An adaptive coding modulation method and system based on expired channel

Abstract

The present invention relates to an adaptive coding modulation method and system based on expired channels, comprising: a receiving end periodically estimates a channel vector and calculates SNR, then maps SNR to CQI based on a channel state index value table and feeds back to a transmitting end; the transmitting end uses the feedback information of the receiving end as input, adopts a DQN algorithm to select the coding modulation level of the current time slot, then adopts an OLLA algorithm to correct the selected coding modulation level, and finally transmits the signal to the receiving end. After receiving the signal, the receiving end demodulates and calculates BLER, and then feeds back the signal demodulation result and BLER to the transmitting end; the transmitting end calculates a reward based on the received feedback information, and based on this, updates the parameters of the OLLA‑DQN algorithm to evolve a mapping model of the measurement amount and the optimal coding modulation level. The present invention realizes accurate coding modulation level selection in the case of inaccurate CSI feedback at the receiving end.

Description

Technical Field

[0001] The present invention relates to the field of wireless communications, and in particular to an adaptive coding modulation method and system based on expired channels. Background Art

[0002] Link adaptation technology refers to a type of technology that adaptively adjusts transmission parameters according to wireless channel conditions. It has strong advantages in improving data transmission rate and spectrum utilization, and is one of the key technologies of current and future mobile communication systems. Link adaptation technology mainly includes two aspects: on the one hand, the acquisition of channel information, accurate and effective acquisition of current channel environment parameters, and determination of channel quality indicators that can reflect channel conditions; on the other hand, the adjustment of transmission parameters, including modulation mode, coding mode, redundant information, transmission power, and time-frequency resources. As an important link adaptation technology, adaptive modulation and coding technology (AMC) has the core idea of ​​dynamically feeding back channel quality indicators according to time-varying channel conditions to change the modulation and coding level (MCS) of the transmitter. When the channel conditions are good, a high-order coding modulation level is used to increase the data transmission rate. When the channel conditions are poor, a low-order coding modulation level is used to ensure the reliability of data transmission, thereby improving the utilization of system resources and obtaining a higher system throughput. The commonly used AMC algorithm is the OLLA outer loop control algorithm, the core idea of ​​which is to set the block error rate (BLER) target value, adjust the reported CQI based on the demodulation result (ACK or NACK) fed back by the user, so that the final BLER converges to the corresponding target value, and then obtain the MCS corresponding to the target value. At present, the OLLA-based AMC method has been widely used in eMBB scenarios, and the effect is good.

[0003] Services in URLLC scenarios require low latency and high reliability, and their end-to-end latency is required to be controlled within 8ms, and the BLER is required to be controlled within 1e-3. Under this BLER constraint, the NACK ratio is low, and the scarcity of NACKs makes it impossible for OLLA to adjust in time, and it is difficult to track the dynamic changes of channels and interference. In addition, the current URLLC indoor factory scenario adopts a semi-static scheduling method, that is, AMC uses a fixed MCS and fixed RB resources. This method cannot match channel changes in real time and has low spectrum utilization. Therefore, since the OLLA solution and semi-static scheduling method have poor performance in URLLC scenarios and cannot meet user needs, it is necessary to design an AMC method for URLLC services to select the MCS that matches the channel to avoid resource waste. Summary of the invention

[0004] Based on this, the present invention proposes an adaptive coding and modulation technology based on OLLA-assisted deep reinforcement learning for high-reliability and low-latency networks. First, the transmitter takes the reported amount of the receiver as input, and uses the DQN algorithm to construct a mapping relationship between the measurement amount and the optimal coding and modulation level, so as to achieve real-time selection of the coding and modulation level of the matching channel without the need for complete CSI, thereby avoiding waste of resources. Then, in order to solve the problem of reduced transmission rate caused by the poor stability of the DQN algorithm, the OLLA scheme corrects the coding and modulation level selected by the DQN algorithm, thereby effectively improving the transmission rate.

