Method and apparatus for reducing transmission delay of small data packet service in mobile communication network

Abstract

The application provides a method and device for reducing small data packet service transmission delay in a mobile communication network, and relates to the technical field of mobile communication. The method comprises the following steps: obtaining the delay requirement and reliability requirement of a user, and constructing a wireless communication model of ultra-reliable low-latency service according to the delay requirement and reliability requirement; modeling an alpha-kappa-mu channel based on a random network calculus method according to the wireless communication model, and obtaining a delay violation probability formula under the alpha-kappa-mu channel; establishing the relationship between delay and reliability according to the delay violation probability formula; determining the service delay component, frame length, service transmission mode, semantic compression ratio and bandwidth allocation result by using an EA3C algorithm according to the delay requirement of the user, making a decision on the delay reduction problem through the relationship between delay and reliability, and outputting the maximum value of delay quality to reduce the small data packet service transmission delay. The application can reduce the small data packet service transmission delay in a mobile communication network.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of mobile communication technology, in particular to a method and device for reducing the transmission delay of small data packet services in a mobile communication network. BACKGROUND

[0002] With the rapid development of mobile communication technology, especially in 5G and future 6G networks, the requirements for communication performance are becoming higher and higher, especially for small data packet services, which have begun to impose extremely strict requirements on latency, requiring latency to be less than 1 millisecond. In order to meet this demand, the communication system must minimize the latency while ensuring transmission quality. Existing technical solutions mainly reduce latency by optimizing transmission protocols, increasing bandwidth, and adjusting transmission methods. However, the above methods are usually difficult to balance the flexibility and efficiency of the system, especially in a dynamically changing network environment, how to adjust the system settings in real time according to different business needs is still a technical challenge. Semantic communication technology, as a new communication mode, can significantly reduce the amount of data transmitted by extracting the core semantic information in the business data, thereby effectively reducing the transmission latency. Semantic communication technology has been applied to a certain extent in data compression and efficient transmission, but due to the need for a certain time in the semantic extraction process, there are still challenges for applications with extremely high latency requirements, especially for small data packet transmission. Therefore, how to flexibly adjust the semantic extraction ratio under different business needs and network states, and reduce the overall latency, has become the focus of current technical research.

[0003] In existing technologies, semantic communication usually reduces the amount of data by extracting all or part of the information to reduce transmission latency. However, the latency of the semantic extraction process and the limitation of transmission bandwidth are still the bottleneck of its application. For example, existing semantic compression techniques do not fully consider dynamic adjustment of latency while ensuring data effectiveness. On the other hand, although some researches reduce latency through bandwidth management and frame length optimization, they lack flexibility and adaptability under different system states, making it difficult to be widely applied in actual scenarios.

[0004] Currently, the requirement for latency in the field of mobile communication, especially in 5G and future 6G networks, is becoming increasingly stringent, especially for small data packet transmission requiring a latency of less than 1 millisecond. To meet this requirement, the existing technology mainly adopts two methods: data compression and bandwidth optimization. Among them, the bandwidth optimization technology reduces the latency by dynamically adjusting the network bandwidth and frame length. In the existing technology, the transmission is usually optimized by fixed bandwidth and frame length, which lacks flexibility and cannot be adjusted in real time according to the network state and user demand, resulting in the inability to maximize latency optimization in some scenarios. Although semantic communication and bandwidth optimization can improve latency and efficiency, they fail to adjust flexibly in practical applications and cannot automatically select the best transmission strategy under various system states, resulting in insufficient performance optimization. In addition, the lack of a unified dynamic adjustment mechanism limits the optimization of latency and reliability. SUMMARY

[0005] The existing technology usually optimizes transmission by fixed bandwidth and frame length, which lacks flexibility and cannot be adjusted in real time according to the network state and user demand, resulting in the technical problem of being unable to maximize latency optimization. The embodiments of the present application provide a method and device for reducing small data packet service transmission latency in a mobile communication network. The technical solution is as follows:

[0006] On the one hand, a method for reducing small data packet service transmission latency in a mobile communication network is provided, which is realized by a device for reducing small data packet service transmission latency in a mobile communication network. The method comprises:

[0007] S1, obtaining the latency requirement and reliability requirement of a user, and constructing a wireless communication model of ultra-reliable low-latency service according to the latency requirement and reliability requirement;

[0008] S2, modeling an alpha-kappa-mu channel based on a stochastic network calculus method according to the wireless communication model of ultra-reliable low-latency service, obtaining a latency violation probability formula under the alpha-kappa-mu channel, and establishing a relationship between latency and reliability according to the latency violation probability formula;

[0009] S3, determining a service latency component, a frame length, a service transmission mode, a semantic compression ratio, and a bandwidth allocation result by using an EA3C algorithm according to the latency requirement of the user, making a decision on the latency reduction problem by the relationship between latency and reliability according to the service latency component, the service transmission mode, the semantic compression ratio, the frame length, and the bandwidth allocation result, and outputting a maximum latency quality to reduce small data packet service transmission latency.

[0010] Optionally, the S2 based on the stochastic network calculus method models the alpha-kappa-mu channel to obtain the latency violation probability formula under the alpha-kappa-mu channel, which comprises:

[0011] S21, defining a queue cumulative arrival process of services to the base station; determining an operating state of the wireless communication system according to the queue cumulative arrival process of services to the base station;

[0012] S22, determining a dynamic change of the queue length according to the operating state of the system, and determining a rule of the queue length changing with time;

[0013] S23, defining a case of delay violation when the queue length exceeds a set maximum value by taking a delay violation probability as a reliability index, and obtaining an upper bound expression of the delay violation probability;

[0014] S24, transforming the upper bound expression of the delay violation probability into an expected form by using Markov inequality and Holder inequality, and obtaining an expression related to the delay violation probability and the expectation;

[0015] S25, analyzing service arrival behaviors by using a Poisson process according to the upper bound expression of the delay violation probability, calculating characteristics of the cumulative arrival process by assuming that a data arrival rate and a packet size conform to a Poisson distribution, and describing behaviors of a wireless channel service process according to a small data packet transmission model;

[0016] S26, processing the expression related to the delay violation probability and the expectation by using a Meijer G function according to steps S21-S25, and obtaining a final expression of the final delay violation probability.

[0017] Optionally, the service delay components include a queuing delay, a transmission delay, a frame alignment delay, a semantic coding delay, and a base station and terminal processing delay.

[0018] The queuing delay is a time for a data packet to wait in a buffer queue, and the queuing delay increases with an increase of a load.

[0019] The transmission delay depends on a data packet size and a transmission rate.

[0020] The frame alignment delay is a delay generated when a data packet arrives and a resource is ready for scheduling, and the system has to wait for a start of a next frame to transmit the data packet.

