Multi-dimensional intelligent quality of service guarantee method and system for 6g hrllc

Abstract

The application discloses a multi-dimensional intelligent quality of service guarantee method and system for 6G HRLLC, and relates to the technical field of communication, comprising the following steps: acquiring the quality of service of multi-dimensional data, and constructing quality of service guarantee indexes of the multi-dimensional data; constructing a feasible region of the multi-dimensional data; performing feature extraction on the quality of service of the multi-dimensional data in the feasible region of the multi-dimensional data, so as to obtain a feature vector of the quality of service of the multi-dimensional data; processing the feature vector of the quality of service of the multi-dimensional data by using a trained deep learning network, obtaining a quality of service collaborative decision of the multi-dimensional data, and converting the quality of service collaborative decision of the multi-dimensional data into executable network instructions; and performing cross-layer scheduling and resource allocation according to the executable network instructions. The application can provide better multi-dimensional intelligent quality of service guarantee.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of communication, and particularly relates to a multi-dimensional intelligent quality of service guarantee method and system for 6G HRLLC. BACKGROUND

[0002] With the gradual development of the sixth generation mobile communication system (6G), it is widely used in industrial automation, Internet of vehicles, remote medical treatment, intelligent power grid, immersive virtual reality and holographic communication, etc. The common feature of these fields is extremely stringent requirements on reliability, time delay, time delay jitter and other indicators.

[0003] However, the existing quality of service guarantee mechanism is mostly single-dimensional statistical modeling, usually only considering time delay violation probability or reliability limit. In the 6G scenario, the traditional quality of service guarantee method often only optimizes a single indicator (such as time delay or error code probability), and cannot simultaneously meet the multi-dimensional constraints of time delay, jitter, peak information freshness, reliability, etc. It also cannot dynamically extract the coupling relationship between indicators, thus wasting a large amount of network resources. In addition, for the difficult-to-predict wireless network state, the traditional static scheduling and resource allocation method is difficult to maintain stable quality of service guarantee, and is mostly limited to local optimization at the physical layer or MAC layer, lacking cross-layer joint modeling and global optimization mechanism. In addition, the existing solution scheme is generally based on convex optimization or iterative solution mechanism, and it is difficult to predictively adjust the resource allocation strategy based on the multi-dimensional quality of service guarantee constraint, resulting in the system being unable to perceive and avoid the multi-dimensional quality of service guarantee violation risk in advance.

[0004] Therefore, it is urgent to provide a multi-dimensional intelligent quality of service guarantee method and system for 6G HRLLC to improve the defects in the prior art. SUMMARY

[0005] In order to solve the above problems existing in the prior art, the application provides a multi-dimensional intelligent quality of service guarantee method and system for 6G HRLLC. The technical problem to be solved by the application is solved by the following technical scheme:

[0006] In a first aspect, the application provides a multi-dimensional intelligent quality of service guarantee method for 6G HRLLC, comprising:

[0007] Obtaining the quality of service of multi-dimensional data, and constructing the quality of service guarantee indicators of multi-dimensional data; wherein the multi-dimensional data includes time delay, peak information freshness, jitter and reliability;

[0008] According to the current network state and service characteristic behavior, and in combination with the quality of service guarantee indicators of multi-dimensional data, the feasible region of multi-dimensional data is obtained; within the feasible region of multi-dimensional data, the quality of service of multi-dimensional data is characterized to obtain the characteristic vector of the quality of service of multi-dimensional data;

[0009] The trained deep learning network is used to process the feature vector of the service quality of the multi-dimensional data, obtain the service quality collaborative decision of the multi-dimensional data, and convert the service quality collaborative decision of the multi-dimensional data into executable network instructions.

[0010] According to the executable network instructions, cross-layer scheduling and resource allocation are performed, and whether the service quality state features of the multi-dimensional data after execution meet the multi-dimensional data service quality guarantee index is checked to form a closed-loop control cycle.

[0011] In the second aspect, the application further provides a multi-dimensional intelligent service quality guarantee system for 6G HRLLC, which is used to realize the multi-dimensional intelligent service quality guarantee method for 6G HRLLC provided above, and includes:

[0012] The data acquisition module is used to acquire the service quality of the multi-dimensional data, wherein the multi-dimensional data includes time delay, peak information freshness, jitter and reliability.

[0013] The basic module is used to construct the service quality guarantee index of the multi-dimensional data.

[0014] The perception module is used to obtain the feasible region of the multi-dimensional data according to the current network state and the service characteristic behavior, in combination with the service quality guarantee index of the multi-dimensional data, and extract the features of the service quality of the multi-dimensional data in the feasible region of the multi-dimensional data to obtain the feature vector of the service quality of the multi-dimensional data.

[0015] The intelligent module is used to process the feature vector of the service quality of the multi-dimensional data by using the trained deep learning network, obtain the service quality collaborative decision of the multi-dimensional data, and convert the service quality collaborative decision of the multi-dimensional data into executable network instructions.

