Intelligent security capacity enhancement method and device for ensuring deterministic delay

Abstract

The application discloses a kind of intelligent safety capacity promotion method and device for guaranteeing deterministic delay, belong to computer network and communication engineering technical field, the method includes: constructing ultra-reliable low-latency communication (uRLLC) network uRLLC communication network model;For uRLLC communication network model, apply random network calculus technique (SNC) to obtain the deterministic delay violation probability general solution suitable for different channels, deduce the delay boundary of uRLLC communication network model;Based on the delay boundary of uRLLC communication network model, using machine learning algorithm, obtain the transmission resource configuration scheme of deterministic delay guarantee, maximize intelligent safety capacity under the premise of guaranteeing channel reliability.The technical scheme provided by the application can maximize channel safety capacity on the basis of guaranteeing deterministic delay, so as to improve network performance.

Description

Technical Field

[0001] This invention relates to the field of computer network and communication engineering technology, and in particular to an intelligent security capacity enhancement method and apparatus for ensuring deterministic latency. Background Technology

[0002] The sixth-generation mobile communication system (6G) is rapidly developing, with its core objective being to provide more flexible and secure services for intelligent uRLLC (uRLLC). uRLLC technology, with typical applications in autonomous driving, industrial automation, and the Internet of Things (IoT) based on touch sensing, places stringent demands on deterministic latency guarantees in networks. Simultaneously, the potential eavesdropping risks of wireless networks necessitate breakthroughs in physical layer security technologies.

[0003] Currently, a common method to improve channel security capacity is artificial noise injection. A typical example of this method is MIMO (Multi-Input Multiple-Output) technology, where, in the presence of an eavesdropper with multiple antennas, a transmitter with multiple antennas sends information to a target receiver with only one antenna. By using a portion of its power to generate "artificial noise," the transmitter can reduce the eavesdropper's channel quality, thus ensuring communication security. While this method can improve channel security capacity, it cannot guarantee deterministic latency.

[0004] Furthermore, current technologies focus on a single optimization objective, failing to simultaneously improve channel security capacity and guarantee deterministic latency, and do not take maximizing security capacity as an optimization objective. Additionally, existing technologies only derive solutions for specific channels, lacking general solutions applicable to different channels. Therefore, with the goal of improving intelligent security capacity, and under the constraint of guaranteeing deterministic latency, finding a general solution for different channels is an urgent problem to be solved. Summary of the Invention

[0005] This invention provides an intelligent security capacity enhancement method and apparatus that guarantees deterministic latency, thereby solving the technical problem that although existing technologies can enhance channel security capacity, they cannot guarantee deterministic latency.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] On the one hand, the present invention provides an intelligent security capacity enhancement method that guarantees deterministic latency, comprising:

[0008] Construct a uRLLC communication network model;

[0009] For the uRLLC communication network model, the deterministic delay violation probability general solution applicable to different channels is obtained by applying SNC, and the delay boundary of the uRLLC communication network model is derived.

[0010] Based on the delay boundary of the uRLLC communication network model, a machine learning algorithm is used to obtain a transmission resource configuration scheme with deterministic delay guarantee, thereby maximizing intelligent security capacity while ensuring channel reliability.

[0011] Furthermore, artificial noise is introduced into the uRLLC communication network model.

[0012] Furthermore, based on the delay boundary of the uRLLC communication network model, a machine learning algorithm is used to obtain a transmission resource configuration scheme with deterministic delay guarantee, maximizing intelligent security capacity while ensuring channel reliability, including:

[0013] A reinforcement learning model is constructed using transmission resource allocation schemes as actions;

[0014] Using the reinforcement learning model, the optimal transmission resource configuration scheme is obtained; wherein, the optimal transmission resource configuration scheme refers to the scheme that maximizes the channel security capacity while satisfying deterministic delay guarantee.