[0021] The semantic coding delay is related to a number of cycles required for calculation, a processing density, and a device calculation capability.

[0022] The base station and terminal processing delay is related to a processing capability of the device itself.

[0023] Optionally, selection factors of the frame length include a delay requirement of a user, a channel quality, and a data packet size.

[0024] The frame duration includes three different frame durations of 0.25 milliseconds, 0.125 milliseconds and 0.0625 milliseconds.

[0025] Optionally, the semantic compression ratio determines the size of the data packet.

[0026] The semantic compression ratio has a value range of 0 to 1, and a larger value indicates a higher compression ratio and a smaller data packet.

[0027] Optionally, the transmission mode of the service is adaptive bit transmission and semantic transmission.

[0028] Optionally, the S3 uses the EA3C algorithm to determine the service delay component, the frame duration, the transmission mode of the service, the semantic compression ratio and the bandwidth allocation result; according to the service delay component, the transmission mode of the service, the semantic compression ratio, the frame duration and the bandwidth allocation result, the relationship between the delay and the reliability is determined to solve the delay reduction problem, including:

[0029] S31, initialize the EA3C algorithm parameters, including: Actor learning rate, Critic learning rate, discount factor and decay factor; initialize the global network parameters and target network parameters; initialize the global shared optimizer; start N parallel environment threads, wherein each thread initializes its local network parameters;

[0030] S32, synchronize parameters, copy global network parameters to local network, initialize local environment, reset local environment and obtain initial state;

[0031] S33, execute actions and collect experiences, and select actions according to the current policy network;

[0032] S34, execute actions, obtain rewards and next states; store the executed actions, initial state rewards and next states into the experience sequence in the thread local;

[0033] S35, update the local network and accumulate the gradient; when the length of the local experience sequence reaches a certain threshold T or encounters a termination state, calculate the reward value from t to t+n, and calculate the update target from back to front: calculate the value function error and the policy gradient, and calculate the total loss;

[0034] S36, update the global network parameters, update the global parameters using the gradient, synchronize the local network parameters to the global parameters; continue to the next state, if the termination state is not reached, continue iteration; when all threads complete training or reach the maximum iteration number, end the algorithm, complete the pre-training of the network; output the pre-trained Actor network and Critic network;

[0035] S37, defining an action space and a state space according to the service delay component, the frame length, the transmission mode of the service, the semantic compression ratio, and the bandwidth allocation result;

[0036] S38, setting a completion degree of the delay as a reward function; according to the action space, the state space, and the reward function, fine-tuning the pre-trained Actor network and Critic network through the relationship between the delay and the reliability, updating the network parameters, outputting a decision on the delay reduction problem, executing the decision, and completing the resource scheduling.

[0037] On the other hand, a device for reducing small packet service transmission delay in a mobile communication network is provided, which is applied to the method for reducing small packet service transmission delay in a mobile communication network, and the device comprises:

[0038] A first acquisition unit is configured to acquire a delay requirement and a reliability requirement of a user, and construct a wireless communication model of an ultra-reliable and low-latency service according to the delay requirement and the reliability requirement;

[0039] A second acquisition unit is configured to model an alpha-kappa-mu channel based on a stochastic network calculus method according to the wireless communication model of the ultra-reliable and low-latency service, obtain a delay violation probability formula under the alpha-kappa-mu channel, and establish a relationship between the delay and the reliability according to the delay violation probability formula.

[0040] An optimization and output unit is configured to determine a service delay component, a frame length, a transmission mode of the service, a semantic compression ratio, and a bandwidth allocation result according to the delay requirement of the user by using an EA3C algorithm, make a decision on a delay reduction problem through the relationship between the delay and the reliability according to the service delay component, the transmission mode of the service, the semantic compression ratio, the frame length, and the bandwidth allocation result, and output a maximum value of delay quality to reduce small packet service transmission delay.

[0041] Optionally, the second acquisition unit is configured to:

[0042] (1) define a queue accumulation arrival process of services to a base station, and determine an operating state of a wireless communication system according to the queue accumulation arrival process of the services to the base station;

[0043] (2) determine a dynamic change of a queue length according to the operating state of the system, and clarify a rule of the queue length changing with time;

[0044] (3) take a delay violation probability as a reliability index, define a case of delay violation when the queue length exceeds a set maximum value, and obtain an upper bound expression of the delay violation probability;

[0045] (4) Using Markov inequality and Holder inequality, the upper bound expression of the delay violation probability is transformed into the form of expectation, and an expression related to the delay violation probability and expectation is obtained;

[0046] (5) According to the upper bound expression of the delay violation probability, the arrival behavior is analyzed by using Poisson process, and the characteristics of the cumulative arrival process are calculated by assuming that the data arrival rate and the packet size conform to Poisson distribution; according to a small data packet transmission model, the behavior of the wireless channel service process is described;

[0047] (6) According to steps (1)-(5), the expression related to the delay violation probability and expectation is processed by using Meijer G function, and a final expression of the delay violation probability is obtained.

[0048] Optionally, the service delay component includes: queuing delay, transmission delay, frame alignment delay, semantic coding delay, and base station and terminal processing delay.

[0049] The queuing delay is the time for a data packet to wait in a buffer queue, and the queuing delay increases with the increase of load.

[0050] The transmission delay depends on the data packet size and the transmission rate.

[0051] The frame alignment delay is the delay generated when the system must wait for the start of the next frame to transmit the data packet when the data packet arrives and the resource is ready for scheduling.

[0052] The semantic coding delay is related to the number of cycles required for calculation, processing density, and device calculation capability.

[0053] The base station and terminal processing delay is related to the processing capability of the device itself.

[0054] Optionally, the selection factors of the frame length include: the delay requirement of the user, the channel quality, and the data packet size.

[0055] The frame length includes three different frame lengths of 0.25 milliseconds, 0.125 milliseconds, and 0.0625 milliseconds.

[0056] Optionally, the semantic compression ratio determines the size of the data packet.

[0057] The semantic compression ratio has a value range of 0 to 1, and the larger the value is, the higher the compression ratio is, and the smaller the data packet is.

[0058] Optionally, the transmission mode of the service is adaptive bit transmission and semantic transmission.