[0016] The control module is used to perform cross-layer scheduling and resource allocation according to the executable network instructions, and check whether the service quality state features of the multi-dimensional data after execution meet the multi-dimensional data service quality guarantee index to form a closed-loop control cycle.

[0017] The application has the following beneficial effects:

[0018] The application provides a multi-dimensional intelligent quality of service guarantee method and system for 6G HRLLC, proposes a multi-dimensional quality of service guarantee method based on delay, peak value information freshness, jitter and reliability, solves the difficulty problem that the existing single-dimensional quality of service guarantee method cannot be applied to various application scenarios and multi-dimensional performance indicators by identifying and defining new statistical multi-dimensional quality of service guarantee indicators. On the basis of multi-dimensional quality of service guarantee constraints, an artificial intelligence algorithm is introduced, so that the system can realize real-time perception of network environment changes and dynamically adjust resource allocation strategies, thereby avoiding the problem of excessive conservatism caused by static design. In addition, the cross-layer collaborative design of the application uses the quality of service guarantee driven by the artificial intelligence algorithm to realize feature sharing and joint optimization between the physical layer, the MAC layer, the transmission layer and the application layer, breaks the layer segmentation, realizes the end-to-end multi-dimensional quality of service guarantee, and meets the real-time demand of 6G ultra-reliable low-latency communication while ensuring low computational complexity.

[0019] The application will be further described in detail below with reference to the accompanying drawings and embodiments. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 is a flow chart of the multi-dimensional intelligent quality of service guarantee method for 6G HRLLC provided by the embodiments of the application;

[0021] Figure 2 is another flow chart of the multi-dimensional intelligent quality of service guarantee method for 6G HRLLC provided by the embodiments of the application;

[0022] Figure 3 is a schematic diagram of the multi-dimensional intelligent quality of service guarantee system for 6G HRLLC provided by the embodiments of the application;

[0023] Figure 4 is a schematic diagram of the performance comparison of the AI-QoS mechanism, the baseline statistical delay QoS scheme and the URLLC optimized scheduler provided by the embodiments of the application;

[0024] Figure 5 is a schematic diagram of the violation distribution of each quality of service indicator of the AI-QoS provided by the embodiments of the application. DETAILED DESCRIPTION

[0025] The application will be further described in detail below with reference to the accompanying drawings and embodiments.

[0026] In the prior art, the multi-dimensional quality of service (QoS) guarantee for 6G ultra-reliable low-latency communication (HRLLC) still has the following deficiencies:

[0027] First, the limitations of single service quality assurance dimension design; existing methods mostly focus on statistical modeling and assurance of a single performance indicator (such as delay), lack joint modeling and optimization of multi-dimensional service quality assurance indicators (including delay, jitter, peak information freshness, reliability), leading to the inability of the system to balance multiple business or service demands when facing emerging complex business scenarios, often only targeting the most stringent indicators for "worst-case design", thus sacrificing overall system efficiency.

[0028] Second, the correlation between multi-dimensional service quality assurance indicators is not fully utilized; the statistical delay service quality assurance method of existing methods only handles the delay indicator independently, failing to capture the correlation and coupling relationship between delay, jitter, information freshness and reliability. This fragmented design limits the optimization effect and cannot effectively reduce the overall violation probability of service quality assurance globally.

[0029] Third, lack of dynamic adaptability and intelligent optimization capability; existing methods of resource scheduling and service quality assurance rely on fixed thresholds and static optimization methods, making it difficult to adapt to highly dynamic channel conditions and traffic fluctuations in 6G environments. Once the network environment changes, the preset scheduling strategy is likely to fail, leading to a decline in service level assurance and even a violation of the constraint probability.

[0030] Fourth, lack of cross-layer coordination and intelligent agent cooperation mechanism; current service quality assurance optimization is mostly limited to a single network layer (such as the MAC layer or the transport layer), lacking cross-layer joint optimization and distributed cooperation mechanisms. In multi-access point, multi-node 6G heterogeneous networks, the lack of coordination mechanisms leads to scattered resource allocation, making it difficult to achieve consistent service assurance globally.

[0031] Fifth, it is difficult to handle non-convex optimization and high-dimensional complex constraints; multi-service quality assurance is essentially a multi-objective, non-convex, time-varying optimization problem. Existing analytical methods are difficult to solve, and existing heuristic algorithms also have difficulty in ensuring convergence and real-time performance, especially in HRLLC scenarios where real-time performance is highly required. The computational complexity of existing algorithms often fails to meet the demand.