[0015] Furthermore, in the reinforcement learning model, the state space includes transmission time interval, bandwidth, and power allocation ratio; the action space includes multiple discrete and continuous actions, each action corresponding to a transmission resource configuration scheme, and each transmission resource configuration scheme includes a channel allocation scheme and transmission priority setting, used to provide the agent with diverse decision options; the reward function is calculated with power, bandwidth, and flexible TTI scheduling duration as variables; positive rewards are given to actions that meet latency, reliability constraints, and resource utilization, while penalties are imposed on those that do not, in order to guide the agent to learn the optimal strategy; where power includes transmission power and artificial noise power.

[0016] Furthermore, the reinforcement learning model combines discrete action selection with continuous parameter optimization, that is: after selecting a specific discrete action, the continuous parameters are further optimized;

[0017] The continuous parameter is a power parameter.

[0018] Furthermore, the discrete action selection is combined with continuous parameter optimization, specifically as follows:

[0019] During each discrete action selection and continuous parameter optimization process, the selected discrete action and the optimized continuous parameters are output through a preset action space decoupling network.

[0020] Furthermore, the action space decoupling network includes discrete action branches and continuous parameter branches; wherein,

[0021] The discrete action branch adopts an attention-based classification network, whose input includes the joint features of user QoS requirements and channel state matrix, and whose output is the probability distribution of each discrete action; wherein, the discrete action corresponding to the highest probability is the currently selected discrete action.

[0022] The continuous parameter branch uses a deep neural network. For the parameter characteristics of power allocation, an independent sub-network is designed to perform regression prediction to output optimized continuous parameters. The parameters share the underlying feature extraction layer to reduce training complexity.

[0023] The action space decoupling network employs joint policy optimization, including: using a proximal policy optimization algorithm, processing policy updates for discrete actions by pruning importance weights, and using Gaussian policy gradients for continuous parameter branches to achieve unbiased estimation of policy gradients in the mixed space.

[0024] On the other hand, the present invention also provides an intelligent secure capacity enhancement device that guarantees deterministic latency, the intelligent secure capacity enhancement device that guarantees deterministic latency includes:

[0025] The network model building module is used to build uRLLC communication network models;

[0026] The delay boundary derivation module is used to apply SNC to obtain a deterministic delay violation probability general solution applicable to different channels for the uRLLC communication network model, and derive the delay boundary of the uRLLC communication network model.

[0027] The resource allocation optimization module is used to obtain a transmission resource allocation scheme with deterministic delay guarantee based on the delay boundary of the uRLLC communication network model derived by the delay boundary derivation module, and to maximize intelligent security capacity while ensuring channel reliability.

[0028] In another aspect, the present invention also provides an electronic device comprising a processor and a memory; wherein the memory stores at least one instruction, which is loaded and executed by the processor to implement the above-described method.

[0029] In another aspect, the present invention also provides a computer-readable storage medium storing at least one instruction, which is loaded and executed by a processor to implement the above method.

[0030] The beneficial effects of the technical solution provided by this invention include at least the following:

[0031] The technical solution of this invention introduces an artificial noise method and obtains a deterministic delay violation probability general solution applicable to different channels based on random network calculus (SNC). This provides theoretical support for ensuring deterministic delay and can meet the requirements of security scenarios with strict delay and reliability requirements, avoiding the waste of resources based on worst-case optimization strategies. At the same time, by using flexible TTI technology, the channel security capacity can be maximized while ensuring deterministic delay. Attached Figure Description

[0032] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0033] Figure 1 This is a flowchart of the intelligent security capacity enhancement method for ensuring deterministic latency provided in an embodiment of the present invention;

[0034] Figure 2 This is a system block diagram of the electronic device provided in the embodiments of the present invention. Detailed Implementation

[0035] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0036] First, it should be noted that in the embodiments of the present invention, the words "exemplarily," "for example," etc., are used to indicate that they are examples, illustrations, or descriptions. Any embodiment or design scheme described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of the term "exemplarily" is intended to present the concept in a specific manner. Furthermore, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either one or the other.