[0005] To achieve the above object, the technical solution of the method of the present invention is:

[0006] An adaptive coding modulation method based on an expired channel comprises the following steps:

[0007] Step 1: The transmitter periodically transmits a pilot signal to the receiver. The receiver estimates the channel state based on the received pilot signal and uses the estimated channel state index value as CQI Feedback to the transmitter;

[0008] Step 2: The transmitter uses the coding modulation method and bit error rate selected in the previous transmission as well as the previous Channel state index value fed back by the user at the moment CQI , select the optimal coding modulation level through the OLLA-DQN algorithm, and then use the selected optimal coding modulation level to transmit the signal to the receiving end;

[0009] Step 3: The transmitter uses the selected optimal coding modulation level to transmit the signal to the receiver. The receiver determines the signal demodulation result based on the received signal and feeds it back to the transmitter. The transmitter updates the OLLA correction coefficient based on the received feedback and calculates the reward. , and then based on this reward, the DQN algorithm is updated to optimize the selection strategy of the coding modulation level to maximize the transmission rate of the entire process.

[0010] Furthermore, the step 1 comprises:

[0011] S11. Determine whether the receiving end receives the pilot signal;

[0012] S12. If the pilot signal is received, the receiving end uses a channel estimation algorithm to estimate the channel vector Make estimates;

[0013] S13. If the pilot signal is not received, the receiving end calculates the channel vector of the current time slot based on the channel vector of the previous time slot:

[0014] S14. Calculate SNR based on the acquired channel vector, , where represents the noise power, Indicates the transmit power of the transmitter;

[0015] S15. Based on the CQI table and combined SNR arrive CQI The mapping rule will be SNR Mapping CQI ;

[0016] S16. The receiving end feeds back the CQI to the transmitting end.

[0017] Further, SNR arrive CQI The mapping rules are as follows:

[0018]

[0019] in, express SNR The quantitative number of N represents the quantization accuracy, Indicates rounding operation; and express SNR The upper and lower bounds of .

[0020] Furthermore, if the channel state estimation is performed in the current time slot, the channel vector of the current time slot is obtained based on the channel state estimation scheme; if the channel state estimation is not performed, the channel vector of the current time slot is calculated based on the channel vector of the previous time slot, that is, the expired channel vector:

[0021]

[0022] In the formula and Respectively represent time slots t and time slot t Channel gain at -1, represents the coherence factor, represents a complex Gaussian random variable.

[0023] Furthermore, the problem of selecting the optimal coding modulation level in step 2 is modeled as:

[0024]

[0025] In the formula represents the optimal coding modulation level, and Indicates the use of m The coding rate and block error rate of the communication system at the coding modulation level are Represents the upper bound of the block error rate.

[0026] Furthermore, the OLLA-DQN algorithm selects the optimal coding modulation level process as follows:

[0027] 1) First, select the coding modulation level of the current time slot according to the current state and strategy of the DQN model ;

[0028] 2) Modify the selected coding modulation level based on the OLLA outer loop control scheme: , and transmit the signal using the modified coding modulation level, is the modified coding modulation level, is the OLLA correction coefficient.

[0029] Furthermore, in step 3, after receiving the transmitted signal, the receiving end feeds back the signal demodulation result and the block error rate to the transmitting end, wherein the block error rate during the transmission process is expressed as:

[0030]

[0031] in N represents the number of information bits, represents the signal-to-noise ratio, represents the block error rate, Indicates the code length occupied by the data packet, represents the inverse function of the Gaussian Q-function, ;

[0032] If the transmitting end transmits a pilot signal in the current time slot for real-time channel estimation, the receiving end uses the real-time channel vector to calculate the block error rate. Otherwise, the outdated channel vector is used to calculate the block error rate. The block error rate based on the outdated channel gain is expressed as:

[0033]

[0034] In the formula is the average signal-to-noise ratio of the channel, is the outdated signal-to-noise ratio of the known channel, is the actual signal-to-noise ratio of the current channel, is the Bessel function;

[0035] After the receiving end calculates the block error rate, it determines whether the block error rate is greater than the preset threshold. If it is greater than the threshold, NACK is returned; if it is less than the threshold, ACK is returned, and finally the signal demodulation result ACK / NACK and block error rate are fed back to the transmitter.