[0032] Therefore, the application provides a multi-dimensional intelligent service quality guarantee method and system for 6G HRLLC, proposes a multi-dimensional service quality guarantee method based on delay, peak information freshness, jitter and reliability, solves the difficulty that the existing single-dimensional service quality guarantee method cannot be applied to various application scenarios and multi-dimensional performance indicators by identifying and defining new statistical multi-dimensional service quality guarantee indicators. On the basis of multi-dimensional service quality guarantee constraints, an artificial intelligence algorithm is introduced, so that the system can realize real-time perception of network environment changes and dynamically adjust resource allocation strategies, thereby avoiding the problem of excessive conservatism caused by static design. In addition, the cross-layer collaborative design proposed by the application uses the service quality guarantee driven by the artificial intelligence algorithm to realize feature sharing and joint optimization between the physical layer, the MAC layer, the transmission layer and the application layer, breaks the layer segmentation, realizes the end-to-end multi-dimensional service quality guarantee, and meets the real-time demand of 6G ultra-reliable low-latency communication while ensuring low computational complexity.

[0033] Please refer to Figure 1 and Figure 2 , Figure 1 is a flowchart of the multi-dimensional intelligent service quality guarantee method for 6G HRLLC provided by the embodiments of the application, Figure 2 is another flowchart of the multi-dimensional intelligent service quality guarantee method for 6G HRLLC provided by the embodiments of the application, and the multi-dimensional intelligent service quality guarantee method for 6G HRLLC provided by the application comprises the following steps:

[0034] S101, acquire the service quality of multi-dimensional data, and construct the service quality guarantee indicators of multi-dimensional data; wherein the multi-dimensional data comprises delay, peak information freshness, jitter and reliability.

[0035] Specifically, in the embodiment, multi-modal data is collected by an intelligent device, and preprocessed, including capturing traffic patterns, user mobility, channel quality indicators, service level agreements (SLA) and device-specific indicators (such as energy level and processing delay), to obtain multi-dimensional data.

[0036] The service quality guarantee indicators of multi-dimensional data are constructed, including:

[0037] The tail behavior of the service quality violation probability of delay, peak information freshness, jitter and reliability is modeled by taking the statistical delay service quality guarantee model as a theoretical benchmark; optionally, the tail behavior refers to the decay law or probability of the probability distribution of the performance indicators (such as queuing delay and peak information freshness) in the extreme region when they exceed a certain large threshold, which describes the service quality of the system in the worst case.

[0038] In order to study the multi-dimensional service quality supply mechanism, the embodiments of the application propose a unified definition of multi-dimensional service quality guarantee indicators:

[0039] Let denote a random process of quality of service for multi-dimensional data, including latency, peak information freshness, jitter and reliability; using the large deviation principle, under sufficient conditions, the probability that the quality of service of multi-dimensional data violates the pre-set threshold satisfies:

[0040] ;

[0041] wherein, denotes the probability of quality of service violation of multi-dimensional data, denotes a random process of quality of service for multi-dimensional data, denotes a decay rate function corresponding to multi-dimensional data, for quantifying the exponential decay rate of violation probability driven by quality of service, denotes a pre-set threshold; as the pre-set threshold increases, the probability of quality of service violation decays exponentially at a fast rate, thereby providing a strict probability limit for extreme quality of service violation events, and the size of the decay rate function directly reflects the strength of quality of service guarantee; The smaller the value is, the slower the exponential decay of violation probability is, indicating that the system can only provide a relatively loose quality of service guarantee; The larger the value is, the faster the decay rate is, thereby being able to support more stringent quality of service requirements.

[0042] It should be noted that when , the violation probability no longer decays, indicating that the system can actually tolerate any degree of violation; when , the violation probability tends to zero, indicating that no violation is tolerated under the pre-set threshold. It should be noted that the exponential decay rate function described herein can be equally applied to the quality of service guarantee indicators of latency, peak information freshness, jitter and reliability described below.

[0043] According to the decay rate function corresponding to multi-dimensional data, a quality of service guarantee indicator for latency, a quality of service guarantee indicator for peak information freshness, a quality of service guarantee indicator for jitter and a quality of service guarantee indicator for reliability are respectively constructed.

[0044] 1. Quantitative characterization of statistical quality of service guarantee indicators oriented to latency. According to the large deviation principle, under sufficient conditions, the queue latency violation probability conforms to exponential decay, and the exponential decay rate is defined as the latency-based quality of service guarantee indicator, which reflects the strictness level of statistical latency boundary quality of service supply as the cache overflow threshold increases, considering the service data volume, cache state and latency violation probability; the larger the latency quality of service guarantee indicator value, the more the network can effectively reduce the latency risk by slightly adjusting its operating indicators, thereby guaranteeing a higher level of quality of service.

[0045] It should be noted that the network mentioned in this embodiment is a 6G network.