[0037] First Embodiment

[0038] This embodiment provides an intelligent secure capacity enhancement method that guarantees deterministic latency. With intelligent secure capacity enhancement as the objective, it solves for general solutions under different channels under the constraint of deterministic latency guarantees. By combining flexible TTI (Time-to-Time Interval) technology and artificial noise technology, it maximizes intelligent secure capacity while ensuring channel reliability. This method can be implemented by electronic devices, and its execution flow is as follows: Figure 1 As shown, it includes the following steps:

[0039] S1, Construct the uRLLC communication network model;

[0040] It should be noted that the uRLLC communication network described in this invention mainly includes the following components:

[0041] (1) Base station: As the core node of the network, it is responsible for scheduling resources and initiating downlink data transmission; In this invention, artificial noise is generated by the base station and injected into the downlink channel to improve physical layer security.

[0042] (2) End-user equipment: including multiple user terminals, such as autonomous vehicles, remote-controlled robotic arms, industrial robots, etc., which are high-reliability, low-latency communication objects in typical uRLLC scenarios; the terminal receives task instructions or control data from the base station in the downlink;

[0043] (3) Potential eavesdroppers: illegal users or unauthorized devices that attempt to monitor downlink data transmission; the algorithm proposed in this invention introduces artificial noise, which significantly reduces the channel quality of eavesdroppers, thereby ensuring communication security.

[0044] (4) Channel model: The base station precodes the data to be transmitted through the α-κ-μ channel and sends it to the target user. At the same time, artificial noise is injected in the space-time dimension. This noise is canceled or ignored on the legitimate user channel, but forms interference in the unlicensed direction.

[0045] Furthermore, considering physical layer security, this embodiment introduces artificial noise into the network.

[0046] S2. For the uRLLC communication network model, SNC is applied to obtain the deterministic delay violation probability general solution applicable to different channels, and the delay boundary of the uRLLC communication network model is derived.

[0047] It's important to note that SNC, as a mathematical tool for analyzing and modeling random traffic behavior and performance boundaries in communication networks, can quantify the statistical characteristics of key performance indicators such as queuing delay and buffer usage of network nodes, providing theoretical support for ensuring deterministic latency. Using SNC to analyze latency violation probabilities in security scenarios can take into account the uncertainty and variability of traffic patterns and channel conditions in the network, better meeting the needs of security scenarios with stringent latency and reliability requirements, and avoiding the waste of resources from worst-case optimization strategies. For example, in 6G Open Radio Access Networks (O-RAN), for Ultra Reliable Low Latency Communication (URLLC) services, SNC can be used to model complex and dynamic O-RAN systems and analyze the latency violation probability of communication performance, i.e., the probability that data packet transmission exceeds the latency limit.

[0048] Based on the above, this embodiment derives the deterministic delay violation probability based on SNC:

[0049] We derive the formula using the SNC method. SNC provides end-to-end performance guarantees for data flows in a network and is suitable for communication networks that need to meet specific Quality of Service (QoS) requirements. This method can be used to analyze the latency performance of latency-sensitive services. Assuming that the service data packets generated in an Industrial IoT scenario adopt a First-In-First-Out (FIFO) transmission mechanism, we use SNC to model the service flow within the buffer to derive the latency boundary.

[0050]

[0051] in, Represents the shape parameters of the channel. The parameter represents shadow fading, and μ represents scaling. Represents the position parameter, where, Let θ represent the gamma function, be a positive parameter, have the subscript k representing the user channel, and have the subscript E representing the eavesdropper channel. Based on random network calculus, and combining the probability density function of the α-κ-μ channel with the Meijer G function, this embodiment ultimately derives a deterministic delay violation probability general solution applicable to different channels.

[0052] S3, based on the delay boundary of the uRLLC communication network model, uses machine learning algorithms to obtain a transmission resource configuration scheme with deterministic delay guarantee, maximizing intelligent security capacity while ensuring channel reliability.