[0059] Optionally, the optimization and output unit is configured to:

[0060] Initializing EA3C algorithm parameters includes: Actor learning rate, Critic learning rate, discount factor and decay factor; initializing global network parameters and target network parameters; initializing a global shared optimizer; starting N parallel environment threads, wherein each thread initializes its local network parameters;

[0061] Each thread synchronizes parameters, copies global network parameters to local networks, initializes a local environment, resets the local environment, and obtains an initial state;

[0062] Performing actions and collecting experiences, selecting actions according to the current policy network;

[0063] Performing actions, obtaining rewards and next states; storing the executed actions, initial state rewards and next states into the experience sequence local to the thread;

[0064] Updating the local network and accumulating gradients; when the length of the local experience sequence reaches a certain threshold T or encounters a termination state, calculating the reward value from t to t+n, and sequentially calculating the update target from back to front: calculating the value function error and policy gradient, and calculating the total loss;

[0065] Updating the global network parameters, updating the global parameters using the gradients, synchronizing the local network parameters to the global parameters; continuing to the next state, if the termination state is not reached, continue iteration; when all threads complete training or reach the maximum number of iterations, end the algorithm, complete pre-training of the network; output the pre-trained Actor network and Critic network;

[0066] Defining the action space and state space according to the business latency components, frame length, transmission mode of the business, semantic compression ratio and bandwidth allocation result;

[0067] Setting the degree of completion of the latency as the reward function; according to the action space, state space and reward function, fine-tuning the pre-trained Actor network and Critic network through the relationship between latency and reliability, updating the network parameters, outputting the decision on the latency reduction problem, executing the decision, and completing resource scheduling.

[0068] On the other hand, a mobile communication network small packet service transmission latency reduction device is provided, which comprises: a processor; a memory, the memory has computer readable instructions stored thereon, when the computer readable instructions are executed by the processor, any one of the above mobile communication network small packet service transmission latency reduction methods is implemented.

[0069] In another aspect, a computer readable storage medium is provided, the storage medium having stored therein at least one instruction, the at least one instruction being loaded and executed by a processor to implement any one of the methods for reducing transmission delay of small data packet service in a mobile communication network.

[0070] The technical scheme provided by the embodiment of the present application has at least the following beneficial effects:

[0071] The embodiment of the present application firstly acquires the delay requirement and reliability requirement of a user, and constructs a wireless communication model of ultra-reliable and low-delay service according to the delay requirement and reliability requirement; secondly, the alpha-kappa-mu channel is modeled based on a random network calculus method according to the wireless communication model of ultra-reliable and low-delay service, and a delay violation probability formula under the alpha-kappa-mu channel is obtained; the relationship between delay and reliability is established according to the delay violation probability formula; finally, the EA3C algorithm is adopted to determine the service delay component, frame length, transmission mode of service, semantic compression ratio and bandwidth allocation result according to the delay requirement of the user; the delay reduction problem is decided through the relationship between delay and reliability according to the service delay component, transmission mode of service, semantic compression ratio, frame length and bandwidth allocation result, and the maximum value of delay quality is output, so that the transmission delay of small data packet service is reduced.

[0072] The present application can enable each user to flexibly select to extract part of information as semantic communication transmission or direct communication transmission, to select to use different frame lengths, and to determine transmission bandwidth. The network determines the optimal selection of each user according to unified optimization scheduling, so that the overall network performance is improved; the embodiment of the present application combines semantic communication and direct communication, and reduces the delay of service by flexibly adjusting the semantic compression ratio, frame length and transmission bandwidth for small data packet service with delay requirement. BRIEF DESCRIPTION OF DRAWINGS

[0073] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0074] Figure 1 is a flow chart of a method for reducing transmission delay of small data packet service in a mobile communication network provided by the embodiment of the present application;

[0075] Figure 2 is a block diagram of a device for reducing transmission delay of small data packet service in a mobile communication network provided by the embodiment of the present application;

[0076] Figure 3The embodiment of the present application provides a structure diagram of a device for reducing small data packet service transmission delay in a mobile communication network. DETAILED DESCRIPTION

[0077] The technical solutions in the present application will be described below with reference to the drawings.

[0078] In the embodiments of the present application, the words such as "example", "for example" and the like are used to represent an example, illustration or description. Any embodiment or design scheme described as "example" in the present application should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. Rather, the word "example" is intended to present the concept in a specific manner. In addition, in the embodiments of the present application, the meaning expressed by "and / or" can be both, or can be one of the two.

[0079] In the embodiments of the present application, "image" and "picture" can be used interchangeably at times, and it should be pointed out that the meanings expressed are consistent when the distinction is not emphasized. "Of", "corresponding" and "corresponding" can be used interchangeably at times, and it should be pointed out that the meanings expressed are consistent when the distinction is not emphasized.

[0080] In the embodiments of the present application, sometimes the subscript such as W1 can be written in the form of non-subscript such as W1, and the meanings expressed are consistent when the distinction is not emphasized.

[0081] To make the technical problems, technical solutions and advantages to be solved by the present application more clear, the following will be described in detail with reference to the drawings and specific embodiments.

[0082] The embodiment of the present application provides a method for reducing small data packet service transmission delay in a mobile communication network, which can be realized by a device for reducing small data packet service transmission delay in a mobile communication network. The device for reducing small data packet service transmission delay in a mobile communication network can be a terminal or a server. As shown in the method flowchart for reducing small data packet service transmission delay in a mobile communication network, the processing flow of the method can include the following steps. Figure 1 The processing flow of the method can include the following steps.

[0083] S1, obtaining the delay requirement and the reliability requirement of a user, and constructing a wireless communication model of an ultra-reliable and low-delay service according to the delay requirement and the reliability requirement.

[0084] The wireless communication model of the ultra-reliable and low-delay service includes N user nodes and 1 base station, and the parameters include the delay requirement, the reliability requirement, the data packet size, the bandwidth and the power of the user node i.

[0085] S2, according to the wireless communication model of the ultra-reliable low-latency service, a channel is modeled based on a stochastic network calculus method to obtain a latency violation probability formula under the channel; and according to the latency violation probability formula, a relationship between latency and reliability is established.

[0086] wherein, when the channel parameters take different values, the alpha-kappa-mu channel can cover most channel cases, including common channels such as Rayleigh channels and Rician channels.

[0087] wherein, since the service has extremely strict latency and reliability requirements, the design of the communication network depends on accurate channel characterization. The alpha-kappa-mu channel is a kind of generalized wireless propagation channel, but due to the complexity of the channel, the probability density function, the cumulative distribution function and the moment generating function contain multiple complex and abstract hypergeometric functions, making it difficult to extend to future 6G networks.

[0088] wherein, the alpha-kappa-mu channel contains four parameters, wherein alpha is a shape parameter that determines the shape of the distribution, kappa is a shadow fading parameter that describes the shadow fading effect, mu is a scale parameter that determines the scale of the distribution, and Omega represents the average power of the signal. When the above parameters take certain specific values, the alpha-kappa-mu channel can be simplified as a common channel, and the specific parameters are shown in Table 1.