[0046] 2. Quantitative characterization of statistical quality of service guarantee indicators oriented to peak information freshness. The strictness of the quality of service guarantee metric of the peak information freshness constraint is determined by its corresponding peak information freshness quality of service guarantee index, which defines the rate of exponential decay of the peak information freshness violation probability as the peak information freshness threshold increases. For a specified peak information freshness threshold, its violation probability can be used to represent the tail characteristics of the peak information freshness. Therefore, a small enough peak information freshness quality of service guarantee index allows the system to accommodate any large peak information freshness, while a large enough index means that the system cannot tolerate any peak information freshness exceeding the threshold.

[0047] 3. Quantitative characterization of statistical quality of service guarantee indicators oriented to jitter. The jitter-constrained quality of service indicator is used to limit the allowable jitter variation range in the network, identify and enforce the deterministic delay violation threshold necessary for different service types to ensure that the network can meet the service level requirements of applications such as Voice over IP (VoIP), video streaming and real-time gaming; this indicator can evaluate and manage the packet delay fluctuations within strict limits to ensure that the delay does not exceed the preset threshold, which is crucial for maintaining data flow integrity. In time-sensitive applications, maintaining predictable low jitter is particularly critical for ensuring reliable quality of service.

[0048] 4. Quantitative characterization of statistical quality of service guarantee indicators oriented to reliability. In the field of finite code length, a measurement standard for reliability-oriented quality of service guarantee mechanism is constructed, which systematically describes the decay of error probability as the code length increases. According to the large deviation principle, when the coding rate is lower than the channel capacity, the error probability-based quality of service guarantee index represents the exponential decay rate of the reliability quality of service violation probability (i.e. error probability), and measures the strictness of statistical reliability quality of service guarantee increasing with block length.

[0049] S102, obtain a feasible region of the multi-dimensional data according to the current network state and the service characteristic behavior, in combination with the service quality guarantee index of the multi-dimensional data; and perform feature extraction on the service quality of the multi-dimensional data in the feasible region of the multi-dimensional data, to obtain a feature vector of the service quality of the multi-dimensional data.

[0050] Specifically, in this embodiment, the feasible region of the multi-dimensional data is obtained according to the network state and the service characteristic behavior, in combination with the service quality guarantee index of the multi-dimensional data, including:

[0051] According to the current network state, the boundary of the service quality of the multi-dimensional data is estimated, and the priority of the service quality guarantee index of the multi-dimensional data is dynamically adjusted according to the service characteristic behavior, in combination with the service quality guarantee index of the multi-dimensional data, to set the feasible region of the multi-dimensional data.

[0052] In this embodiment, the obtained multi-dimensional data is converted into an operable quality index, and the network state (including channel change, service arrival process and mobility dynamics) is mapped into an optimal operation strategy, for example, the adaptive scheduling, resource allocation and cooperative retransmission strategy of the current network are taken as the optimal operation strategy. Based on models such as queuing theory, effective capacity analysis and finite block code (FBC) theory, the statistical performance boundary such as delay upper bound, peak AoI distribution, reliability probability and jitter threshold is estimated under the current network state, and the statistical guarantee ability of the key service quality performance index is further evaluated. According to the service characteristic behavior, the priority of the service quality guarantee index of the multi-dimensional data is dynamically adjusted, in combination with the service quality guarantee index of the multi-dimensional data, to construct the feasible region of the multi-dimensional data, which is used to guide the feature extraction process.

[0053] It should be noted that the feasible region of the multi-dimensional data represents the internal trade-off relationship between the service quality guarantee indexes, for example, achieving an extremely low delay limit may weaken the control ability of the jitter, reflecting the mutual restriction between delay and its variance; in the case of limited spectrum resources, forcing an extreme peak information freshness may need to pay the price of reducing the reliability index. Through perceptual learning, a multivariate statistical learning method is used to capture the dependency relationship between delay, jitter, reliability and peak information freshness, and to identify the joint statistical structure between the multi-dimensional service quality dimensions.

[0054] Further, in this embodiment, feature extraction is performed on the service quality of the multi-dimensional data in the feasible region of the multi-dimensional data, to obtain a feature vector of the service quality of the multi-dimensional data, including:

[0055] The service quality oriented to delay is subjected to feature extraction, to obtain a delay distribution function, representing the sensitivity to delay;

[0056] The service quality oriented to peak information freshness is subjected to feature extraction, to obtain an information freshness statistic, representing the requirement for peak information freshness;

[0057] Feature extraction is performed on the quality of service (QoS) oriented towards jitter to obtain jitter variance, which characterizes the periodicity of traffic.

[0058] Feature extraction is performed on reliability-oriented service quality to obtain reliability curves, which characterize reliability constraints.

[0059] In this embodiment, the feasible domain of multidimensional data defines the acceptable latency, peak information freshness, jitter, and reliability boundaries. The feature extraction process is performed within this feasible domain, ensuring that all learning and adaptation processes are always feasible within this probability region.

[0060] S103. The trained deep learning network is used to process the feature vector of service quality of multidimensional data to obtain the service quality collaborative decision of multidimensional data, and the service quality collaborative decision of multidimensional data is transformed into executable network instructions.

