Wireless resource service quality full-dimension evaluation method and system, storage medium and terminal

Abstract

The invention provides a wireless resource service quality full-dimensional evaluation method and system, a storage medium and a terminal. The method comprises the following steps: acquiring multi-dimensional network service quality sensing information provided by a user terminal for spectrum resources, environment and computing power resources; based on the multi-dimensional network service quality perception information, obtaining an evaluation result of the service quality of the wireless resources according to a multi-task learning neural network, so that a scheduling module performs priority division and dynamic rescheduling on the wireless resources based on the evaluation result; and obtaining a scheduling execution result of the dynamic rescheduling, and correcting the multi-task learning neural network based on the scheduling execution result and the multi-dimensional network service quality perception information. According to the wireless resource service quality full-dimension evaluation method and system, the storage medium and the terminal, panoramic perception and dynamic optimization of frequency spectrums, environments and computing power resources are realized based on a full-dimension index system, and resource allocation intelligence and service quality multi-dimension guarantee can be realized.

Description

Technical Field

[0001] The present invention relates to the technical field of wireless communication, and particularly to a method, a system, a storage medium and a terminal for comprehensively evaluating the quality of service of wireless resources in all dimensions. Background Art

[0002] With the accelerating evolution of wireless communication technology from the fifth Generation Mobile Communication System (5G) to the sixth Generation Mobile Communication System (6G), the wireless network is facing unprecedented challenges in scenario complexity and service diversification.

[0003] On the application demand side, emerging scenarios such as the Internet of Things, industrial Internet, intelligent transportation, and holographic communication have put forward strict and refined requirements for their service quality. For example, industrial automation control requires extreme performance with an end-to-end delay lower than 1 ms and a jitter error less than 1 μs; holographic communication requires dual guarantees of a peak rate of the order of 10 Gbps and a spatial synchronization accuracy of sub-millimeter level. These requirements have far exceeded the capacity boundary of the traditional wireless resource evaluation system. The limitations of the existing evaluation system centered on single-dimensional indicators such as throughput and packet loss rate are systematically amplified in the 6G multi-domain fusion scenario. It can neither dynamically quantify the collaborative efficiency of multi-domain resources such as spectrum, environment, and computing power, nor can it meet the differentiated service requirements of scenarios such as ultra-reliable low-latency communication and Massive Machine Type Communications (mMTC), revealing the fundamental contradiction between the static evaluation model and the dynamic service environment.

[0004] On the technology supply side, 6G is driving a structural transformation in the wireless resource management paradigm by deeply integrating disruptive frontier technologies such as artificial intelligence (AI), terahertz communication, intelligent metasurfaces, and quantum communication. For example, the AI-driven dynamic spectrum sharing strategy can fuse deep reinforcement learning methods to achieve real-time sensing and reasonable allocation of spectrum holes, which requires the introduction of new metrics such as dynamic spectrum efficiency and anti-interference tolerance. Terahertz band communication is vulnerable to environmental temperature and humidity, as well as obstacle occlusion, and it is necessary to build metrics for the fusion accuracy of multi-modal environmental data and the real-time performance of modeling to maintain link stability. The deep collaboration between edge computing and network functions requires metrics such as the flexibility of heterogeneous computing power scheduling and the resource fragmentation index to achieve the global optimum of task offloading and resource allocation. Such technological evolution has made the intelligent, high-dimensional, and heterogeneous characteristics of wireless resources more prominent. Due to defects such as staticity and isolation, the traditional evaluation system has difficulty adapting to the needs of new technologies for dynamic environment response, cross-domain collaborative optimization, and fine-grained parameter measurement, and has even become a bottleneck restricting the release of the huge potential of 6G.

[0005] Therefore, it can be seen that for the existing wireless resource service quality (QoRS) evaluation system centered on single-dimensional metrics such as throughput, packet loss rate, and spectrum efficiency, its limitations are systematically amplified in the 6G multi-domain fusion scenario, and the following deficiencies have been exposed.

[0006] a) Traditional models only evaluate single resource dimensions such as spectrum, power, and time slots in isolation, and cannot quantify the collaborative efficiency of cross-domain resources (such as the joint optimization of terahertz band dynamic blockage avoidance and edge computing power scheduling).