[0053] It should be noted that this embodiment derives the deterministic delay violation probability based on SNC. In model building, deterministic delay is incorporated into the model's constraints to optimize safety capacity, and this is combined with Flexible Transmission Time Interval (Flexible TTI) technology (a mechanism that dynamically adjusts the transmission time interval of data packets in the network according to different needs) and artificial noise technology. A dual-network reinforcement learning architecture is used: a Q_Actor network and a ParamNet network are designed, and training stability is improved through experience replay buffer and dual network updates. A composite reliability assessment is performed based on bandwidth usage, TTI scheduling, power data, and safety capacity, allowing the agent to maximize safety capacity under constraints.

[0054] Specifically, in this embodiment, the implementation process of S3 is as follows:

[0055] S31, using the transmission resource allocation scheme as an action, constructs a reinforcement learning model;

[0056] S32, using a reinforcement learning model to obtain the optimal transmission resource allocation scheme; whereby the optimal transmission resource allocation scheme refers to the scheme that maximizes the channel security capacity while satisfying deterministic delay guarantees.

[0057] Reinforcement learning algorithms:

[0058] In the environmental modeling section, the state space encompasses key data such as transmission time interval, transmission mode, bandwidth, and power allocation ratio. The transmission time interval directly affects communication latency, and different transmission modes correspond to different transmission efficiencies and resource consumption. Bandwidth and power allocation ratio are core elements of resource allocation; they are interrelated and jointly determine communication quality. The action space sets 24 discrete actions, corresponding to various transmission resource configuration schemes, from channel allocation to transmission priority setting, providing the agent with diverse decision-making options.

[0059] Key performance indicators (KPIs) are calculated using power (channel power and artificial noise power), bandwidth, and TTI (Time-to-Interception) as variables. Power allocation must balance communication quality and energy consumption; artificial noise power is used to interfere with enemy eavesdropping and ensure communication security; dynamic bandwidth allocation determines the data transmission rate; flexible TTI scheduling can adjust transmission duration according to real-time service requirements and channel conditions, improving resource utilization. A composite reward mechanism integrates latency, reliability, and resource utilization objectives, providing positive rewards for actions that meet latency and reliability constraints and efficiently utilize resources, while penalizing actions that fail to meet these constraints, thus guiding the agent to learn the optimal strategy. The composite reward calculation formula is as follows:

[0060]

[0061] C1: This constraint is the probabilistic derivation of deterministic delay.

[0062] C2: It is a large-scale fading factor. For small-scale channel power gain, The transmit power allocated by the base station to user k. For one-sided noise power spectral density, For user k, the subcarrier bandwidth For user k's bandwidth, Refers to the power of artificial noise.

[0063] C3: This refers to artificial noise power, used to interfere with eavesdropping channels. Refers to the legal signal power. The total power limit is designed to meet the actual hardware capability requirements.

[0064] C4: The transmission delay for user k (the transmission time of data packets in the channel).

[0065] : Processing delay of user k (the calculation and processing time of data packets at the sending or receiving end).

[0066] Queuing delay for user k (waiting time for data packets in the buffer).

[0067] : The maximum allowed end-to-end delay for user k.

[0068] C5: Represents transmission errors for k users. This represents the maximum allowable transmission error.

[0069] C6: The amount of resources (bandwidth) allocated to user k. This represents the maximum total bandwidth of the system.

[0070] C7: It supports a variety of pre-configured TTIs (such as 0.125ms, 0.25ms, etc.).

[0071] The algorithm's key technologies are highly innovative. Hybrid action space processing combines discrete action selection with continuous parameter optimization. For example, after selecting a specific transmission mode (discrete action), power allocation parameters (continuous parameters) are further optimized, achieving joint optimization through an improved algorithm of Deep Q-Network (DQN). Channel modeling comprehensively considers both large-scale fading and small-scale Rayleigh fading. Large-scale fading describes the long-term attenuation trend of the signal with changes in distance and environment, while small-scale Rayleigh fading simulates the rapid signal fluctuations caused by multipath effects, more closely resembling the real channel environment. Reliability calculation is based on the α-κ-μ channel model, deriving the deterministic delay violation probability. This model accurately characterizes the signal transmission characteristics under complex channel conditions, providing a solid theoretical foundation for reliability assessment.

