step S3 includes:

[0212] Determine the quality of the current channel and confirm the transmission mode. The specific steps are as follows:

[0213] When the signal-to-noise ratio prediction value When the channel quality is high, the data is directly transmitted after verification coding, as shown in the diagram. Figure 3 As shown in the figure, where Th1 is the SNR threshold corresponding to a lower bit error rate. The specific transmission steps are as follows:

[0214] 1) The sending end performs a simple checksum on the data before sending it;

[0215] 2) When the receiving end successfully decodes, the receiving end sends a confirmation signal to the sending end to indicate that it has been received correctly and no retransmission is required; when the receiving end fails to decode, the receiving end sends an error signal to the sending end to indicate that the reception has failed, and the sending end repeats the corresponding data.

[0216] When the signal-to-noise ratio prediction value When the channel quality is average, the hybrid automatic request and repeat transmission method is used to perform error correction coding on the data before transmission, as shown in the diagram. Figure 3 The specific transmission steps are as follows:

[0217] 1) The sending end performs error correction encoding on the data before sending it;

[0218] 2) The receiving end receives the encoded data and buffers it, then decodes it based on the received error-correction code. Upon successful decoding, the receiving end sends an acknowledgment signal to the sending end, indicating successful reception and no further retransmission is required. If decoding fails, the receiving end sends an error signal to the sending end, indicating the failure, and the sending end retransmits the corresponding data.

[0219] When the signal-to-noise ratio prediction value When the channel quality is poor, the data is encoded and transmitted using a hybrid automatic request retransmission method with repeated coding, as shown in the diagram. Figure 3 The specific transmission steps are as follows:

[0220] 1) The sending end performs error correction coding on the data;

[0221] 2) The sender copies the same code and sends it together;

[0222] 3) receiving end receives encoding and buffering. After the receiving end successfully decodes the first segment encoding, the receiving end discards the repeatedly received encoding and sends an acknowledgement signal to the sending end to indicate correct reception and no longer needs retransmission. After the receiving end fails to decode the first segment encoding, the receiving end performs Chase combining on the two repeated encodings. After successfully decoding the combined two encodings, the receiving end sends an acknowledgement signal to the sending end to indicate correct reception and no longer needs retransmission. After failing to decode the combined two encodings, the receiving end sends an error signal to the sending end to indicate reception failure, and the sending end repeatedly sends the corresponding data.

[0223] In order to verify the reliability of the method, the polar code is used as error correction coding for verification, and the specific steps are as follows:

[0224] For the construction of the polar code of the polar channel, the specific steps are as follows:

[0225] For the polar code with length N=2 n (n≥0), the following formula is used for calculation:

[0226]

[0227] Wherein, is the bit sequence after polar code encoding, is the uncoded bit sequence, G N is the generating sequence. Wherein, G N is calculated by the following formula:

[0228]

[0229] Wherein, B N is the bit inversion of the uncoded bit sequence, is the n-th Kronecker power of F2. Wherein,

[0230] For the transition probability of the polar code of the polar channel, the specific steps are as follows:

[0231] For a specific modulation and demodulation mode, the SNR of the channel W and the bit error rate have a fixed relationship:

[0232]

[0233] Wherein, B M (·) is the function of the bit error rate of the specific modulation and demodulation mode, E b is the energy per bit signal, N0 is the noise power density, R b is the data transmission quantity, B N is the noise bandwidth.

[0234] For the polar code of the polar channel with length N=2 nA polar code with n >= 0, and the channel capacity of each polar channel is:

[0235]

[0236]

[0237] wherein, 2 n = 1,

[0238] In order to verify the reliability of the method, the following simulation settings are carried out: the length of the polar code is set to 64 bits, the code rate is set to 0.5, the maximum number of retransmissions is set to two, and the modulation and demodulation mode is set to BPSK. More than one bit of a frame of data (64 bits) is not successfully decoded, and the success probability of the proposed method and the existing BPSK modulation and demodulation method direct transmission, the first transmission success probability of the existing polar code encoding, the data retransmission number, and the frame error rate are respectively shown in Figure 4 、 Figure 5 and Figure 6

[0239] It can be seen that the polar code encoding transmission has a gain of about 10 dB compared with the direct transmission; and the proposed method has a gain of about 4 dB compared with the polar code transmission. The first transmission success probability, the data retransmission number, and the frame error rate of the proposed method are better than those of the existing method.