[0007] b) The static threshold evaluation mechanism (such as the signal-to-noise ratio measurement with a fixed threshold) is seriously out of touch with the dynamic service environment. In complex scenarios with sudden interference, the system bit error rate will face a sharp increase.

[0008] c) The existing system lacks measurement standards for new resource management parameters driven by AI (such as the convergence degree of the federated learning model and the beamforming accuracy of intelligent reflectors), resulting in a large efficiency loss in the actual deployment of AI-enabled spectrum sharing strategies.

[0009] The international standardization process has further highlighted the urgency of system innovation. ITU-R lists "native intelligent radio resource management" as one of the core capabilities of 6G in the "IMT-2030 Framework Recommendation", and 3GPP focuses on resource optimization models empowered by machine learning. These measures all point to a consensus: there is an urgent need to establish an evaluation system that is compatible with the standardization process and can quantify the effectiveness of new technologies. At the same time, the demand for "deterministic service quality" (such as bounded latency and self-healing of faults) in scenarios such as industrial metaverse and digital twin is driving the network architecture to leap from best-effort to precise guarantee. Summary of the Invention

[0010] In view of the above problems, the purpose of the present invention is to provide a full-dimensional evaluation method, system, storage medium and terminal for the service quality of radio resources, which can realize panoramic perception and dynamic optimization of spectrum, environment and computing power resources based on a full-dimensional index system, and help to realize intelligent resource allocation and multi-dimensional guarantee of service quality.

[0011] In a first aspect, the present invention provides a full-dimensional evaluation method for the service quality of radio resources, and the method includes the following steps: obtaining multi-dimensional network service quality perception information provided by a user terminal for spectrum resources, environment and computing power resources; based on the multi-dimensional network service quality perception information, obtaining an evaluation result of the service quality of radio resources according to a multi-task learning neural network, so that a scheduling module performs priority division and dynamic rescheduling on radio resources based on the evaluation result; obtaining the scheduling execution result of the dynamic rescheduling, and correcting the multi-task learning neural network based on the scheduling execution result and the multi-dimensional network service quality perception information.

[0012] In one implementation manner of the first aspect, obtaining the multi-dimensional network service quality perception information provided by the user terminal includes the following steps: Performing normalization processing on spectrum resource perception information, environment perception information and computing power resource perception information to obtain normalized perception information; Performing spatio-temporal alignment on the normalized perception information to obtain aligned perception information; Performing semantic coding on the aligned perception information to obtain perception features; Performing vector compression on the perception features to obtain a multi-dimensional network service quality perception vector as the multi-dimensional network service quality perception information.

[0013] In an implementation of the first aspect, the evaluation result includes the total score of the quality of wireless resource service and the score values of different evaluation dimensions; the evaluation dimensions include the performance, security, overhead, and autonomy of wireless resources; the performance includes digital performance sub-items, algorithm performance sub-items, computing power performance sub-items, and connection performance sub-items; the security includes data security sub-items and algorithm security sub-items; the overhead includes storage overhead sub-items, computing overhead sub-items, transmission overhead sub-items, and energy consumption overhead sub-items; the autonomy includes data autonomy sub-items and algorithm autonomy sub-items.

[0014] In an implementation of the first aspect, based on the multi-dimensional network service quality perception information, obtaining the evaluation result of the wireless resource service quality according to the multi-task learning neural network includes the following steps: Input the multi-dimensional network service quality perception information into the trained multi-task learning neural network; Obtain the total score of the wireless resource service quality output by the multi-task learning neural network , where N represents the total number of sub-items of each evaluation dimension, x i represents the i th multi-dimensional network service quality perception information corresponding to the sub-items of the evaluation dimension, f i (·) represents the sub-score function corresponding to the i th sub-item of the evaluation dimension, α i represents the i th weight coefficient corresponding to the sub-items of the evaluation dimension and ∑ α i = 1; Obtain the score value of any evaluation dimension , where M represents the total number of sub-items corresponding to the evaluation dimension, x j represents the j th multi-dimensional network service quality perception information corresponding to the sub-items of the evaluation dimension, f j (·) represents the sub-score function corresponding to the j th sub-item of the evaluation dimension, α j represents the j th weight coefficient corresponding to the sub-items of the evaluation dimension; Take the total score of the wireless resource service quality and the score value of the evaluation dimension as the evaluation result.