[0072] The training process is rigorous and scientific. During environment initialization, initial state parameters for each user are set; training parameters, including learning rate and discount factor, affect the agent's learning speed and long-term planning ability. In each step, the agent selects an action from the action space based on its current state. After the environment executes the action, a new state, reward, and completion flag are calculated based on the physical layer model. Experience data is stored in a memory pool, and an experience replay mechanism is used to break the correlation between data, improving training stability. Network parameter updates utilize a deep neural network to fit the value function, and the backpropagation algorithm is used to optimize the parameters. The training process lasts for 200 epochs, with a maximum of 1000 steps per epoch. During this period, indicators such as reward, energy consumption, and completion rate are recorded in real time to evaluate algorithm performance and analyze optimization directions.

[0073] Furthermore, to address the challenges of handling hybrid action spaces, this embodiment introduces an Action Decoupling Network (ADNet) to structurally separate discrete action selection from continuous parameter optimization. The specific implementation is as follows:

[0074] In this embodiment, the Action Space Decoupling Network (ADNet) includes a discrete action branch and a continuous parameter branch. The discrete action branch employs an attention-based classification network, taking into input joint features including user QoS requirements and the channel state matrix, and outputting the probability distribution of 24 discrete actions, thus addressing the action space sparsity problem caused by traditional One-Hot encoding. The continuous parameter branch constructs a multi-task deep neural network, designing independent sub-networks for regression prediction based on different parameter characteristics such as power allocation ratio and TTI scheduling duration. The parameters share the underlying feature extraction layer to reduce training complexity. Furthermore, the network employs joint policy optimization: using the Proximal Policy Optimization (PPO) algorithm, it processes policy updates for discrete actions by pruning importance weights, and applies Gaussian policy gradients to the continuous parameter branch to achieve unbiased estimation of policy gradients in the mixed space.

[0075] Specifically, the network input and output are defined as follows:

[0076] 1. Input:

[0077] User-side QoS requirements include maximum allowable latency and minimum reliability requirements.

[0078] Channel State Information (CSI) matrix: may include downlink channel gain, channel matrix H, noise power, etc. for multiple users;

[0079] Historical resource scheduling information: such as auxiliary features like the action selection at the previous moment and system feedback;

[0080] The above inputs are concatenated or embedded into a joint feature vector, which is then fed into the shared feature extraction layer of ADNet.

[0081] 2. Output:

[0082] Discrete action branch output: a 24-dimensional vector corresponding to 24 predefined discrete combination actions (such as different TTI lengths, number of subcarriers, artificial noise power allocation ratios, etc.), with each dimension being a probability value representing the probability of selecting that action at the current moment;

[0083] Continuous parameter branch output: Transmission power value, the output is a real regression value of the corresponding parameter dimension, which can be normalized according to the actual settings.

[0084] Furthermore, it should be noted that the Action Space Decoupling Network (ADNet) in this embodiment is a customized design based on a multi-task neural network architecture, optimized for hybrid action space structures. Its specific structure is as follows:

[0085] 1. Feature extraction layer (shared):

[0086] It consists of 2-3 layers of fully connected neural networks (each layer contains ReLU activation); the input is a joint feature vector, and the output is a shared high-dimensional semantic representation; the output of this layer serves as the input for both discrete and continuous branches.

[0087] 2. Discrete Action Branching (Structured Classification Network):

[0088] An attention mechanism module is introduced to dynamically assign weights to inputs of different dimensions, highlighting key QoS features; a fully connected layer is connected to output a 24-dimensional softmax probability distribution; compared with traditional One-Hot encoding, this structure can capture the potential correlation between actions and alleviate the sparsity problem.