[0240] Based on the same inventive concept, the embodiment also provides a channel adaptive hybrid automatic repeat request system, comprising:

[0241] A training unit is configured to train a recurrent neural network prediction model to obtain a channel quality prediction model by using a self-supervised mechanism;

[0242] A prediction unit is configured to predict a channel quality parameter based on the channel quality prediction model;

[0243] A transmission unit is configured to compare the prediction result of the channel quality parameter with a preset threshold, determine the channel quality according to the comparison result, and determine the encoding and transmission mode of the data according to the channel quality for transmission.

[0244] Further, the training of the recurrent neural network prediction model by using the self-supervised mechanism to obtain the channel quality prediction model comprises:

[0245] Obtaining long-time channel quality historical data in a target scene, and constructing training samples of the recurrent neural network prediction model by using the channel quality historical data;

[0246] ​The sequence of channel quality historical data of a plurality of previous time points in the training sample is taken as input, and the channel quality historical data of the next time point is taken as output, and a self-supervised mechanism is used to train the recurrent neural network prediction model.

[0247] In the training process of the recurrent neural network prediction model, the cost function of the recurrent neural network prediction model in each iteration process is obtained, and when the cost function of the recurrent neural network prediction model is less than a preset threshold, the training is ended, and the channel quality prediction model is obtained.

[0248] Further, the self-supervised mechanism comprises the following settings:

[0249] A supervision signal is calculated, which is used to supervise the training process of the recurrent neural network prediction model, and the supervision signal calculation step is as follows:

[0250] According to the current channel quality parameter value, the following is calculated:

[0251] E b / N0=f1(CQ(t))

[0252] Wherein, E b is the energy per bit of signal, N0 is the noise power density, f1 is used to map the channel quality parameter CQ(t) of the current time t to the signal-to-noise ratio, and f1 is determined according to the type of channel quality to be predicted;

[0253] According to E b / N0, the bit error rate is calculated:

[0254] BER=f2(E b / N0)

[0255] Wherein, BER is the bit error rate, f2 is the bit error rate function, and the bit error rate function is determined according to the modulation / demodulation mode used;

[0256] The data packet reception rate is calculated:

[0257] PRR=(1-BER) Nbit

[0258] Wherein, PRR is the data packet reception rate, and Nbit is the number of bits of the data packet;

[0259] The supervision signal is calculated:

[0260] e=PRR

[0261] Wherein, e is the supervision signal.

[0262] Further, the training of the recurrent neural network prediction model comprises:

[0263] Initialize the recurrent neural network prediction model, specifically:

[0264] Set the number of input layer neurons in the recurrent neural network prediction model as p1, the number of hidden layer neurons as p2, and the number of output layer neurons as p3. The number of input layer neurons and output layer neurons are determined by the number of inputs and outputs, and the number of hidden layer neurons p2 can be calculated as follows:

[0265]

[0266] Where a and b are parameters;

[0267] Initialize the weights, specifically including setting the weight from the input layer neuron to the hidden layer neuron as v ij , representing the weight from the i-th input layer neuron to the j-th hidden layer neuron; the weight from the hidden layer neuron to the output layer neuron as ω jk , representing the weight from the j-th hidden layer neuron to the k-th output layer neuron; the weight from the previous time hidden layer neuron to the current time hidden layer neuron as u jl , representing the weight from the l-th hidden layer neuron at the previous time to the j-th hidden layer neuron at the current time; and selecting the Sigmoid function as the activation function of the hidden layer neuron;

[0268] Initialize the neuron nodes, and set the j-th hidden layer neuron as:

[0269] h(j) = 0

[0270] Initialize the k-th output layer neuron as:

[0271] y(k) = 0

[0272] Set the target cost function as follows:

[0273]

[0274] Where e is the supervision signal, p is the model output, H is the binary cross-entropy, α is the penalty coefficient, W is the parameter set of the used neural network, and is the exclusive or logical operation. The binary cross-entropy formula is as follows:

[0275]

[0276] Where p is the probability value vector, p m is the m-th element of the probability value vector, with a value range of [0, 1]; y is the discrete label vector, y m is the m-th element of the discrete label vector, with a value of 0 or 1, and n is the vector length;

[0277] Input the training sample into the model, and calculate the output of the output layer neuron as follows:

[0278] The input of the jth hidden layer neuron is:

[0279]

[0280] The output of the jth hidden layer neuron is:

[0281] h j (t) = f(net j (t))

[0282] The input of the kth output layer neuron is:

[0283]

[0284] The output of the kth output layer neuron is:

[0285] y k (t) = f(net k (t))

[0286] Wherein, f represents a Sigmoid function;

[0287] The cost back propagation is set as a gradient descent algorithm, the learning rate is η, η>0, the momentum factor is λ, 0<λ<1, and the update formula of the connection weight is as follows:

[0288]

[0289] Wherein:

[0290]

[0291] Further, the channel quality parameter is predicted based on the channel quality prediction model, including:

[0292] An empty channel quality parameter sequence CQ(N) = {} is constructed;

[0293] Based on a time window T W The channel quality parameters are counted to obtain N channel quality parameter sequences [CQ(t-(N-1)×T W ),...,CQ(t-T W ),CQ(t)], t is the current time;

[0294] The channel quality parameter sequence is input into the channel quality prediction model to calculate the predicted value of the channel quality at the next time

[0295]

[0296] Further, the method comprises: comparing the prediction result of the channel quality parameter with a preset threshold value; determining the channel quality according to the comparison result; and determining the encoding and transmission mode of the data according to the channel quality.

[0297] The preset threshold value Th1 and the preset threshold value Th2 corresponding to the channel quality parameter are set, wherein Th1 is greater than Th2;

[0298] When the channel quality parameter prediction value is greater than Th1 and less than Th2, it is determined that the channel quality is high, and the data is directly transmitted after being check-encoded;

[0299] When the channel quality parameter prediction value is less than Th2, it is determined that the channel quality is general, and the data is transmitted after being error-corrected encoded by using a hybrid automatic repeat request transmission method;

[0300] When the channel quality parameter prediction value is less than Th1, it is determined that the channel quality is poor, and the data is encoded and transmitted by using a hybrid automatic repeat request transmission method with repetition encoding.

[0301] Further, the method comprises:

[0302] The data is sent after being check-encoded by the sending end;

[0303] When the receiving end successfully decodes, the receiving end sends an acknowledgement signal to the sending end to indicate correct reception, and retransmission is not needed; when the receiving end fails to decode, the receiving end sends an error signal to the sending end to indicate reception failure, and the sending end repeatedly sends corresponding data.

[0304] Further, the method comprises:

[0305] The data is sent after being error-corrected encoded by the sending end;

[0306] The receiving end receives and buffers the encoding and decodes according to the received error-correcting code; when the receiving end successfully decodes, the receiving end sends an acknowledgement signal to the sending end to indicate correct reception, and retransmission is not needed; when the receiving end fails to decode, the receiving end sends an error signal to the sending end to indicate reception failure, and the sending end repeatedly sends corresponding data.