[0015] In an implementation of the first aspect, the dynamic scheduling module performs priority division and dynamic rescheduling of radio resources based on the evaluation result, including the following steps: Obtain the i radio resource scheduling priority of the th user, where QoRS (i) is the total radio resource service quality score of the i th user, β ∈[0,1] is an adjustable weight coefficient; , QoRS (i) best is the optimal total score of the radio resource service quality of the i th user within the historical window, ε is a small positive number introduced to prevent the denominator from being zero; Based on the radio resource scheduling priority p i and the schedulable resource set, the system obtains the dynamic scheduling result , where A = a ij , a ij ∈{0,1} indicates whether to allocate the resource j to the user i , w ij The resource j allocated to the user i 's weight.

[0016] In an implementation of the first aspect, based on the scheduling execution result and the multi-dimensional network service quality perception information, correcting the multi-task learning neural network includes the following steps: Calculate the deviation index between the scheduling execution result and the multi-dimensional network service quality perception information , where QoS post (i) represents the actual measurement value of the i th multi-dimensional network service quality perception information, QoS pred (i) represents the predicted value of the i th multi-dimensional network service quality perception information before this round of scheduling, δ i represents the deviation index between the i th multi-dimensional network service quality perception information and the corresponding scheduling execution result; When the deviation index is greater than the preset threshold, retrain or fine-tune the multi-task learning neural network and adjust the weight coefficient of the radio resource priority.

[0017] Second aspect, the present invention provides a system for all-dimensional evaluation of wireless resource service quality, the system comprising an acquisition module, an evaluation module and a feedback module; The acquisition module is used to acquire multi-dimensional network service quality perception information provided by a user terminal for spectrum resources, environment and computing power resources; The evaluation module is used to obtain an evaluation result of the wireless resource service quality based on the multi-dimensional network service quality perception information according to a multi-task learning neural network, so that a scheduling module performs priority division and dynamic rescheduling on the wireless resources based on the evaluation result; The feedback module is used to acquire the scheduling execution result of the dynamic rescheduling, and correct the multi-task learning neural network based on the scheduling execution result and the multi-dimensional network service quality perception information.

[0018] Third aspect, the present invention provides a storage medium, on which a computer program is stored, and when the program is executed by a processor, the above-mentioned all-dimensional evaluation method for wireless resource service quality is implemented.

[0019] Fourth aspect, the present invention provides an evaluation terminal, comprising: a processor and a memory; The memory is used to store a computer program; The processor is used to execute the computer program stored in the memory, so that the evaluation terminal executes the above-mentioned all-dimensional evaluation method for wireless resource service quality.

[0020] Fifth aspect, the present invention provides a system for all-dimensional evaluation of wireless resource service quality, comprising the above-mentioned evaluation terminal, a perception module, a scheduling module and an edge control module; The perception module is used to collect multi-dimensional network service quality perception information provided by a user evaluation terminal for spectrum resources, environment and computing power resources, and provide it to the evaluation terminal; The scheduling module is used to perform priority division and dynamic rescheduling on the wireless resources based on the evaluation result provided by the evaluation terminal; The edge control module is used to generate the scheduling execution result of the dynamic rescheduling, and provide it to the evaluation terminal.

[0021] As described above, the all-dimensional evaluation method, system, storage medium and terminal for wireless resource service quality of the present invention have the following beneficial effects.

[0022] (1)A multi-dimensional and multi-granularity evaluation index system is constructed, integrating four core dimensions of performance, security, overhead, and autonomy, covering multiple levels from the physical state of the link, resource usage to service response performance, and user experience perception. For typical application scenarios in 6G, a customizable and evolvable quality of service (QoS) index system can be realized to ensure the accuracy and practicality of evaluation under complex communication conditions. Through the multi-dimensional integration based on quality of service (QoS), quality of experience (QoE), and network capabilities (such as resource status, channel quality, energy consumption), the impact of wireless resource allocation on service quality in the 6G network environment can be more accurately characterized, improving the evaluation accuracy and coverage breadth.