[0089] Table 1

[0090]

[0091] wherein, the latency violation probability formula under the alpha-kappa-mu channel provides a solid foundation for accurate analysis of latency performance under different channel conditions. By focusing on the cumulative arrival process and the cumulative service process of the queue, the arrival process of the service is analyzed using Poisson distribution, the service process of the service is analyzed using finite block length coding mode, and based on the theory of stochastic network calculus, the relationship between latency and reliability is established.

[0092] Optionally, the specific implementation process of S2 can include S21-S26:

[0093] S21, defining a queue cumulative arrival process of a service to a base station; and determining a running state of a wireless communication system according to the queue cumulative process of the service to the base station;

[0094] wherein, the queue cumulative process of the service to the base station includes a cumulative process of data arrival and service process. By defining the cumulative process, the system running state is described, which provides a basis for subsequent analysis.

[0095] wherein, the representation of the queue length involves the balance relationship of data arrival and service and data departure.

[0096] S22, according to the running state of the system, determine the dynamic change of the queue length, and determine the law of the change of the queue length with time;

[0097] S23, define the case of delay violation when the queue length exceeds the set maximum value by taking the delay violation probability as the reliability index, and obtain the upper bound expression of the delay violation probability;

[0098] S24, transform the upper bound expression of the delay violation probability into the expected form by using Markov inequality and Holder inequality, and obtain the expression related to the expectation of the delay violation probability;

[0099] S25, according to the upper bound expression of the delay violation probability, analyze the service arrival behavior by using Poisson process, calculate the characteristics of the cumulative arrival process by assuming that the data arrival rate and packet size conform to Poisson distribution, and describe the behavior of the wireless channel service process according to the small data packet transmission model;

[0100] For example, the transmission model based on channel state and data block length; the channel state is represented by an α-κ-μ channel, and the probability density function of the channel is used for analysis.

[0101] S26, according to steps S21-S25, process the expression related to the expectation of the delay violation probability by using Meijer G function, and obtain the final expression of the final delay violation probability.

[0102] Wherein, the Meijer G function is used to transform the complex integral and expression into a processable form; by simplifying the expression of the Meijer G function, the expression is further simplified to adapt to different communication channel conditions and system parameters.

[0103] Wherein, the final expression of the final delay violation probability can adapt to different channel conditions, and provide the possibility for delay analysis in complex scenarios.

[0104] Wherein, the final expression of the final delay violation probability can be expressed by the following formula (1):

[0105]

[0106] Wherein,

[0107] Wherein,

[0108] Wherein,

[0109] Wherein, ε k represents the delay violation probability, θ is a positive parameter of stochastic network calculus, N k represents the block length, denotes inverse function of Gaussian Q function, and ε c denotes coding error rate, denotes time delay requirement of user, and B k W k denotes bandwidth of user; B k denotes bandwidth of resource block; W k denotes number of resource blocks; G denotes Meijer G function; α denotes first parameter of channel; Ω denotes second parameter of channel; Γ(G) denotes gamma function of G; Γ(Z) denotes gamma function of Z; wherein, in order to make expression of time delay violation probability more clear, relevant variables are replaced by Z, U, Y and C; Z denotes first replacement relevant variable; U denotes second replacement relevant variable; Y denotes third replacement relevant variable; C denotes fourth replacement relevant variable; G G denotes power form of Meijer G function.

[0110] wherein, according to Meijer G function formula, by setting α=2, κ=0, μ=1, Ω=1, Meijer G function of Rayleigh channel is obtained, and can be expressed by following formula (2):

[0111]

[0112] wherein, G' denotes Meijer G function of Rayleigh channel;

[0113] wherein, according to Meijer G function formula, by setting α=2, κ=0, μ=2, Ω=1, Meijer G function of Nakagami-m channel is obtained, and can be expressed by following formula (3):

[0114]

[0115] wherein, G'' denotes Meijer G function of Nakagami-m channel.

[0116] S3, obtaining time delay requirement of user; according to time delay requirement of user, adopting EA3C algorithm, determining service time delay component, frame length, transmission mode of service, semantic compression ratio and bandwidth allocation result; according to service time delay component, transmission mode of service, semantic compression ratio, frame length and bandwidth allocation result, making decision on time delay reduction problem through relationship between time delay and reliability, outputting maximum value of time delay quality, so as to reduce time delay of small data packet service transmission.

[0117] Optionally, service time delay component comprises: queuing time delay, transmission time delay, frame alignment time delay, semantic coding and decoding time delay and base station and terminal processing time delay.

[0118] The queuing delay is the time for the data packet to wait in the buffer queue, and the queuing delay increases with the increase of the load.

[0119] The transmission delay depends on the data packet size and the transmission rate.

[0120] The frame alignment delay is the delay generated when the data packet arrives and the resource is ready to be scheduled, and the system has to wait for the start of the next frame to transmit the data packet.

[0121] The semantic coding and decoding delay is related to the number of cycles required for calculation, processing density, and device computing power.

[0122] The base station and terminal processing delay is related to the processing capability of the device itself.

[0123] The semantic communication mode reduces the transmission delay, but increases the semantic coding and decoding delay.

[0124] Optionally, the transmission mode of the service is adaptively selected from bit transmission and semantic transmission.

[0125] The semantic transmission compresses the data packet, which can save bandwidth and improve anti-interference capability in a low signal-to-noise ratio condition, but the semantic transmission introduces coding and decoding delay, which is not friendly to users with extremely strict delay indicators. Therefore, the application can flexibly select the transmission mode of the service according to the delay requirement of the user and the channel condition.

[0126] Optionally, the factors for selecting the frame length include the delay requirement of the user, the channel quality, and the data packet size.

[0127] The frame length includes three different frame lengths of 0.25 milliseconds, 0.125 milliseconds, and 0.0625 milliseconds.

[0128] In a feasible implementation, for delay-sensitive services, the traditional 1ms frame length cannot meet the delay requirement of the user. A larger frame length seriously increases the size of the frame alignment delay; selecting a small frame length will result in frequent signaling overhead and reduce the spectrum efficiency. The factors for selecting the frame length mainly include the delay requirement of the user, the channel quality, and the data packet size; when the delay requirement of the user is strict, a small frame length is preferred; the control signaling includes scheduling authorization and related information for decoding; when the channel quality deteriorates, the control signaling overhead will be larger; and the size of the data packet will directly affect the transmission delay of the user, and a large data packet tends to use a larger frame length for transmission. Considering the delay requirement of the user, the channel quality, and the data packet size, the frame length size is selected for the user. In the application, the frame length can be selected from three different frame lengths of 0.25 milliseconds, 0.125 milliseconds, and 0.0625 milliseconds.