[0036] Furthermore, in step 3, the transmitter updates the OLLA correction coefficient based on the received feedback and calculates the reward include:

[0037] 1) The transmitter adjusts the OLLA correction coefficient based on the signal from the receiver Update, the update method is ,in , step represents the step size of algorithm update, Indicates the demodulation result, , IBlerTarget is the preset hyperparameter;

[0038] 2) Calculate rewards based on demodulation result feedback , the reward function Set to:

[0039] ,

[0040] in represents the magnification factor, represents the threshold coefficient, is the coding rate of the communication system when using the selected coding modulation level, is the block error rate;

[0041] 3) Based on the feedback demodulation result, the sample is stored in the experience pool. If the feedback demodulation result is ACK, the sample is stored in the experience pool If the demodulation result of the feedback is NACK, the sample is stored in the experience pool. middle.

[0042] Furthermore, in step 3, the model extracts samples, of which samples will be selected in reverse chronological order, and the remaining samples will be randomly selected from the experience pool;

[0043] Based on the samples selected in the experience pool, the gradient descent algorithm is used to optimize and update the DQN model parameters to optimize the selection of the coding modulation method to obtain a greater reward, where the loss function is defined as:

[0044]

[0045] In the formula Indicates that the transmitter is The rewards for the time slot, represents the discount factor, and Represents the factors that affect the parameters of the neural network model, Indicates the value of the action taken by the transmitter in the current channel state value, Indicates that the user takes action in the current channel state in the target network of value.

[0046] On the other hand, the present invention provides an adaptive coding modulation system, comprising

[0047] Module 1: It is used for the transmitter to periodically transmit a pilot signal to the receiver. The receiver estimates the channel state based on the received pilot signal and uses the estimated channel state index value as CQI Feedback to the transmitter;

[0048] Module 2: It is used by the transmitter to select the coding modulation mode and bit error rate selected in the previous transmission and the previous Channel state index value fed back by the user at the moment CQI , select the optimal coding modulation level through the OLLA-DQN algorithm, and then use the selected optimal coding modulation level to transmit the signal to the receiving end;

[0049] Module 3: The transmitter uses the selected optimal coding modulation level to transmit the signal to the receiver. The receiver determines the signal demodulation result based on the received signal and feeds it back to the transmitter. The transmitter updates the parameters of the OLLA algorithm based on the received feedback and calculates the reward. , and then based on this reward, the DQN algorithm is updated to optimize the selection strategy of the coding modulation level to maximize the transmission rate of the entire process.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] 1. To address the problem of high reliability in URLLC scenarios, in the context of the current OLLA scheme being almost ineffective and transmission resources being semi-statically fixed, the present invention adaptively adjusts the coding modulation level based on dynamic user feedback information to maximize resource utilization.

[0052] 2. The adaptive coding and modulation level selection problem at the transmitter in the case of expired CSI and inaccurate CSI feedback is considered. Compared with the coding and modulation level selection based on complete CSI, it is more in line with the actual URLLC scenario.

[0053] 3. Considering the sample unevenness problem of ACK feedback and NACK feedback in the real URLLC scenario, the DQN algorithm is combined with the OLLA scheme to improve the system transmission rate while ensuring communication stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 This is a flow chart of an adaptive coding modulation method based on OLLA-DQN according to an embodiment of the present application;

[0055] Figure 2 A flowchart of the CQI operation of the receiving end in accordance with an embodiment of the present application;

[0056] Figure 3 A flowchart of the coding modulation level selection of the OLLA-DQN algorithm in an embodiment of the present application;

[0057] Figure 4 This is a parameter update flow chart of the OLLA-DQN algorithm according to an embodiment of the present application. DETAILED DESCRIPTION

[0058] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians of the present invention without making creative work are within the scope of protection of the present invention.

[0059] Example 1

[0060] Reference Figure 1 The embodiment of the present application provides an adaptive coding modulation method based on OLLA-DQN, comprising the following specific steps:

[0061] S1. The receiving end estimates the channel vector and calculates the SNR, then maps the SNR to the CQI based on the CQI table, and finally feeds the CQI back to the transmitting end.

[0062] S2. Use the OLLA-DQN algorithm to select the coding modulation level of the current time slot, and use the selected coding modulation level to transmit the signal. After receiving the signal, the receiving end demodulates and calculates the BLER, and then feeds back the signal demodulation result (ACK / NACK) to the transmitting end.

[0063] S3. The transmitter updates the parameters of the OLLA algorithm based on the received feedback, then updates the state of the DQN algorithm and calculates the reward , and finally the obtained reward is used to update the DQN correction coefficient to maximize the transmission rate of the whole process.