[0089] 3. Continuous parameter branch (multi-task sub-network):

[0090] The regression objective is divided into different sub-tasks based on its physical meaning (such as power, noise ratio, and TTI duration). Each sub-task is constructed with a separate 1-2 layer regression sub-network, with shared features as input and continuous parameters as output. The parameters of different sub-networks are independent, but they share the same feature extraction layer, which helps the model generalize.

[0091] 4. Joint optimization strategy:

[0092] The discrete branch updates the policy using the policy gradient method in the PPO algorithm; the continuous branch uses the Gaussian policy gradient to predict the mean and sample actions based on the standard deviation; importance sampling weights are introduced to prune the uniform gradient update process and ensure the convergence of the mixed action space policy.

[0093] In summary, this embodiment considers physical layer security and latency requirements, and introduces a deterministic delay guarantee-based intelligent security capacity enhancement scheme. This scheme introduces artificial noise, and based on random network calculus, combines the probability density function of the α-κ-μ channel with the Meijer G function to derive a general solution for delay violation probabilities applicable to different channels. In simulations, this embodiment incorporates flexible TTI technology, using power (including channel power and artificial noise power), bandwidth, and flexible TTI scheduling as variables, and employs machine learning algorithms to maximize intelligent security capacity while ensuring channel reliability.

[0094] Second Embodiment

[0095] This embodiment provides an intelligent secure capacity enhancement device that ensures deterministic latency, including:

[0096] The network model building module is used to build uRLLC communication network models;

[0097] The delay boundary derivation module is used to apply SNC to obtain a deterministic delay violation probability general solution applicable to different channels for the uRLLC communication network model, and derive the delay boundary of the uRLLC communication network model.

[0098] The resource allocation optimization module is used to obtain a transmission resource allocation scheme with deterministic delay guarantee based on the delay boundary of the uRLLC communication network model derived by the delay boundary derivation module, and to maximize intelligent security capacity while ensuring channel reliability.

[0099] It should be noted that the intelligent security capacity enhancement device for ensuring deterministic latency in this embodiment corresponds to the intelligent security capacity enhancement method for ensuring deterministic latency in the first embodiment described above; the functions implemented by each functional module in the intelligent security capacity enhancement device for ensuring deterministic latency in this embodiment correspond one-to-one with each process step in the intelligent security capacity enhancement method for ensuring deterministic latency in the first embodiment described above; therefore, they will not be described again here.

[0100] Third Embodiment

[0101] This embodiment provides an electronic device, such as... Figure 2 As shown, the electronic device includes a processor and a memory; wherein the processor and the memory can be connected via a communication bus; the memory stores at least one instruction, which is loaded and executed by the processor to implement the method of the first embodiment described above. Furthermore, the electronic device may also include a transceiver, the processor and the transceiver can be connected via a communication bus, and the transceiver is used to communicate with other devices.

[0102] Below, in conjunction with Figure 2 A detailed introduction to each component of this electronic device is provided below:

[0103] The processor is the control center of the electronic device. The electronic device may include multiple processors, each of which can be a single-core processor (single-CPU) or a multi-core processor (multi-CPU). The term "processor" can refer to a single processor or a collective term for multiple processing elements. For example, a processor can be one or more central processing units (CPUs), other general-purpose processors, application-specific integrated circuits (ASICs), or one or more integrated circuits configured to implement embodiments of the present invention, such as one or more digital signal processors (DSPs), one or more field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor can perform various functions of the electronic device by running or executing software programs stored in memory and by calling data stored in memory.

[0104] In a specific implementation, as one example, the processor may include one or more CPUs, for example... Figure 2 CPU0 and CPU1 shown are, of course, merely illustrative examples.

[0105] The memory is used to store the software program that executes the solution of the present invention, and the processor controls its execution. For specific implementation methods, please refer to the above method embodiments, which will not be repeated here.

[0106] Optionally, the memory may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory may be integrated with the processor or exist independently, and may be accessed through the interface circuit of the electronic device ( Figure 2 (Not shown in the image) is coupled to the processor; however, this embodiment of the invention does not impose specific limitations on this.