[0129] The calculation process of the semantic coding delay can be represented by the following formula (4):

[0130]

[0131] wherein, represents the calculation delay of semantic coding and decoding, represents the number of calculation cycles required for data recovery; β represents the processing density of the calculation terminal; f is the processing capacity of the calculation resource, such as the CPU frequency.

[0132] wherein, since there is a correlation between the semantic information recovery and the data packet size, the number of calculation cycles required for data recovery at the receiving end can be represented by the following formula (5):

[0133]

[0134] wherein: v k represents the size of the data packet; a1, a2 and a3 are constant parameters obtained through simulation; the reward function is 0 when the constraint condition of the algorithm is not met.

[0135] Optionally, the semantic compression ratio determines the size of the data packet.

[0136] wherein, the value range of the semantic compression ratio is between 0 and 1, and the greater the value indicates the higher the compression ratio and the smaller the data packet.

[0137] In a feasible implementation, the semantic compression ratio determines the size of the data packet, which will directly affect the size of the transmission delay. At the same time, the size of the semantic compression ratio determines the number of calculation cycles required for semantic compression, which will also affect the coding and decoding delay of the semantic part. Therefore, the present application can flexibly select the semantic compression ratio according to the delay requirement of the user and the channel condition.

[0138] In a feasible implementation, regarding the allocation of bandwidth, the total amount of transmission bandwidth in the system is certain, and all users use the transmission bandwidth together, and the bandwidth occupied by each user cannot coincide; in the present application, since different frame lengths are involved, the size of the resource block should be fully considered when allocating bandwidth; the allocation of bandwidth in the communication system involves time-frequency resources, when the frame length of the resource block is determined, the size of the corresponding frequency domain is also determined, and the frequency domain size is different for resource blocks of different frame lengths, and the larger the time domain is, the smaller the frequency domain is.

[0139] In a feasible implementation, the application adopts the EA3C algorithm to obtain the selection that maximizes the delay quality of the network, and the EA3C algorithm takes the completion degree of delay as an optimization target, and the specific completion degree of delay is represented by the ratio of the difference between the actually calculated delay and the delay requirement to the delay requirement. The constraint conditions of the EA3C algorithm include: (1) the delay violation probability calculated by the stochastic network calculus meets the reliability requirement; (2) the delay is less than the maximum delay tolerance limit of the user; and (3) the bandwidth of the user is less than the total bandwidth of the system.

[0140] In the method, the selection of the user in each round includes: a transmission mode of the service, a semantic compression ratio, a frame length, and a bandwidth allocation result as a policy input policy space, an expected return of taking a specific action in a given state is evaluated, an operation is selected from each state and the selected operation is evaluated in real time, and the selection of the user in the next round is selected based on the evaluation result.

[0141] S6, according to the transmission mode of the service, the semantic compression ratio, the frame length, and the bandwidth allocation result, the EA3C algorithm is adopted to make a decision on the delay reduction problem through the relationship between the delay and the reliability, and the maximum value of the delay quality is output; and the delay of the small data packet service transmission is reduced according to the maximum value of the delay quality.

[0142] The EA3C algorithm is an asynchronous advantage actor-critic algorithm, which is an algorithm widely used in the field of reinforcement learning, combines the policy gradient method and the learning of the value function, is used to approximately solve the Markov decision process problem, and is a method mastered by those skilled in the art, which will not be further described in the application.

[0143] Optionally, the EA3C algorithm is adopted in S3 to determine the service delay component, the frame length, the transmission mode of the service, the semantic compression ratio, and the bandwidth allocation result; the relationship between the delay and the reliability is used to make a decision on the delay reduction problem according to the service delay component, the transmission mode of the service, the semantic compression ratio, the frame length, and the bandwidth allocation result, including:

[0144] S31, the EA3C algorithm parameters are initialized, including: an actor learning rate, a critic learning rate, a discount factor, and a decay factor; global network parameters and target network parameters are initialized; a global shared optimizer is initialized; N parallel environment threads are started, wherein each thread initializes its local network parameters;

[0145] S32, each thread synchronizes parameters, copies the global network parameters to the local network, initializes the local environment, resets the local environment, and obtains an initial state;

[0146] S33, an action is performed and experience is collected, and an action is selected according to the current policy network;

[0147] S34, performing an action, obtaining a reward and a next state; storing the performed action, the initialized state reward and the next state into the experience sequence local to the thread;

[0148] S35, updating the local network and accumulating the gradient; when the length of the local experience sequence reaches a certain threshold T or a terminal state is encountered, the reward value from t to t+n is calculated, and the update target is calculated from back to front; the value function error and the policy gradient are calculated, and the total loss is calculated;

[0149] S36, updating the global network parameters, updating the global parameters using the gradient, synchronizing the local network parameters to the global parameters; continue to the next state, if the termination state is not reached, continue iteration; when all threads complete training or reach the maximum number of iterations, end the algorithm, complete the pre-training of the network; output the pre-trained Actor network and Critic network;

[0150] S37, defining the action space and state space according to the business delay component, the frame length, the transmission mode of the business, the semantic compression ratio and the bandwidth allocation result;

[0151] Wherein, the action space is defined as a={a1,a2,a3,a4};

[0152] Wherein, a1 is the transmission mode decision; a1=0 represents semantic transmission, and a1=1 represents bit transmission; a2 is the frame length size; a3 is the resource block quantity, i.e. bandwidth; a4 is the semantic compression rate.

[0153] In a feasible implementation, after each step of the algorithm is executed, the network state is updated and stored in the experience pool, and the state space is represented as s={s1,s2,s2,s4,s5}; wherein, s1 is the available resource, including the current time free bandwidth and available power; S2 is the channel quality, which has an important influence on the selection of transmission mode and bandwidth quantity; S3 is the user demand, including the delay and reliability requirement; S4 is the constraint violation times, which is used to quantify the negative influence of the action on the reward; S5 is the action selection of the last time.

[0154] S38, setting the completion degree of the delay as the reward function; according to the action space, the state space and the reward function, fine-tuning the pre-trained Actor network and Critic network through the relationship between the delay and the reliability, updating the network parameters, outputting the decision on the delay reduction problem, and executing the decision to complete the resource scheduling.

[0155] Wherein, the decision is executed, the maximum value of the delay quality is output, and the delay of the small data packet business transmission is reduced according to the maximum value of the delay quality.

[0156] The application embodiment firstly acquires the time delay requirement and reliability requirement of a user, constructs a wireless communication model of the ultra-reliable low time delay service according to the time delay requirement and reliability requirement, secondly, according to the wireless communication model of the ultra-reliable low time delay service, models the alpha-kappa-mu channel based on the random network calculus method, obtains the time delay violation probability formula under the alpha-kappa-mu channel, establishes the relationship between the time delay and the reliability according to the time delay violation probability formula, finally, adopts the EA3C algorithm to determine the service time delay component, frame length, service transmission mode, semantic compression ratio and bandwidth allocation result according to the time delay requirement of the user, makes a decision on the time delay reduction problem through the relationship between the time delay and the reliability according to the service time delay component, service transmission mode, semantic compression ratio, frame length and bandwidth allocation result, and outputs the maximum value of the time delay quality, so as to reduce the small data packet service transmission time delay.