[0061] Specifically, in this embodiment, the training process of the deep learning network includes:

[0062] Obtain training samples to construct a training dataset; wherein, the training samples include feature vectors of service quality of pre-defined multidimensional data, and the true labels of the training samples include the dependencies between multidimensional data;

[0063] Input a portion of the samples from the training dataset into the first... The deep learning network to be trained is then trained to obtain the first... The prediction results output during this training process;

[0064] According to the The prediction results output during the training process are the same as those during the training phase. The true labels of the samples for the next deep learning network to be trained are used to calculate the loss and serve as the first... Loss during each training session;

[0065] According to the The loss from the training process is backpropagated to update the loss from the training process. The network parameters of the deep learning network to be trained are obtained. The deep learning network to be trained is iterated until the number of training iterations or the degree of convergence meets the preset conditions, and a trained deep learning network is obtained.

[0066] Furthermore, the process of obtaining the dependencies between multidimensional data includes:

[0067] By employing a multivariate statistical learning method, the dependencies between latency, peak information freshness, jitter, and reliability are constructed through learning and approximating the Pareto optimal strategy.

[0068] In this embodiment, an offline multi-objective strategy based on multi-dimensional quality of service guarantee probability constraint is generated, and a deep reinforcement learning algorithm is used to select a suitable strategy from the offline strategy library and dynamically adjust it in combination with real-time network status to obtain a quality of service collaborative decision for multi-dimensional data, which decides how to allocate time-frequency resource blocks, transmit power and cache space and other resources to meet diversified quality of service guarantee requirements.

[0069] It should be noted that when training the deep learning network, the strategy in the offline strategy library is used.

[0070] S104, according to the executable network instruction, cross-layer scheduling and resource allocation are performed; at the same time, whether the service quality state characteristics of the multi-dimensional data after execution meet the multi-dimensional data service quality guarantee index is checked to form a closed loop control cycle.

[0071] Specifically, in this embodiment, the collaborative decision is converted into an executable network instruction, and the collaborative decision is not generated independently at a certain network layer, but is jointly optimized to ensure end-to-end multi-dimensional quality of service performance. For example, when a high-priority service flow encounters increased jitter, the system can simultaneously shrink the transmission interval, increase the signal power and adjust the processing queue of the edge server. In addition, the execution result is monitored and intelligently optimized, that is, the actual performance index after decision, such as the real-time delay experienced by the user, the packet loss event, etc., will be collected and fed back to the learning process, thereby forming a closed loop control cycle.

[0072] To sum up, the multi-dimensional intelligent quality of service guarantee method for 6G HRLLC provided by the present application has the following beneficial effects:

[0073] First, the present application fills the gap in the performance index of the traditional quality of service guarantee, re-identifies and defines the performance index system of multi-dimensional quality of service guarantee, and constructs an intelligent-driven unified multi-dimensional quality of service guarantee framework, and first proposes a multi-dimensional quality of service guarantee mechanism based on artificial intelligence, which effectively reflects the diversified and extreme performance index in special application scenarios, supports complete end-to-end multi-application scenarios in future wireless networks, and adapts to the dynamic environment of heterogeneous 6G networks.

[0074] Secondly, the application designs a security constraint mechanism based on a statistical multi-dimensional service quality guarantee boundary. Under network uncertainty conditions, interactive verification is performed through a boundary constraint model, a unified safety margin of the multi-dimensional service quality guarantee statistics is defined, it is ensured that the probability service guarantee is not violated, and the generated resource scheduling strategy is always within the mathematical boundary capable of providing the probability service guarantee, thereby realizing "intelligentization with constraints".

[0075] Based on the same inventive concept, please refer to Figure 3 , Figure 3 is a schematic diagram of a multi-dimensional intelligent service quality guarantee system for 6G HRLLC provided by an embodiment of the application. The application further provides a multi-dimensional intelligent service quality guarantee system for 6G HRLLC, which is used to implement the multi-dimensional intelligent service quality guarantee method for 6G HRLLC provided by the above embodiment. The embodiments of the method can be referred to above, and will not be described here again. The system comprises:

[0076] a data acquisition module configured to acquire the service quality of multi-dimensional data; wherein the multi-dimensional data comprises latency, peak information freshness, jitter and reliability;

[0077] a basic module configured to construct service quality guarantee indexes of the multi-dimensional data;

[0078] a perception module configured to obtain a feasible region of the multi-dimensional data according to a current network state and a service characteristic behavior, in combination with the service quality guarantee indexes of the multi-dimensional data; and perform feature extraction on the service quality of the multi-dimensional data in the feasible region of the multi-dimensional data to obtain a feature vector of the service quality of the multi-dimensional data;

[0079] an intelligent module configured to process the feature vector of the service quality of the multi-dimensional data by using a trained deep learning network to obtain a service quality collaborative decision of the multi-dimensional data, and convert the service quality collaborative decision of the multi-dimensional data into executable network instructions;

[0080] a control module configured to execute cross-layer scheduling and resource allocation according to the executable network instructions; and check whether a service quality state feature of the multi-dimensional data after execution meets the service quality guarantee indexes of the multi-dimensional data to form a closed-loop control cycle.