[0023] (2)A state awareness mechanism based on machine learning is introduced, which can realize the real-time monitoring of current space-air-ground integrated network resources (such as low-earth orbit satellite channels, edge computing nodes, air platform spectrum resources, etc.), be able to track the changes in wireless resource status and user requirements in real time, extract features and optimize models by combining machine learning methods, support the rapid evaluation before scheduling decisions, improve the forward-looking and accuracy of resource scheduling, and dynamically update the evaluation model through a feedback mechanism to achieve a rapid response to network changes, improving the agility and adaptability of the system.

[0024] (3)By taking the multi-dimensional QoRS evaluation results as the key input of the resource scheduling algorithm, the upper-layer intelligent control module can realize the cross-layer and cross-domain intelligent collaborative optimization of various resources such as spectrum, power, cache, and routing. It not only significantly improves the system throughput and energy efficiency but also effectively reduces the communication delay, realizes the optimization of key indicators, and enhances the robustness and service reliability of the network in complex scenarios.

[0025] (4)It can provide differentiated and multi-level service quality guarantee strategies according to different business scenarios, including enhanced mobile broadband (eMBB), massive machine type communication (mMTC), ultra-reliable low-latency communication (uRLLC), etc., realize the transformation from "resource priority" to "service priority", and ultimately improve the comprehensive user perception experience.

[0026] (5)Compared with the traditional "black box" scheduling strategy, it provides an evaluation framework with clear structure and logical closed-loop, making the 6G network operation process more verifiable, the scheduling results more interpretable, enhancing the trust of both service providers and users, and contributing to the construction of a transparent, fair, and reliable new network governance ecosystem.

[0027] (6)It provides core methodological support for the paradigm upgrade of the 6G network from "connection empowerment" to "service intelligent connection". Its innovation lies in upgrading the network resource configuration from "experience-driven" to "service-defined", realizing the cross-domain elastic scheduling of spectrum-environment-computing power resources, and providing a quantifiable, verifiable, and scalable methodological foundation for the 6G communication network. Description of the Drawings

[0028] Figure 1 Shown is a flowchart of the method for comprehensively evaluating the quality of wireless resource service of the present invention in an embodiment.

[0029] Figure 2 Shown is a schematic diagram of the evaluation dimensions and metrics of the quality of wireless resource service of the present invention in an embodiment.

[0030] Figure 3 Shown is a schematic structural diagram of the comprehensive evaluation system for the quality of wireless resource service of the present invention in an embodiment.

[0031] Figure 4 Shown is a schematic structural diagram of the evaluation terminal of the present invention in an embodiment.

[0032] Figure 5 Shown is a schematic structural diagram of the comprehensive evaluation system for the quality of wireless resource service of the present invention in another embodiment. Detailed Embodiments

[0033] The following uses specific specific examples to illustrate the embodiments of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, without conflict, the following embodiments and the features in the embodiments can be combined with each other.

[0034] It should be noted that the diagrams provided in the following embodiments only schematically illustrate the basic concept of the present invention. Therefore, only the components related to the present invention are shown in the diagrams, rather than being drawn according to the number, shape, and size of the components in actual implementation. The type, quantity, and ratio of each component in actual implementation can be arbitrarily changed, and the component layout type may also be more complex.

[0035] The full-dimensional evaluation method, system, storage medium and terminal for the quality of wireless resource service of the present invention are aimed at intelligent scheduling of wireless resources and quality of service guarantee in multi-service scenarios such as enhanced mobile broadband (eMBB), massive machine-type communication (mMTC), and ultra-reliable low-latency communication (uRLLC) in a wireless network, and aim to provide a unified and scalable solution for multi-dimensional quality of service evaluation in a complex and dynamic wireless environment. By integrating multiple core dimensions including but not limited to performance, security, overhead, autonomy, etc., the present invention maps them to the technical implementation layer, such as the quality of service (QoS) of spectrum resources, environmental information QoS, computing power resource QoS, etc., and finally establishes a unified evaluation framework covering the entire life cycle of wireless resources, thereby breaking through the contradiction between the static evaluation paradigm and the dynamic technology evolution, providing a quality of service guarantee benchmark with both accuracy and adaptability for the wireless network, and realizing panoramic perception and dynamic optimization of wireless resources. It should be noted that the full-dimensional evaluation method, system, storage medium and terminal for the quality of wireless resource service of the present invention are particularly applicable to the 6G network.