[0157] The application can make each user flexibly select to extract part of information as semantic communication transmission or direct communication transmission, can select to use different frame lengths, and can determine the transmission bandwidth. The network determines the optimal selection of each user according to the unified optimization scheduling, so as to improve the overall network performance; the embodiment of the application combines the semantic communication and the direct communication, reduces the time delay of the service by flexibly adjusting the semantic compression ratio, frame length and transmission bandwidth for the small data packet service with time delay requirement.

[0158] Figure 2 It is a device block diagram for reducing small data packet service transmission time delay in a mobile communication network according to an exemplary embodiment, and the device is used for the method for reducing small data packet service transmission time delay in a mobile communication network. Referring to Figure 2 The device includes a first acquisition unit 210, a second acquisition unit 220 and an optimization and output unit 230. Wherein:

[0159] The first acquisition unit 210 is used for acquiring the time delay requirement and reliability requirement of a user, and constructing a wireless communication model of the ultra-reliable low time delay service according to the time delay requirement and reliability requirement;

[0160] The second acquisition unit 220 is used for modeling the alpha-kappa-mu channel based on the random network calculus method according to the wireless communication model of the ultra-reliable low time delay service, obtaining the time delay violation probability formula under the alpha-kappa-mu channel, and establishing the relationship between the time delay and the reliability according to the time delay violation probability formula;

[0161] An optimization and output unit 230 is configured to determine a service delay component, a frame length, a transmission mode of the service, a semantic compression ratio, and a bandwidth allocation result by using an EA3C algorithm according to the delay requirement of the user, and to make a decision on delay reduction by using a relationship between delay and reliability according to the service delay component, the transmission mode of the service, the semantic compression ratio, the frame length, and the bandwidth allocation result, and output a maximum value of delay quality to reduce the delay of small-packet service transmission.

[0162] Optionally, the second acquisition unit 220 is configured to:

[0163] (1) define a queue accumulation arrival process of the service to the base station, and determine the running state of the wireless communication system according to the queue accumulation process of the service to the base station;

[0164] (2) determine the dynamic change of the queue length according to the running state of the system, and determine the law of the change of the queue length with time;

[0165] (3) define the case of delay violation when the queue length exceeds the set maximum value by taking the delay violation probability as a reliability index, and obtain an upper bound expression of the delay violation probability;

[0166] (4) convert the upper bound expression of the delay violation probability into an expected form by using Markov inequality and Holder inequality, and obtain an expression related to the expected value of the delay violation probability;

[0167] (5) analyze the service arrival behavior by using a Poisson process according to the upper bound expression of the delay violation probability, calculate the characteristics of the cumulative arrival process by assuming that the data arrival rate and the packet size conform to the Poisson distribution, and describe the behavior of the service process of the wireless communication system according to the communication model;

[0168] (6) process the expression related to the expected value of the delay violation probability by using a Meijer G function according to steps (1)-(5), and obtain a final expression of the final delay violation probability.

[0169] Optionally, the service delay component includes a queuing delay, a transmission delay, a frame alignment delay, a semantic coding and decoding delay, and a base station and terminal processing delay.

[0170] The queuing delay is the time for which the data packet waits in the buffer queue, and the queuing delay increases with the increase of the load.

[0171] The transmission delay depends on the data packet size and the transmission rate.

[0172] The frame alignment delay is the delay generated when the system has to wait for the start of the next frame to transmit the data packet when the data packet arrives and the resource is ready for scheduling.

[0173] The semantic coding delay is related to the number of cycles required for calculation, processing density and device calculation capability.

[0174] The base station and terminal processing delay is related to the processing capability of the device itself.

[0175] Optionally, the selection factors of the frame length include: user's delay requirement, channel quality and data packet size.

[0176] The frame length includes three different frame lengths of 0.25 milliseconds, 0.125 milliseconds and 0.0625 milliseconds.

[0177] Optionally, the semantic compression ratio determines the size of the data packet.

[0178] The semantic compression ratio is in the range of 0 to 1, and the greater the value indicates the higher the compression ratio and the smaller the data packet.

[0179] Optionally, the transmission mode of the service is adaptive bit transmission and semantic transmission.

[0180] Optionally, the optimization and output unit 230 is configured to:

[0181] The initialization of the EA3C algorithm parameters includes: Actor learning rate, Critic learning rate, discount factor and decay factor; the initialization of the global network parameters and the target network parameters; the initialization of the global shared optimizer; the start of N parallel environment threads, wherein each thread initializes its local network parameters;

[0182] Each thread synchronizes parameters, copies global network parameters to local networks, initializes local environments, resets local environments, and obtains initial states.

[0183] Perform actions and collect experiences, and select actions according to the current policy network;

[0184] Perform actions, obtain rewards and next states; store the executed actions, initial states, rewards and next states into the experience sequence in the thread local;

[0185] Update the local network and accumulate the gradient; when the length of the local experience sequence reaches a certain threshold T or encounters a termination state, calculate the reward value from t to t+n, and calculate the update target from back to front: calculate the value function error and the policy gradient, and calculate the total loss.

[0186] Update the global network parameters, use the gradient to update the global parameters, and synchronize the local network parameters to the global parameters; continue to the next state, and if the terminal state has not been reached, continue to iterate; when all threads have completed training or the maximum number of iterations has been reached, end the algorithm and complete the pre-training of the network; output the pre-trained Actor network and Critic network;

[0187] Define the action space and state space based on the service delay components, frame duration, service transmission mode, semantic compression ratio, and bandwidth allocation results;

[0188] The degree of delay completion is set as the reward function. Based on the action space, state space, and reward function, the pre-trained Actor network and Critic network are fine-tuned through the relationship between delay and reliability, the network parameters are updated, and the decision on delay reduction is output. The decision is executed and resource scheduling is completed.

[0189] The implementation of the present invention first obtains the user's delay requirements and reliability requirements, and constructs a wireless communication model for ultra-reliable low-latency services based on the delay requirements and reliability requirements; secondly, based on the wireless communication model for ultra-reliable low-latency services and using a stochastic network calculation method, an α-κ-μ channel is modeled to obtain a delay violation probability formula under the α-κ-μ channel; based on the delay violation probability formula, a relationship between delay and reliability is established; finally, based on the user's delay requirements, an EA3C algorithm is used to determine service delay components, frame duration, service transmission mode, semantic compression ratio, and bandwidth allocation results; based on the service delay components, service transmission mode, semantic compression ratio, frame duration, and bandwidth allocation results, a decision is made on the delay reduction issue through the relationship between delay and reliability, and a maximum delay quality is output, thereby reducing the transmission delay of small data packet services.