[0081] In an optional embodiment of the application, the system further comprises a coordination module configured to judge whether the executable network instructions are coordinated with the service quality guarantee indexes of the multi-dimensional data.

[0082] In an alternative embodiment of the application, the control module comprises a cross-layer scheduling module, a resource allocation module and a multi-dimensional data quality of service state feature guarantee module; wherein the cross-layer scheduling module is used to realize cross-layer scheduling of the network, the resource allocation module is used to realize resource allocation according to executable network instructions, and the multi-dimensional data quality of service state feature guarantee module is used to obtain the quality of service state feature of the executed multi-dimensional data.

[0083] In summary, the application proposes a hierarchical processing method, which integrates intelligent feature extraction, adaptive learning, federal collaboration and predictive resource management, provides agile, stable and customized low tail QoS guarantee services for quality of service requirements such as delay, peak information freshness, jitter, error code probability, etc., and is composed of the following hierarchical modules:

[0084] The basic module is responsible for collecting the underlying network resources, including channel state information (CSI), queue length, traffic intensity and historical multi-dimensional quality of service guarantee indicators.

[0085] The perception module: the data collected by the basic layer is subjected to feature extraction and modeling, and a multi-dimensional quality of service guarantee feature vector is generated, such as delay distribution function, jitter variance, information freshness statistics and reliability curve.

[0086] The intelligent module: the principle of "offline training, online decision" is adopted, that is, the generated offline strategy is used to train the deep learning network, and the trained deep learning network is used to learn the strategy online, multi-objective reinforcement learning and deep reinforcement learning are integrated, and high-dimensional network observation data is converted into dynamic control and resource allocation strategy. Among them, multi-objective reinforcement learning is used to explore the Pareto optimal strategy that can balance the mutually restrictive service demands to generate an offline strategy; deep reinforcement learning realizes low-delay decision-making under high-dynamic conditions, that is, the deep reinforcement learning trained based on the input of the perception layer can generate fine control operations in real time, including scheduling, resource slicing and power adaptation, etc.

[0087] The control module: the decision of the intelligent layer is converted into specific operations at the network level. The statistical multi-dimensional quality of service guarantee model is used as a filter of the action space to constrain and check the generated decision, and decisions on transmission power control, channel access control and priority scheduling are made. By enforcing the low tail bound guarantee of violation probability, it is ensured that the feasible operation region is always maintained, so that the adaptive process improves the performance while not damaging the statistical reliability.

[0088] The collaboration module: realizes federal learning and collaborative optimization between distributed intelligent agents, ensures end-to-end multi-dimensional quality of service guarantee, and realizes predictive risk control in the global range of the network.

[0089] To sum up, the multi-agent and multi-target optimization driven hierarchical and collaborative strategy provided by the application designs a hierarchical architecture (including a perception layer, an intelligence layer, a control layer and a collaboration layer), and realizes a complete closed loop from feature extraction, learning optimization, strategy execution and multi-agent collaboration.

[0090] In an optional embodiment of the application, the effect of the multi-dimensional intelligent service quality guarantee method for 6G HRLLC provided in the above embodiment is verified through simulation experiments, specifically:

[0091] Please refer to Figure 4 , Figure 4 is a schematic diagram of performance comparison of the AI-QoS mechanism provided by the embodiment of the application and the baseline statistical delay QoS scheme and the URLLC optimized scheduler, the AI-QoS mechanism (proposed AI-QoS) is the mechanism provided by the above embodiment of the application, the baseline statistical delay QoS scheme (baseline statistical QoS) and the URLLC optimized scheduler (URLLC optimized scheduler) are both schemes existing in the prior art, the baseline statistical delay QoS scheme only considers the delay boundary QoS constraint, and the URLLC optimized scheduler only considers the delay and error rate constraints, the performance evaluation adopts a multi-dimensional service quality violation probability (Multi-QoS Violation Probability, MQVP) as a unified index, and the index is used to quantify the probability of violating any condition in the delay, jitter, reliability and peak information freshness multi-service quality guarantee constraints. Figure 4The curves of the MQVP of the three schemes vary with the normalized network load level are shown, under the condition of low load (network load = 1), all schemes perform well, but the AI-QoS framework has shown obvious advantages: its MQVP drops to 2%, while the baseline statistical delay QoS scheme and the URLLC optimized scheduler reach 15% and 10% respectively; as the load increases, the performance difference gradually becomes significant: under the condition of medium load (network load = 3), the MQVP of the baseline statistical delay QoS scheme rises sharply to 60%, while the AI-QoS framework only maintains at 10%, indicating that it still has strong robustness in peak traffic environment. Under the condition of the highest load (network load = 5), the baseline statistical delay QoS scheme reaches complete saturation (MQVP = 100%), and the URLLC optimized scheduler also rises to 95%, while the AI-QoS framework remains at 30%. Compared with the baseline statistical delay QoS scheme and the URLLC optimized scheduler, the violation probability is reduced by 70% and 68.4% respectively; these improvements are due to the AI-QoS framework which can jointly predict and control multi-dimensional QoS indicators, dynamically reallocate resources and adaptively adjust the scheduling strategy according to the learned traffic pattern.