[0107] The transceiver may include a receiver and a transmitter. Figure 2 (Not shown separately). The receiver is used to implement the receiving function, and the transmitter is used to implement the transmitting function. The transceiver can be integrated with the processor or exist independently, and can be connected through the interface circuit of the electronic device (…). Figure 2 (Not shown in the image) is coupled to the processor, and this embodiment of the invention does not specifically limit this.

[0108] In addition, it should be noted that, Figure 2 The structure of the electronic device shown is not intended to limit the device. Actual devices may include more or fewer components than shown, or combine certain components, or have different component arrangements. Furthermore, the technical effects achieved by this electronic device when performing the method of the first embodiment described above can be referenced to the technical effects described in the first embodiment; therefore, they will not be repeated here.

[0109] Fourth embodiment

[0110] This embodiment provides a computer-readable storage medium storing at least one instruction, which is loaded and executed by a processor to implement the method of the first embodiment described above. The computer-readable storage medium may be a ROM, random access memory, CD-ROM, magnetic tape, floppy disk, or optical data storage device, etc. The instruction stored therein can be loaded and executed by a processor in a terminal.

[0111] Furthermore, it should be noted that the present invention can be provided as a method, apparatus, or computer program product. Therefore, embodiments of the present invention can take the form of a completely or partially hardware embodiment, a completely or partially software embodiment, or an embodiment combining software and hardware aspects. Moreover, when implemented in software, embodiments of the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any usable medium accessible to a computer or a data storage device such as a server or data center containing one or more sets of usable media. 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. A semiconductor medium can be a solid-state drive (SSD).

[0112] Embodiments of the present invention are described with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0113] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1The functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable data processing terminal equipment to cause a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0114] It should also be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. The terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element. Furthermore, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone, where A and B can be singular or plural. Additionally, the character " / " in this text generally indicates an "or" relationship between the preceding and following objects, but it can also indicate an "AND / OR" relationship. Please refer to the context for specific interpretations. "At least one" refers to one or more items, while "more than" refers to two or more items. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or multiple items. For example, at least one of a, b, or c can be represented as: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0115] Furthermore, it is understood that in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0116] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented 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 implementations should not be considered beyond the scope of this invention.

[0117] In the several embodiments provided by this invention, 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 illustrative. For instance, the division of functional modules / units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the shown or discussed mutual couplings or direct couplings or communication connections may be through some interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms. Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, i.e., they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs. Additionally, the functional units in the various embodiments of this invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0118] If the method is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0119] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention. It should be pointed out that although preferred embodiments of the present invention have been described, those skilled in the art, once they understand the basic inventive concept of the present invention, can make several improvements and modifications without departing from the principles described herein. These improvements and modifications should also be considered within the scope of protection of the present invention. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present invention.