[0190] This invention enables each user to flexibly choose to extract partial information for semantic communication or direct communication, use different frame durations, and determine the transmission bandwidth. The network determines the optimal choice for each user based on unified optimization scheduling, thereby improving overall network performance. This embodiment combines semantic communication with direct communication, reducing latency for small data packet services with latency requirements by flexibly adjusting the semantic compression ratio, frame duration, and transmission bandwidth.

[0191] Figure 3 is a structural diagram of a device for reducing transmission delay of small data packets in a mobile communication network provided by an embodiment of the present invention, such as Figure 3 As shown, the device for reducing the transmission delay of small data packets in the mobile communication network may include the above Figure 2The device for reducing small data packet service transmission delay in a mobile communication network is shown. Optionally, the device for reducing small data packet service transmission delay in a mobile communication network 310 can comprise a first processor 2001.

[0192] Optionally, the device for reducing small data packet service transmission delay in a mobile communication network 310 can further comprise a memory 2002 and a transceiver 2003.

[0193] The first processor 2001 is connected with the memory 2002 and the transceiver 2003, for example, through a communication bus.

[0194] The following will be specifically introduced in combination with Figure 3 The various components of the device for reducing small data packet service transmission delay in a mobile communication network 310 will be specifically introduced as follows:

[0195] The first processor 2001 is the control center of the device for reducing small data packet service transmission delay in a mobile communication network 310, and can be one processor or a collective term of multiple processing elements. For example, the first processor 2001 is one or more central processing units (CPU), or is an application specific integrated circuit (ASIC), or is one or more integrated circuits configured to implement the embodiments of the present application, such as one or more digital signal processors (DSP), or one or more field programmable gate arrays (FPGA).

[0196] Optionally, the first processor 2001 can execute various functions of the device for reducing small data packet service transmission delay in a mobile communication network 310 by running or executing software programs stored in the memory 2002 and calling data stored in the memory 2002.

[0197] In a specific implementation, as an embodiment, the first processor 2001 can comprise one or more CPUs, for example, the CPU0 and the CPU1 shown in Figure 3

[0198] In a specific implementation, as an embodiment, the device for reducing small data packet service transmission delay in a mobile communication network 310 can also comprise multiple processors, for example, the CPU0, the CPU1 and the CPU2 shown in Figure 3 ​1 and 2. The first processor 2001 and the second processor 2004 are shown in FIG. Each of these processors can be a single-core processor (single-CPU) or a multi-core processor (multi-CPU). A processor herein can refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0199] The memory 2002 is used to store the software program for executing the solution of the present invention, and is controlled by the first processor 2001 for execution. The specific implementation method can refer to the above method embodiment and will not be repeated here.

[0200] Alternatively, the memory 2002 may be a read-only memory (ROM) or other type of static storage device that can store static information and instructions, a random access memory (RAM) or other type of dynamic storage device that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compact disc, laser disc, optical disc, digital versatile disc, Blu-ray disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory 2002 may be integrated with the first processor 2001 or exist independently and be connected to the first processor 2001 through the interface circuit ( Figure 3 (not shown) is coupled to the first processor 2001, which is not specifically limited in this embodiment of the present invention.

[0201] The transceiver 2003 is used to communicate with a network device or a terminal device.

[0202] Optionally, the transceiver 2003 may include a receiver and a transmitter ( Figure 3 (not shown separately in the figure). The receiver is used to implement a receiving function, and the transmitter is used to implement a sending function.

[0203] Optionally, the transceiver 2003 may be integrated with the first processor 2001 or may exist independently and be connected to the first processor 2001 through the interface circuit of the device 310 for reducing the transmission delay of small data packets in the mobile communication network ( Figure 3 (not shown) is coupled to the first processor 2001, which is not specifically limited in this embodiment of the present invention.

[0204] It should be noted that, Figure 3 The structure of the mobile communication network small data packet service transmission delay reduction device 310 shown in the figure does not constitute a limitation on the router, and the actual knowledge structure identification device can include more or fewer components than the figure, or combine certain components, or different component arrangements.

[0205] In addition, the technical effects of the mobile communication network small data packet service transmission delay reduction device 310 can refer to the technical effects of the mobile communication network small data packet service transmission delay reduction method described in the above method embodiments, which will not be repeated here.

[0206] It should be understood that the first processor 2001 in the embodiments of the present application can be a central processing unit (CPU), and the processor can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0207] It should also be understood that the memory in the embodiments of the present application can be volatile or nonvolatile memory, or can include both volatile and nonvolatile memory. The nonvolatile memory can be read-only memory (ROM), programmable ROM (PROM), erasable PROM (EPROM), electrically EPROM (EEPROM), or flash memory. The volatile memory can be random access memory (RAM) used as external cache. By way of example, and not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous dynamic RAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), Synchlink DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0208] The above-described embodiments can be implemented in whole or in part by software, hardware (e.g., circuitry), firmware, or any combination thereof. When implemented in software, the above-described embodiments can be implemented in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded and executed on a computer, the processes or functions described in the embodiments of the present application are wholly or partially generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions can be transferred from one website, computer, server, or data center to another website, computer, server, or data center through wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible by a computer or a data storage device such as a server, data center, etc. containing one or more available medium collections. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. The semiconductor medium can be a solid-state disk.

[0209] It should be understood that the term "and / or" herein merely describes an association relationship of associated objects, which means that there can be three relationships, for example, A and / or B can represent three cases of A alone, A and B together, and B alone, where A and B can be singular or plural. In addition, the character " / " herein generally represents an "or" relationship between the front and rear associated objects, but can also represent an "and / or" relationship, which can be understood in the context before and after.

[0210] In the present application, "at least one" means one or more, and "multiple" means two or more. "At least one of the following" or the like means any combination of the items, including any combination of single or multiple items. For example, at least one of a, b, or c can represent a, b, c, a-b, a-c, b-c, or a-b-c, where a, b, and c can be single or multiple.

[0211] It should be understood that in various embodiments of the present application, the size of the sequence number of the above-described processes does not mean the order of execution, and the execution order of the processes should be determined by their functions and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0212] Those skilled in the art can clearly understand that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0213] Those skilled in the art can clearly understand that, for the convenience and brevity of the description, the specific working processes of the devices, apparatuses and units described above can refer to the corresponding processes in the foregoing method embodiments, which will not be repeated here.