[0092] See Figure 5 , Figure 5 is a schematic diagram of the AI-QoS Fraction of Violations of each quality of service indicator provided by the embodiment of the application, which shows the relationship between different violation conditions of multi-dimensional quality of service indicators based on latency, peak information freshness (AoI), jitter and reliability and the network load level. When the network load increases, the proportion of peak information freshness remains basically unchanged, while the proportion of latency violation condition shows a clear downward trend. At the same time, the proportions of reliability and jitter violation conditions increase significantly. This shows that the greater the network load, the more stringent the requirements for jitter and reliability of the service.

[0093] It is to be understood that the terminology used herein such as first and second, and the like, is only intended to distinguish between one

[0094] In the description of the present specification, the description of the terms "one embodiment", "some embodiments", "an example", "a specific example", or "some examples" and the like means that the specific features or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In the present specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Also, the specific features or characteristics described can be combined in any suitable manner in one or more embodiments or examples. In addition, those skilled in the art can combine and combine different embodiments or examples described in the present specification.

[0095] The above is a further detailed description of the present application in conjunction with specific preferred embodiments, and cannot be considered as limiting the specific implementation of the present application to these descriptions. For those skilled in the art, without departing from the concept of the present application, a number of simple deductions or substitutions can be made, which should be considered as falling within the scope of protection of the present application.

Claims

1. A multi-dimensional intelligent quality of service guarantee method for 6G HRLLC, characterized in that, The method comprises the following steps: Obtaining the quality of service of multi-dimensional data, and constructing quality of service guarantee indexes of the multi-dimensional data; wherein the multi-dimensional data comprises latency, peak information freshness, jitter and reliability; wherein the construction of the quality of service guarantee indexes of the multi-dimensional data comprises: Using the large deviation principle, under sufficient conditions, the probability that the quality of service of the multi-dimensional data violates the preset threshold satisfies: ； wherein, a quality of service violation probability of the multidimensional data, a stochastic transformation process of the quality of service oriented to the multidimensional data, a decay rate function corresponding to the multidimensional data, used for quantifying an exponential decay rate of the quality of service driven violation probability, a preset threshold value; According to the decay rate function corresponding to the multi-dimensional data, the quality of service guarantee indexes for latency, the quality of service guarantee indexes for peak information freshness, the quality of service guarantee indexes for jitter and the quality of service guarantee indexes for reliability are respectively constructed; According to the current network state and the business characteristic behavior, the feasible region of the multi-dimensional data is obtained in combination with the quality of service guarantee indexes of the multi-dimensional data; the feature vector of the quality of service of the multi-dimensional data is obtained by performing feature extraction on the quality of service of the multi-dimensional data in the feasible region of the multi-dimensional data; wherein the feasible region of the multi-dimensional data is obtained in combination with the quality of service guarantee indexes of the multi-dimensional data according to the network state and the business characteristic behavior, and comprises: According to the current network state, the boundary of the quality of service of the multi-dimensional data is estimated, and the priority of the quality of service guarantee indexes of the multi-dimensional data is dynamically adjusted according to the business characteristic behavior; in combination with the quality of service guarantee indexes of the multi-dimensional data, the feasible region of the multi-dimensional data is set; The feature vector of the quality of service of the multi-dimensional data is processed by using the trained deep learning network, the quality of service collaborative decision of the multi-dimensional data is obtained, and the quality of service collaborative decision of the multi-dimensional data is converted into executable network instructions; According to the executable network instructions, cross-layer scheduling and resource allocation are performed; at the same time, it is checked whether the quality of service state characteristics of the multi-dimensional data after execution meet the quality of service guarantee indexes of the multi-dimensional data, so as to form a closed loop control cycle.

2. The multi-dimensional intelligent quality of service assurance method for 6G HRLLC of claim 1, wherein In the feasible region of the multi-dimensional data, the feature vector of the quality of service of the multi-dimensional data is obtained by performing feature extraction on the quality of service of the multi-dimensional data, and comprises: Performing feature extraction on the quality of service for latency to obtain a latency distribution function; Performing feature extraction on the quality of service for peak information freshness to obtain an information freshness statistical quantity; Performing feature extraction on the quality of service for jitter to obtain a jitter variance; Performing feature extraction on the quality of service for reliability to obtain a reliability curve.