[0307] Further, the method comprises:

[0308] The data is error-corrected encoded by the sending end;

[0309] The same encoding is copied by the sending end and sent together; ​​​

[0310] The receiving end receives and buffers the encoding; after the first segment of the encoding is successfully decoded by the receiving end, the receiving end discards the repeatedly received encoding and sends an acknowledgement signal to the sending end to indicate correct reception and no need for retransmission; after the two segments of the encoding are unsuccessfully decoded by the receiving end, the receiving end performs Chase combining on the two segments of the encoding; after the two segments of the encoding are successfully decoded after being combined, the receiving end sends an acknowledgement signal to the sending end to indicate correct reception and no need for retransmission; after the two segments of the encoding are unsuccessfully decoded after being combined, the receiving end sends an error signal to the sending end to indicate reception failure, and the sending end repeatedly sends corresponding data.

[0311] Based on the same inventive concept, the embodiment further provides a computer device, comprising: one or more processors; the processor is configured to store one or more programs; when the one or more programs are executed by the one or more processors, the channel adaptive hybrid automatic repeat request method according to any one of the above is implemented.

[0312] Based on the same inventive concept, the embodiment further provides a computer readable storage medium, which has a computer program stored thereon, and the computer program is executed to implement the channel adaptive hybrid automatic repeat request method according to any one of the above.

Claims

1. A channel-adaptive hybrid automatic repeat request method, characterized in that: include: The channel quality prediction model is obtained by training the recurrent neural network prediction model using a self-supervisory mechanism; including: Obtain long-term historical channel quality data in the target scenario and use it to construct training samples for the recurrent neural network prediction model; The self-supervisory mechanism is used to train the recurrent neural network prediction model, taking a sequence of channel quality history data of several consecutive previous moments in the training sample as input and the channel quality history data of the next moment as output; During the training process of the recurrent neural network prediction model, a cost function of the recurrent neural network prediction model is obtained in each iteration process. When the cost function of the recurrent neural network prediction model is less than a preset threshold, the training is terminated to obtain a channel quality prediction model. The self-supervision mechanism includes the following settings: Calculate the supervisory signal, which is used to supervise the training process of the recurrent neural network prediction model. The supervisory signal calculation steps are as follows: Calculated based on the current channel quality parameter value: E b / N0=f1(CQ(t)) Among them, E b is the signal energy per bit, N0 is the noise power density, and f1 is used to map the channel quality parameter CQ(t) at the current time t to the signal-to-noise ratio. f1 is determined according to the channel quality type to be predicted; According to E b / N0 calculates the bit error rate: BER=f2(E b / N0) Where BER is the bit error rate, f2 is the bit error rate function, and the bit error rate function is determined by the modulation / demodulation method used; Calculate the packet reception rate: PRR=(1-BER) Nbit Where PRR is the packet reception rate, Nbit is the number of bits in the packet; Calculate the supervisory signal: e=PRR Among them, e is the supervision signal; Predicting channel quality parameters based on a channel quality prediction model; The predicted result of the channel quality parameter is compared with a preset threshold, the channel quality is determined according to the comparison result, and the data encoding and transmission mode are determined according to the channel quality for transmission.

2. The channel-adaptive hybrid automatic repeat request method according to claim 1, wherein: The training of the recurrent neural network prediction model includes: Initialize the recurrent neural network prediction model, specifically: Set the number of input layer neurons in the recurrent neural network prediction model to p1, the number of hidden layer neurons to p2, and the number of output layer neurons to p3. The number of input layer neurons and the number of output layer neurons are determined by the number of inputs and outputs. The number of hidden layer neurons p2 can be calculated as follows: Among them, a and b are parameters; Initialize the weights, specifically setting the weights from the input layer neurons to the hidden layer neurons to v ij , represents the weight from the i-th input layer neuron to the j-th hidden layer neuron; the weight from the hidden layer neuron to the output layer neuron is ω jk , represents the weight from the jth hidden layer neuron to the kth output layer neuron; the weight from the previous hidden layer neuron to the current hidden layer neuron is u jl , represents the weight from the lth hidden layer neuron at the previous moment to the jth hidden layer neuron at the current moment; the Sigmoid function is selected as the activation function of the hidden layer neurons; Initialize the neuron nodes and set the jth hidden layer neuron to: h(j)=0 Initialize the kth output layer neuron as: y(k)=0 The target cost function is set as follows: Among them, e is the supervision signal, p is the model output, H is the binary cross entropy, α is the penalty coefficient, W is the parameter set of the neural network used, ⊙ is the exclusive OR logical operation, and the binary cross entropy formula is as follows: Among them, p is the probability value vector, p m is the mth element of the probability value vector, ranging from [0,1]; y is the discrete label vector, y m is the mth element of the discrete label vector, with a value of 0 or 1, and n is the length of the vector; The training samples are input into the model and the output of the output layer neurons is calculated as follows: The input to the jth hidden layer neuron is: The output of the jth hidden layer neuron is: h j (t)=f(net j (t)) The input to the kth output layer neuron is: The output of the kth output layer neuron is: y k (t)=f(net k (t)) Where f represents the Sigmoid function; Set the cost back propagation to the gradient descent algorithm, the learning rate to η, η>0, the momentum factor to λ, 0<λ<1, and the update formula of the connection weight is as follows: in:

3. The channel-adaptive hybrid automatic repeat request method according to claim 1, wherein: The channel quality prediction model is used to predict channel quality parameters, including: Construct an empty channel quality parameter sequence CQ(N)={}; Based on the time window T W The channel quality parameters are counted to obtain N channel quality parameter sequences [CQ(t-(N-1)×T W ),...,CQ(tT W ), CQ(t)], t is the current moment; Input the channel quality parameter sequence into the channel quality prediction model to calculate the predicted value of the channel quality at the next moment 4. The channel-adaptive hybrid automatic repeat request method according to claim 1, wherein: The step of comparing the predicted result of the channel quality parameter with a preset threshold, determining the channel quality based on the comparison result, and determining the encoding and transmission mode of the data for transmission based on the channel quality includes: Setting a preset threshold Th1 and a preset threshold Th2 for determining a corresponding channel quality parameter, wherein Th1 is greater than Th2; When the channel quality parameter prediction value When the channel quality is determined to be high, the data is directly transmitted after verification encoding; When the channel quality parameter prediction value When the channel quality is judged to be average, the hybrid automatic repeat request transmission method is used to perform error correction coding on the data before transmission; When the channel quality parameter prediction value When the channel quality is determined to be poor, the data is encoded and transmitted using a hybrid automatic repeat request transmission method with repetition coding.

5. The channel-adaptive hybrid automatic repeat request method according to claim 4, wherein: The step of directly transmitting the data after performing verification encoding comprises: The sending end verifies and encodes the data before sending it; When the receiving end successfully decodes, the receiving end sends a confirmation signal to the sending end to indicate that the data is received correctly and no retransmission is required; when the receiving end fails to decode, the receiving end sends an error signal to the sending end to indicate that the data is received incorrectly and the sending end repeats the transmission of the corresponding data.

6. The channel-adaptive hybrid automatic repeat request method according to claim 4, wherein: The hybrid automatic repeat request transmission method is used to transmit data after error correction coding, including: The sending end performs error correction coding on the data before sending it; The receiving end receives the code and caches it, and decodes it according to the received error correction code. After successful decoding, the receiving end sends a confirmation signal to the sending end to indicate that it has been received correctly and no retransmission is required. If the receiving end fails to decode, the receiving end sends an error signal to the sending end to indicate that the reception has failed, and the sending end repeats the corresponding data.

7. The channel-adaptive hybrid automatic repeat request method according to claim 4, wherein: The hybrid automatic repeat request transmission method using repetition coding encodes and transmits data, including: The sending end performs error correction coding on the data; The sender copies the same code and sends it together; The receiving end receives the code and caches it; after successfully decoding the first segment of the code, the receiving end discards the repeated code and sends a confirmation signal to the sending end to indicate that it is correctly received and no longer needs to be retransmitted; after the receiving end fails to decode the first segment of the code, the receiving end performs a Zeiss merger on the two repeated segments of the code; after successfully decoding the merged two segments of the code, the receiving end sends a confirmation signal to the sending end to indicate that it is correctly received and no longer needs to be retransmitted; after failing to decode the merged two segments of the code, the receiving end sends an error signal to the sending end to indicate that the reception failed, and the sending end repeats the corresponding data.