Claims

1. A method for intelligent security capacity enhancement that guarantees deterministic latency, characterized in that, include: A uRLLC communication network model based on the α-κ-μ channel is constructed; artificial noise is introduced into the uRLLC communication network model. For the uRLLC communication network model, the deterministic delay violation probability general solution applicable to different channels is obtained by applying SNC, and the delay boundary of the uRLLC communication network model is derived. Based on the delay boundary of the uRLLC communication network model, a machine learning algorithm is used to obtain a transmission resource configuration scheme with deterministic delay guarantee, maximizing intelligent security capacity while ensuring channel reliability, including: A reinforcement learning model is constructed using transmission resource allocation schemes as actions; Using the reinforcement learning model, the optimal transmission resource allocation scheme is obtained; wherein, the optimal transmission resource allocation scheme refers to the scheme that maximizes the channel security capacity while satisfying deterministic delay guarantee. In the reinforcement learning model, the state space includes transmission time interval, bandwidth, and power allocation ratio; the action space includes multiple discrete and continuous actions, each action corresponding to a transmission resource configuration scheme, and each transmission resource configuration scheme includes a channel allocation scheme and transmission priority setting, used to provide the agent with diverse decision options; the reward function is calculated using power, bandwidth, and flexible TTI scheduling duration as variables; positive rewards are given to actions that meet latency, reliability constraints, and resource utilization, while penalties are imposed to guide the agent to learn the optimal strategy; wherein, power includes transmission power and artificial noise power; In each discrete action selection and continuous parameter optimization process, the reinforcement learning model outputs the selected discrete action and the optimized continuous parameters through a pre-defined action space decoupling network; wherein, the continuous parameters are power parameters. The action space decoupling network includes discrete action branches and continuous parameter branches; wherein... The discrete action branch adopts an attention-based classification network, whose input includes the joint features of user QoS requirements and channel state matrix, and whose output is the probability distribution of each discrete action; wherein, the discrete action corresponding to the highest probability is the currently selected discrete action. The continuous parameter branch uses a deep neural network. For the parameter characteristics of power allocation, an independent sub-network is designed to perform regression prediction to output optimized continuous parameters. The parameters share the underlying feature extraction layer to reduce training complexity. The action space decoupling network employs joint policy optimization, including: using a proximal policy optimization algorithm, processing policy updates for discrete actions by pruning importance weights, and using Gaussian policy gradients for continuous parameter branches to achieve unbiased estimation of policy gradients in the mixed space.

2. An intelligent secure capacity enhancement system that guarantees deterministic latency, characterized in that, include: The network model building module is used to construct the uRLLC communication network model; artificial noise is introduced into the uRLLC communication network model. The delay boundary derivation module is used to apply SNC to obtain a deterministic delay violation probability general solution applicable to different channels for the uRLLC communication network model, and derive the delay boundary of the uRLLC communication network model. The resource allocation optimization module is used to obtain a transmission resource allocation scheme with deterministic delay guarantee based on the delay boundary of the uRLLC communication network model derived by the delay boundary derivation module, using machine learning algorithms, and maximizing intelligent security capacity while ensuring channel reliability. This includes: A reinforcement learning model is constructed using transmission resource allocation schemes as actions; Using the reinforcement learning model, the optimal transmission resource allocation scheme is obtained; wherein, the optimal transmission resource allocation scheme refers to the scheme that maximizes the channel security capacity while satisfying deterministic delay guarantee. In the reinforcement learning model, the state space includes transmission time interval, bandwidth, and power allocation ratio; the action space includes multiple discrete and continuous actions, each action corresponding to a transmission resource configuration scheme, and each transmission resource configuration scheme includes a channel allocation scheme and transmission priority setting, used to provide the agent with diverse decision options; the reward function is calculated using power, bandwidth, and flexible TTI scheduling duration as variables; positive rewards are given to actions that meet latency, reliability constraints, and resource utilization, while penalties are imposed to guide the agent to learn the optimal strategy; wherein, power includes transmission power and artificial noise power; In each discrete action selection and continuous parameter optimization process, the reinforcement learning model outputs the selected discrete action and the optimized continuous parameters through a pre-defined action space decoupling network; wherein, the continuous parameters are power parameters. The action space decoupling network includes discrete action branches and continuous parameter branches; wherein... The discrete action branch adopts an attention-based classification network, whose input includes the joint features of user QoS requirements and channel state matrix, and whose output is the probability distribution of each discrete action; wherein, the discrete action corresponding to the highest probability is the currently selected discrete action. The continuous parameter branch uses a deep neural network. For the parameter characteristics of power allocation, an independent sub-network is designed to perform regression prediction to output optimized continuous parameters. The parameters share the underlying feature extraction layer to reduce training complexity. The action space decoupling network employs joint policy optimization, including: using a proximal policy optimization algorithm, processing policy updates for discrete actions by pruning importance weights, and using Gaussian policy gradients for continuous parameter branches to achieve unbiased estimation of policy gradients in the mixed space.

Images