3. The multi-dimensional intelligent quality of service assurance method for 6G HRLLC of claim 1, wherein, The training process of the deep learning network comprises: Obtaining training samples to construct a training data set; wherein the training samples comprise preset feature vectors of the quality of service of the multi-dimensional data, and the true labels of the training samples comprise the dependency relationship between the multi-dimensional data; inputting part of the samples in the training data set into a first deep learning network to be trained to obtain a first prediction result output in a first training process; inputting part of the samples in the training data set into a first deep learning network to be trained to obtain a first prediction result output in a first training process; inputting part of the samples in the training data set into a first deep learning network to be trained to obtain a first prediction result output in a first According to the first The prediction result output in the second training process is compared with the real label of the sample of the deep learning network to be trained, the loss is calculated, and the loss is taken as the loss of the first The prediction result output in the second training process is compared with the real label of the sample of the deep learning network to be trained, the loss is calculated, and the loss is taken as the loss of the first The prediction result output in the second training process is compared with the real label of the sample of the deep learning network to be trained According to the first loss of the first training process is back-propagated to update the network parameters of the deep learning network to be trained in the first time, to obtain the deep learning network to be trained in the first time; and the iteration is performed until the training times or the convergence degree meets the preset condition, to obtain the trained deep learning network.

4. The multi-dimensional intelligent quality of service assurance method for 6G HRLLC of claim 3, wherein, The process of obtaining the dependency relationship between the multi-dimensional data comprises: Using a multivariate statistical learning method, the dependency relationship between latency, peak information freshness, jitter and reliability is constructed by learning and approximating a Pareto optimal strategy.

5. A multi-dimensional intelligent service quality guarantee system for 6G HRLLC, configured to implement the multi-dimensional intelligent service quality guarantee method for 6G HRLLC according to any one of claims 1-4. The method comprises the following steps: A data acquisition module is configured to acquire the quality of service of multi-dimensional data; wherein the multi-dimensional data comprises latency, peak information freshness, jitter and reliability; A basic module is configured to construct a quality of service guarantee index of multi-dimensional data; wherein the quality of service guarantee index of multi-dimensional data includes: By using a large deviation principle, under sufficient conditions, a probability that a quality of service violation probability of multi-dimensional data is greater than a preset threshold satisfies: ； wherein, a quality of service violation probability of the multidimensional data, a random transformation process of the quality of service oriented to the multidimensional data, a decay rate function corresponding to the multidimensional data, used for quantifying an exponential decay rate of the quality of service driven violation probability, a preset threshold value; According to a decay rate function corresponding to the multi-dimensional data, a quality of service guarantee index for time delay, a quality of service guarantee index for peak information freshness, a quality of service guarantee index for jitter, and a quality of service guarantee index for reliability are respectively constructed; A perception module is configured to obtain a feasible region of multi-dimensional data according to a current network state and a service characteristic behavior, in combination with the quality of service guarantee index of multi-dimensional data; and perform feature extraction on the quality of service of the multi-dimensional data in the feasible region of multi-dimensional data to obtain a feature vector of the quality of service of multi-dimensional data; wherein the feasible region of multi-dimensional data is obtained according to the network state and the service characteristic behavior, in combination with the quality of service guarantee index of multi-dimensional data, and includes: According to the current network state, a boundary of the quality of service of multi-dimensional data is estimated, and a priority of the quality of service guarantee index of multi-dimensional data is dynamically adjusted according to the service characteristic behavior; in combination with the quality of service guarantee index of multi-dimensional data, the feasible region of multi-dimensional data is set; An intelligent module is configured to process the feature vector of the quality of service of multi-dimensional data by using a trained deep learning network to obtain a quality of service collaborative decision of multi-dimensional data, and convert the quality of service collaborative decision of multi-dimensional data into executable network instructions; A control module is configured to execute cross-layer scheduling and resource allocation according to the executable network instructions; and check whether a quality of service state feature of multi-dimensional data after execution meets the quality of service guarantee index of multi-dimensional data to form a closed-loop control cycle.

6. The 6G HRLLC-oriented multi-dimensional intelligent quality of service assurance system according to claim 5, characterized in that, Further comprising: A coordination module is configured to determine whether the executable network instructions are coordinated with the quality of service guarantee index of multi-dimensional data.

7. The 6G HRLLC-oriented multi-dimensional intelligent quality of service assurance system of claim 5, wherein, The control module includes a cross-layer scheduling module, a resource allocation module, and a quality of service state feature guarantee module of multi-dimensional data; wherein the cross-layer scheduling module is configured to implement cross-layer scheduling of a network, the resource allocation module is configured to implement resource allocation according to the executable network instructions, and the quality of service state feature guarantee module of multi-dimensional data is configured to obtain the quality of service state feature of multi-dimensional data after execution.

Images