[0036] Next, the technical solutions in the embodiments of the present invention will be described in detail with reference to the accompanying drawings in the embodiments of the present invention.

[0037] As Figure 1 shown, in one embodiment, the full-dimensional evaluation method for the quality of wireless resource service of the present invention includes steps S1 to S3.

[0038] Step S1, obtain multi-dimensional network service quality perception information provided by a user terminal for spectrum resources, environment, and computing power resources.

[0039] Specifically, in the present invention, based on the multi-source heterogeneous access mechanism of the perception module, through real-time data collection and collaborative feedback of global nodes such as ground base stations, satellite links, and unmanned platforms, global monitoring of core parameters such as the dynamic distribution of spectrum resources, environmental interference fluctuations, and computing power load status is realized, multi-dimensional network service quality perception information (multi-dimensional network QoRS perception information) is obtained, and it is uploaded through a 6G access network gateway node.

[0040] In terms of spectrum resource perception, by identifying the cyclostationary characteristics of wireless signals, cyclic autocorrelation operations are performed on user signals to extract parameters such as the spectrum utilization rate, idle frequency band distribution, channel occupancy, and interference characteristics in the area where the user is located, reduce the influence of negative effects such as false alarm fluctuations, detection probability distortion, and user misidentification, and realize non-intrusive perception and classification identification of the spectrum state.

[0041] In terms of environmental perception, multi-source environmental information collected by user terminals and edge nodes is fused, and combined with indicators such as dynamic channel quality, occlusion change trend, and user movement trajectory to construct an environmental state evolution model, which is used to assist in identifying the risk of channel degradation and potential service interruption in complex scenarios.

[0042] In terms of computing power resource perception, the status of computing resource nodes such as edge nodes and access base stations associated with the user access path is perceived, covering key indicators such as the utilization, operating load, and response capabilities of heterogeneous computing power resources such as CPUs, GPUs, DPUs, and NPUs. It can achieve standardized classification of computing power nodes and quantitative evaluation of resource capabilities. It can also identify the service priority and required resource capabilities through computing power demand instructions, realizing dynamic adaptation and efficient matching of services and computing power.

[0043] In one embodiment, obtaining the multi-dimensional network service quality perception information provided by the user terminal includes the following steps.

[0044] 11) Normalize the spectrum resource perception information, environmental perception information, and computing power resource perception information to eliminate the dimensional differences between different indicators, and obtain the normalized perception information.

[0045] 12) Perform spatio-temporal alignment on the normalized perception information to obtain the aligned perception information.

[0046] Among them, combining the user location movement trajectory and network topology evolution information, synchronously align the normalized perception information in the time and space dimensions to ensure the temporal consistency and context relevance of the input information.

[0047] 13) Perform semantic encoding on the aligned perception information to obtain perception features.

[0048] Among them, use an embedding encoder to perform semantic annotation and mapping on the heterogeneous information in the aligned perception information, and convert discrete information such as behavior preferences, service types, and node attributes into computable vector forms; for continuous indicators, use a sliding window aggregation strategy for feature extraction to retain trend features and short-term fluctuation patterns, thereby obtaining perception features.

[0049] 14) Compress the perception features into a multi-dimensional network service quality perception vector as the multi-dimensional network service quality perception information.

[0050] Among them, all perception features are mapped through vector compression into a multi-dimensional network QoRS perception vector in a unified format. This vector constitutes a comprehensive multi-dimensional network service quality perception view of the user resource status, behavior characteristics, and environmental conditions, providing an interpretable and quantifiable input basis for QoRS evaluation, and significantly improving the sensitivity and response ability of the evaluation system to service quality assurance.

[0051] Step S2: Based on the multi-dimensional network service quality perception information, obtain the evaluation result of the wireless resource service quality according to the multi-task learning neural network, so that the scheduling module divides the priorities of the wireless resources and performs dynamic rescheduling based on the evaluation result.