8. A channel-adaptive hybrid automatic repeat request system, characterized in that: include: A training unit is used to train the recurrent neural network prediction model using a self-supervision mechanism to obtain a channel quality prediction model; it includes: Obtain long-term historical channel quality data in the target scenario and use it to construct training samples for the recurrent neural network prediction model; The self-supervisory mechanism is used to train the recurrent neural network prediction model, taking a sequence of channel quality history data of several consecutive previous moments in the training sample as input and the channel quality history data of the next moment as output; During the training process of the recurrent neural network prediction model, a cost function of the recurrent neural network prediction model is obtained in each iteration process. When the cost function of the recurrent neural network prediction model is less than a preset threshold, the training is terminated to obtain a channel quality prediction model. The self-supervision mechanism includes the following settings: Calculate the supervisory signal, which is used to supervise the training process of the recurrent neural network prediction model. The supervisory signal calculation steps are as follows: Calculated based on the current channel quality parameter value: E b / N0=f1(CQ(t)) Among them, E b is the signal energy per bit, N0 is the noise power density, and f1 is used to map the channel quality parameter CQ(t) at the current time t to the signal-to-noise ratio. f1 is determined according to the channel quality type to be predicted; According to E b / N0 calculates the bit error rate: BER=f2(E b / N0) Where BER is the bit error rate, f2 is the bit error rate function, and the bit error rate function is determined by the modulation / demodulation method used; Calculate the packet reception rate: PRR=(1-BER) Nbit Where PRR is the packet reception rate, Nbit is the number of bits in the packet; Calculate the supervisory signal: e=PRR Among them, e is the supervision signal; A prediction unit, configured to predict a channel quality parameter based on a channel quality prediction model; The transmission unit is used to compare the prediction result of the channel quality parameter with a preset threshold, determine the channel quality according to the comparison result, and determine the coding and transmission mode of the data for transmission according to the channel quality.

9. The channel-adaptive hybrid automatic repeat request system according to claim 8, wherein: The training of the recurrent neural network prediction model includes: Initialize the recurrent neural network prediction model, specifically: Set the number of input layer neurons in the recurrent neural network prediction model to p1, the number of hidden layer neurons to p2, and the number of output layer neurons to p3. The number of input layer neurons and the number of output layer neurons are determined by the number of inputs and outputs. The number of hidden layer neurons p2 can be calculated as follows: Among them, a and b are parameters; Initialize the weights, specifically setting the weights from the input layer neurons to the hidden layer neurons to v ij , represents the weight from the i-th input layer neuron to the j-th hidden layer neuron; the weight from the hidden layer neuron to the output layer neuron is ω jk , represents the weight from the jth hidden layer neuron to the kth output layer neuron; the weight from the previous hidden layer neuron to the current hidden layer neuron is u jl , represents the weight from the lth hidden layer neuron at the previous moment to the jth hidden layer neuron at the current moment; the Sigmoid function is selected as the activation function of the hidden layer neurons; Initialize the neuron nodes and set the jth hidden layer neuron to: h(j)=0 Initialize the kth output layer neuron as: y(k)=0 The target cost function is set as follows: Among them, e is the supervision signal, p is the model output, H is the binary cross entropy, α is the penalty coefficient, W is the parameter set of the neural network used, ⊙ is the exclusive OR logical operation, and the binary cross entropy formula is as follows: Among them, p is the probability value vector, p m is the mth element of the probability value vector, ranging from [0,1]; y is the discrete label vector, y m is the mth element of the discrete label vector, with a value of 0 or 1, and n is the length of the vector; The training samples are input into the model and the output of the output layer neurons is calculated as follows: The input to the jth hidden layer neuron is: The output of the jth hidden layer neuron is: h j (t)=f(net j (t)) The input to the kth output layer neuron is: The output of the kth output layer neuron is: y k (t)=f(net k (t)) Where f represents the Sigmoid function; Set the cost back propagation to the gradient descent algorithm, the learning rate to η, η>0, the momentum factor to λ, 0<λ<1, and the update formula of the connection weight is as follows: in:

10. The channel-adaptive hybrid automatic repeat request system according to claim 8, wherein: The channel quality prediction model is used to predict channel quality parameters, including: Construct an empty channel quality parameter sequence CQ(N)={}; Based on the time window T W The channel quality parameters are counted to obtain N channel quality parameter sequences [CQ(t-(N-1)×T W ),...,CQ(tT W ), CQ(t)], t is the current moment; Input the channel quality parameter sequence into the channel quality prediction model to calculate the predicted value of the channel quality at the next moment 11. The channel-adaptive hybrid automatic repeat request system according to claim 8, wherein: The step of comparing the predicted result of the channel quality parameter with a preset threshold, determining the channel quality based on the comparison result, and determining the encoding and transmission mode of the data for transmission based on the channel quality includes: Setting a preset threshold Th1 and a preset threshold Th2 for determining a corresponding channel quality parameter, wherein Th1 is greater than Th2; When the channel quality parameter prediction value When the channel quality is determined to be high, the data is directly transmitted after verification encoding; When the channel quality parameter prediction value When the channel quality is judged to be average, the hybrid automatic repeat request transmission method is used to perform error correction coding on the data before transmission; When the channel quality parameter prediction value When the channel quality is determined to be poor, the data is encoded and transmitted using a hybrid automatic repeat request transmission method with repetition coding.

12. The channel-adaptive hybrid automatic repeat request system according to claim 11, wherein: The step of directly transmitting the data after performing verification encoding comprises: The sending end verifies and encodes the data before sending it; When the receiving end successfully decodes, the receiving end sends a confirmation signal to the sending end to indicate that the data is received correctly and no retransmission is required; when the receiving end fails to decode, the receiving end sends an error signal to the sending end to indicate that the data is received incorrectly and the sending end repeats the transmission of the corresponding data.

13. The channel-adaptive hybrid automatic repeat request system according to claim 11, wherein: The hybrid automatic repeat request transmission method is used to transmit data after error correction coding, including: The sending end performs error correction coding on the data before sending it; The receiving end receives the code and caches it, and decodes it according to the received error correction code. After successful decoding, the receiving end sends a confirmation signal to the sending end to indicate that it has been received correctly and no retransmission is required. If the receiving end fails to decode, the receiving end sends an error signal to the sending end to indicate that the reception has failed, and the sending end repeats the corresponding data.

14. The channel-adaptive hybrid automatic repeat request system according to claim 11, wherein: The hybrid automatic repeat request transmission method using repetition coding encodes and transmits data, including: The sending end performs error correction coding on the data; The sender copies the same code and sends it together; The receiving end receives the code and caches it; after successfully decoding the first segment of the code, the receiving end discards the repeated code and sends a confirmation signal to the sending end to indicate that it is correctly received and no longer needs to be retransmitted; after the receiving end fails to decode the first segment of the code, the receiving end performs a Zeiss merger on the two repeated segments of the code; after successfully decoding the merged two segments of the code, the receiving end sends a confirmation signal to the sending end to indicate that it is correctly received and no longer needs to be retransmitted; after failing to decode the merged two segments of the code, the receiving end sends an error signal to the sending end to indicate that the reception failed, and the sending end repeats the corresponding data.

15. A computer device, characterized in that: include: one or more processors; The processor is configured to store one or more programs; When the one or more programs are executed by the one or more processors, the channel-adaptive hybrid automatic repeat request method according to any one of claims 1 to 7 is implemented.

16. A computer-readable storage medium, characterized in that A computer program is stored thereon, and when the computer program is executed, the channel adaptive hybrid automatic repeat request method according to any one of claims 1 to 7 is implemented.

Images