[0064] like Figure 2 As shown, the specific steps for the receiving end to estimate the channel vector and obtain the CQI in step S1 are:

[0065] S11. Determine whether the receiving end has received the pilot signal.

[0066] S12. If the pilot signal is received, the receiving end uses a channel estimation algorithm to estimate the channel vector Make an estimate.

[0067] S13. If the pilot signal is not received, the receiving end calculates the (inaccurate) channel vector of the current time slot based on the channel vector of the previous time slot (i.e., the outdated channel vector): , where and represent the channel gains at time slot t and time slot t-1 respectively, represents the coherence factor, represents a complex Gaussian random variable.

[0068] S14. Calculate SNR based on the acquired channel vector, , where represents the noise power, Indicates the transmit power of the transmitter.

[0069] S15. Based on the CQI table and in combination with the following rules, the SNR is mapped to the CQI. The mapping rule from SNR to CQI is as follows:

[0070]

[0071] In the formula Indicates the quantization number of SNR, N indicates the quantization accuracy, Indicates a rounding operation. and Indicates the upper and lower bounds of the SNR.

[0072] S16. The receiving end feeds back the CQI to the transmitting end.

[0073] like Figure 3 As shown, the specific steps of selecting the coding modulation level of the transmitting end and the information feedback of the receiving end in step 2 are:

[0074] S21. The transmitting end feeds back the CQI received from the receiving end, and sends the MCS selected for the last transmission, demodulation result, OLLA correction coefficient, current time slot and previous time slot. The CQI of each time slot is used as status information.

[0075] S22. Input the state information into DQN to select the coding modulation level.

[0076] S23. Input the coding modulation selected by DQN into the OLLA algorithm to modify the selected coding modulation level .

[0077] S24. After receiving the transmitted signal, the receiving end determines whether real-time channel estimation has been performed before the current communication time slot.

[0078] Right now .

[0079] S25. If real-time channel estimation is performed ( ), the real-time SNR is used to calculate the BLER, where the BLER during transmission can be expressed as:

[0080]

[0081] Where N represents the number of information bits, represents SNR, represents BLER, Indicates the code length occupied by the data packet, represents the inverse function of the Gaussian Q-function, .

[0082] S26. If real-time channel estimation is not performed ( ), the outdated channel vector is used to calculate the BLER. The BLER based on the outdated channel gain can be expressed as:

[0083]

[0084] In the formula is the average signal-to-noise ratio of the channel, is the outdated signal-to-noise ratio of the known channel, is the actual signal-to-noise ratio of the current channel, is the Bessel function.

[0085] S27. Determine whether the BLER is greater than a preset threshold, i.e., whether :

[0086] S28. If , then returns NACK.

[0087] S29. If , returns ACK.

[0088] S30. The receiving end feeds back the signal demodulation result (ACK / NACK) and BLER to the transmitting end.

[0089] like Figure 4 As shown, the specific steps for updating the parameters of the OLLA-DQN algorithm at the transmitter in step 3 are:

[0090] S31. The transmitter adjusts the OLLA correction coefficient based on the demodulation result of the signal at the receiver. The OLLA correction coefficient is updated as follows: ,in , Step represents the step length of algorithm update, Indicates the demodulation result, .

[0091] S32. The transmitter calculates the reward based on the feedback from the receiver , the reward function Defined as ,in and Represents the amplification factor, which is used to ensure that the reward obtained when the transmission is successful is greater than the reward when it fails.

[0092] S33. Store the sample data into the experience pool based on the feedback demodulation result.

[0093] S34. If the feedback demodulation result is ACK, the sample is stored in experience pool 1.

[0094] S35. If the feedback demodulation result is NACK, the obtained samples are stored in experience pool 2.

[0095] S36. Determine whether the current time slot is a period of N, that is, .

[0096] S37. If , then draw from experience pool 1 When training with samples, samples will be selected in reverse chronological order, and the remaining Samples will be randomly selected from the experience pool.

[0097] S38. If , drawn from experience pool 2 samples.