[0214] In several embodiments provided by the present application, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely schematic, for example, the division of the units is only a logical function division, and actual implementation can have another division manner, for example, multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0215] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or can be distributed on multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment scheme.

[0216] In addition, each functional unit in each embodiment of the present application can be integrated into a processing unit, or each unit can exist physically independently, or two or more units can be integrated into one unit.

[0217] If the functions are realized in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application or the parts of the present application that essentially contribute to the prior art or the parts of the technical solutions can be embodied in the form of software products. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the various embodiments of the present application. The aforementioned storage medium includes a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0218] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical scope disclosed by the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for reducing transmission delay of small data packets in a mobile communication network, characterized in that: The method comprises: S1. Obtain the user's latency and reliability requirements, and build an ultra-reliable low-latency service wireless communication model based on the latency and reliability requirements; S2. According to the ultra-reliable low-latency wireless communication model, based on the random network calculation method, Channel modeling is performed to obtain A delay violation probability formula under a channel; establishing a relationship between delay and reliability based on the delay violation probability formula; S3. Based on the user's latency requirements, the experience-based asynchronous dominant actor-critic (EA3C) algorithm is used to determine the service latency components, frame duration, service transmission mode, semantic compression ratio, and bandwidth allocation results. Based on the service latency components, service transmission mode, semantic compression ratio, frame duration, and bandwidth allocation results, and taking into account the relationship between latency and reliability, a decision is made on latency reduction, and the maximum latency quality is output to reduce the transmission latency of small data packet services. The S3 employs the EA3C algorithm to determine service delay components, frame duration, service transmission mode, semantic compression ratio, and bandwidth allocation results. Based on the service delay components, service transmission mode, semantic compression ratio, frame duration, and bandwidth allocation results, and taking into account the relationship between delay and reliability, a decision on delay reduction is made, including: S31. Initialize EA3C algorithm parameters including: Actor learning rate, Critic learning rate, discount factor, and decay factor; initialize global network parameters and target network parameters; initialize the global shared optimizer; start N parallel environment threads, where each thread initializes its local network parameters; S32. Each thread synchronizes parameters, copies global network parameters to the local network, initializes the local environment, resets the local environment, and obtains the initial state; S33, execute actions and collect experience, and select actions based on the current strategy network; S34. Execute the action, obtain the reward and the next state; store the executed action, the initialization state reward, and the next state in the thread-local experience sequence; S35. Update the local network and accumulate gradients. When the length of the local experience sequence reaches a certain threshold T or encounters a termination state, calculate the reward value from t to t+n, and calculate the updated target from back to front: calculate the value function error and policy gradient, and calculate the total loss. S36. Update the global network parameters, use the gradient to update the global parameters, and synchronize the local network parameters to the global parameters; continue to the next state. If the terminal state has not been reached, continue to iterate; when all threads have completed training or the maximum number of iterations has been reached, end the algorithm and complete the pre-training of the network; output the pre-trained Actor network and Critic network; S37. Define an action space and a state space based on the service delay components, frame duration, service transmission mode, semantic compression ratio, and bandwidth allocation results; S38. Set the degree of delay completion as the reward function; based on the action space, state space, and reward function, fine-tune the pre-trained Actor network and Critic network through the relationship between delay and reliability, update the network parameters, output the decision on the delay reduction problem, execute the decision, and complete resource scheduling.

2. The method for reducing transmission delay of small data packets in a mobile communication network according to claim 1, wherein: The S2 is based on the random network calculus method, Channel modeling is performed to obtain The delay violation probability formula under the channel includes: S21. Define a queue accumulation process of services arriving at a base station; and determine an operating state of the wireless communication system based on the queue accumulation process of services arriving at the base station. S22. Determine the dynamic changes in queue length based on the system's operating status and clarify the pattern of queue length changes over time; S23. Using the delay violation probability as a reliability indicator, define a situation where a delay violation occurs when the queue length exceeds a set maximum value, and obtain an upper bound expression for the delay violation probability. S24. Using Markov's inequality and Holder's inequality, the upper bound expression of the delay violation probability is converted into the expected form, and the expression of the relationship between the delay violation probability and the expectation is obtained; S25. Based on the upper bound expression for the delay violation probability, a Poisson process is used to analyze the service arrival behavior. By assuming that the data arrival rate and packet size follow a Poisson distribution, the characteristics of the cumulative arrival process are calculated. Based on the small data packet transmission model, the behavior of the wireless channel service process is described. S26 . According to steps S21 - S25 , the Meijer G function is used to process the expression related to the delay violation probability and the expectation to obtain a final expression of the delay violation probability.

3. The method for reducing transmission delay of small data packets in a mobile communication network according to claim 1, wherein: The components of the service delay include: queuing delay, transmission delay, frame alignment delay, semantic encoding and decoding delay, and base station and terminal processing delay; The queuing delay is the time a data packet waits in the buffer queue, and the queuing delay increases with the increase of load; Wherein, the transmission delay depends on the data packet size and the transmission rate; The frame alignment delay is the delay incurred when the system must wait for the start of the next frame to transmit the data packet after the data packet arrives and resources are ready for scheduling. The semantic encoding and decoding delay is related to the number of cycles required for calculation, processing density and device computing power; The base station and terminal processing delays are related to the processing capabilities of the devices themselves.

4. The method for reducing transmission delay of small data packets in a mobile communication network according to claim 1, wherein: The factors for selecting the frame duration include: user's delay requirement, channel quality and data packet size; The frame duration includes three different frame durations: 0.25 milliseconds, 0.125 milliseconds, and 0.0625 milliseconds.

5. The method for reducing transmission delay of small data packets in a mobile communication network according to claim 1, wherein: The semantic compression ratio determines the size of the data packet; The semantic compression ratio ranges from 0 to 1. A larger value indicates a higher compression ratio and a smaller data packet.

6. The method for reducing transmission delay of small data packets in a mobile communication network according to claim 1, wherein: The transmission mode of the service is adaptive selection of bit transmission and semantic transmission.

7. A device for reducing transmission delay of small data packets in a mobile communication network, characterized in that: The device for reducing the transmission delay of small data packet services in the mobile communication network includes: The first acquisition unit, the second acquisition unit, and the optimization and output unit are used to implement the method for reducing the transmission delay of small data packet services in a mobile communication network as described in any one of claims 1-6.

8. A device for reducing transmission delay of small data packets in a mobile communication network, characterized in that: The device for reducing the transmission delay of small data packet services in the mobile communication network includes: processor; A memory having computer-readable instructions stored thereon, wherein when the computer-readable instructions are executed by the processor, the method according to any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores program code, which can be called by a processor to execute the method according to any one of claims 1 to 6.

Images