[0098] S39. Based on the samples selected from the experience pool, use the gradient descent algorithm to optimize the model parameters of DQN to obtain greater rewards. The loss function is defined as:

[0099]

[0100] In the formula Indicates that the transmitter is The rewards for the time slot, represents the discount factor, and represents the factors that affect the parameters of the neural network model, Indicates the value of the action taken by the transmitter in the current channel state value, Indicates that the user takes action in the current channel state in the target network of value.

[0101] Example 2

[0102] This embodiment provides an adaptive coding modulation system based on OLLA-DQN, including:

[0103] Module 1: It is used for the transmitter to periodically transmit a pilot signal to the receiver. The receiver estimates the channel state based on the received pilot signal and uses the estimated channel state index value as CQI Feedback to the transmitter;

[0104] Module 2: It is used by the transmitter to select the coding modulation mode and bit error rate selected in the previous transmission and the previous Channel state index value fed back by the user at the moment CQI , select the optimal coding modulation level through the OLLA-DQN algorithm, and then use the selected optimal coding modulation level to transmit the signal to the receiving end;

[0105] Module 3: The transmitter uses the selected optimal coding modulation level to transmit the signal to the receiver. The receiver determines the signal demodulation result based on the received signal and feeds it back to the transmitter. The transmitter updates the parameters of the OLLA algorithm based on the received feedback and calculates the reward. , and then based on this reward, the DQN algorithm is updated to optimize the selection strategy of the coding modulation level to maximize the transmission rate of the entire process.

[0106] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present invention.

[0107] Obviously, those skilled in the art can make various changes and modifications to the embodiments of the present invention without departing from the spirit and scope of the embodiments of the present invention. Thus, if these modifications and variations of the embodiments of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

[0108] Other parts not described in detail are all prior art.

Claims

1. An adaptive coding modulation method based on expired channels, characterized in that: The following steps are involved: Step 1: The transmitter periodically transmits a pilot signal to the receiver. The receiver estimates the channel state based on the received pilot signal and uses the estimated channel state index value as CQI Feedback to the transmitter; Step 2: The transmitter selects the coding modulation mode and bit error rate selected in the previous transmission and the previous Channel state index value fed back by the user at the moment CQI , select the optimal coding modulation level through the OLLA-DQN algorithm, and then use the selected optimal coding modulation level to transmit the signal to the receiving end; Step 3: The transmitter uses the selected optimal coding modulation level to transmit the signal to the receiver. The receiver determines the signal demodulation result based on the received signal and feeds it back to the transmitter. The transmitter updates the OLLA correction coefficient based on the received feedback and calculates the reward. , and then based on this reward, the DQN algorithm is updated to optimize the selection strategy of the coding modulation level to maximize the transmission rate of the entire process.

2. The adaptive coding modulation method based on expired channel according to claim 1, characterized in that: The step 1 comprises: S11. Determine whether the receiving end receives the pilot signal; S12. If the pilot signal is received, the receiving end uses a channel estimation algorithm to estimate the channel vector Make estimates; S13. If the pilot signal is not received, the receiving end calculates the channel vector of the current time slot based on the channel vector of the previous time slot: S14. Calculate SNR based on the acquired channel vector, , where represents the noise power, Indicates the transmit power of the transmitter; S15. Based on the CQI table and combined SNR arrive CQI The mapping rule will be SNR Mapping CQI ; S16. The receiving end feeds back the CQI to the transmitting end.

3. The adaptive coding modulation method based on expired channel according to claim 2, characterized in that: SNR arrive CQI The mapping rules are as follows: in, express SNR The quantitative number of N represents the quantization accuracy, Indicates rounding operation; and express SNR The upper and lower bounds of .

4. The adaptive coding modulation method based on expired channel according to claim 2, characterized in that: If the channel state estimation is performed in the current time slot, the channel vector of the current time slot is obtained based on the channel state estimation scheme; if the channel state estimation is not performed, the channel vector of the current time slot is calculated based on the channel vector of the previous time slot, that is, the expired channel vector: In the formula and Respectively represent time slots t and time slot t Channel gain at -1, represents the coherence factor, represents a complex Gaussian random variable.

5. The adaptive coding modulation method based on expired channel according to claim 1, characterized in that: The problem of selecting the optimal coding modulation level in step 2 is modeled as follows: In the formula represents the optimal coding modulation level, and Indicates the use of m The coding rate and block error rate of the communication system at the coding modulation level are Represents the upper bound of the block error rate.

6. The adaptive coding modulation method based on expired channel according to claim 1, characterized in that: In step 2, the OLLA-DQN algorithm selects the optimal coding modulation level as follows: 1) First, select the coding modulation level of the current time slot according to the current state and strategy of the DQN model ; 2) Modify the selected coding modulation level based on the OLLA outer loop control scheme: , and transmit the signal using the modified coding modulation level, is the modified coding modulation level, is the OLLA correction coefficient.

7. The adaptive coding modulation method based on expired channel according to claim 1, characterized in that: In step 3, after receiving the transmitted signal, the receiving end feeds back the signal demodulation result and block error rate to the transmitting end, where the block error rate during the transmission process is expressed as: in N represents the number of information bits, represents the signal-to-noise ratio, represents the block error rate, Indicates the code length occupied by the data packet, represents the inverse function of the Gaussian Q-function, ; If the transmitting end transmits a pilot signal in the current time slot for real-time channel estimation, the receiving end uses the real-time channel vector to calculate the block error rate. Otherwise, the outdated channel vector is used to calculate the block error rate. The block error rate based on the outdated channel gain is expressed as: In the formula is the average signal-to-noise ratio of the channel, is the outdated signal-to-noise ratio of the known channel, is the actual signal-to-noise ratio of the current channel, is the Bessel function; After the receiving end calculates the block error rate, it determines whether the block error rate is greater than the preset threshold. , if it is greater than the threshold, NACK is returned; If it is less than the threshold, ACK is returned, and finally the signal demodulation result ACK / NACK and block error rate are fed back to the transmitter.

8. The adaptive coding modulation method based on expired channel according to claim 1, characterized in that: In step 3, the transmitter updates the OLLA correction coefficient based on the received feedback and calculates the reward include: 1) The transmitter adjusts the OLLA correction coefficient based on the signal from the receiver Update, the update method is ,in , step represents the step size of algorithm update, Indicates the demodulation result, , IBlerTarget is the preset hyperparameter; 2) Calculate rewards based on demodulation result feedback , the reward function Set to: ， in represents the magnification factor, represents the threshold coefficient, is the coding rate of the communication system when using the selected coding modulation level, is the block error rate; 3) Based on the feedback demodulation result, the sample is stored in the experience pool. If the feedback demodulation result is ACK, the sample is stored in the experience pool If the demodulation result of the feedback is NACK, the sample is stored in the experience pool. middle.

9. The adaptive coding modulation method based on expired channel according to claim 8, characterized in that: In step 3, the model extracts samples, of which samples will be selected in reverse chronological order, and the remaining samples will be randomly selected from the experience pool; Based on the samples selected in the experience pool, the gradient descent algorithm is used to optimize and update the DQN model parameters to optimize the selection of the coding modulation method to obtain a greater reward, where the loss function is defined as: In the formula Indicates that the transmitter is The rewards for the time slot, represents the discount factor, and represents the factors that affect the parameters of the neural network model, Indicates the value of the action taken by the transmitter in the current channel state value, Indicates that the user takes action in the current channel state in the target network of value.

10. An adaptive coding modulation system based on an expired channel, characterized in that: include Module 1: It is used for the transmitter to periodically transmit a pilot signal to the receiver. The receiver estimates the channel state based on the received pilot signal and uses the estimated channel state index value as CQI Feedback to the transmitter; Module 2: It is used by the transmitter to calculate the coding modulation method and bit error rate selected in the previous transmission and the previous Channel state index value fed back by the user at the moment CQI , select the optimal coding modulation level through the OLLA-DQN algorithm, and then use the selected optimal coding modulation level to transmit the signal to the receiving end; Module 3: The transmitter uses the selected optimal coding modulation level to transmit the signal to the receiver. The receiver determines the signal demodulation result based on the received signal and feeds it back to the transmitter. The transmitter updates the parameters of the OLLA algorithm based on the received feedback and calculates the reward. , and then based on this reward, the DQN algorithm is updated to optimize the selection strategy of the coding modulation level to maximize the transmission rate of the entire process; The adaptive coding modulation system based on expired channels is used to execute the steps in the adaptive coding modulation method based on expired channels described in any one of claims 1-9.

Images