[0052] Specifically, in the evaluation module of the present invention, input the multi-dimensional network service quality perception information into the trained multi-task learning neural network to obtain the evaluation result of the wireless resource service quality.

[0053] The evaluation result includes the total score of the wireless resource service quality and the dimension score values of different evaluation dimensions. As Figure 2 shown, in one embodiment, the evaluation dimensions include the performance, security, overhead, and autonomy of the wireless resources. The performance includes digital performance sub-items, algorithm performance sub-items, computing power performance sub-items, and connection performance sub-items. The digital performance sub-items include data quality and accuracy guarantee and data update and timeliness management. The algorithm performance sub-items include the selection and application of efficient algorithms and algorithm optimization and performance improvement. The computing power performance sub-items include the reasonable allocation of computing power resources and the technical means for computing power improvement. The connection performance sub-items include the technical implementation of stable connection and the optimization strategy of connection speed. The security includes data security sub-items and algorithm security sub-items. The data security sub-items include data encryption and privacy protection and data backup and recovery mechanisms. The algorithm security sub-items include the detection and prevention of algorithm vulnerabilities and the evaluation system of algorithm security. The overhead includes storage overhead sub-items, computing overhead sub-items, transmission overhead sub-items, and energy consumption overhead sub-items. The storage overhead sub-items include the selection and configuration of storage devices and the control method of storage costs. The computing overhead sub-items include the efficient utilization of computing resources and the optimization strategy of computing costs. The transmission overhead sub-items include the optimized selection of transmission protocols and the measures for reducing transmission costs. The energy consumption overhead sub-items include the application practice of energy-saving technologies and the management plan of energy consumption costs. The autonomy includes data autonomy sub-items and algorithm autonomy sub-items. The data autonomy sub-items include the mode of data autonomous management and the implementation path of digital autonomy. The algorithm autonomy sub-items include the mechanism of algorithm self-optimization and the evaluation index of algorithm autonomy.

[0054] In one embodiment, based on the multi-dimensional network service quality perception information, obtaining the evaluation result of the wireless resource service quality according to the multi-task learning neural network includes the following steps.

[0055] 21) Input the multi-dimensional network service quality perception information into the trained multi-task learning (Multi-Task Learning, MTL) neural network.

[0056] Among them, the multi-task learning neural network takes multi-dimensional network QoRS perception information as input, constructs feature prediction sub-channels corresponding to each dimension respectively, and introduces an adaptive attention mechanism in the shared layer to model and dynamically adjust the weights of the implicit correlation relationships between the features of each dimension; realizes the alignment of heterogeneous perception information through feature normalization and the embedding layer, and combines with the scenario weight factor to enhance the adaptation ability to typical business scenarios (such as eMBB, uRLLC, mMTC) while maintaining the generalization ability of the metrics, so as to realize the evaluation of the quality of wireless resource services.

[0057] When training the multi-task learning neural network, using historical multi-dimensional network QoS perception data, network behavior feedback, and user satisfaction evaluation as annotation signals, constructing a multi-objective loss function and jointly optimizing it, so that the multi-task learning neural network has the ability to highly sensitively identify abnormal states such as performance mutations, channel degradations, and computing congestions. The output of the multi-task learning neural network includes the total score of the quality of wireless resource services and the dimension score values of the evaluation dimensions, and performs risk annotation on the evaluation dimensions that significantly deviate from the target values, providing accurate and interpretable decision-making basis for subsequent resource scheduling.

[0058] 22) Obtain the total score of the quality of wireless resource services output by the multi-task learning neural network , where N represents the total number of sub-items of each evaluation dimension, x i represents the multi-dimensional network service quality perception information corresponding to the i th sub-item of the evaluation dimension, f i (·) represents the sub-score function corresponding to the i th sub-item of the evaluation dimension, α i represents the weight coefficient corresponding to the i th sub-item of the evaluation dimension and ∑ α i = 1.

[0059] Among them, the sub-score function f i (·) can be implemented by a neural network sub-model or a regressor, α i and is dynamically adjusted according to the task scenario, user preferences, or historical feedback.

[0060] 23) Obtain the dimension score value of any evaluation dimension , where M represents the total number of sub-items corresponding to the evaluation dimension, x j represents the multi-dimensional network service quality perception information corresponding to the j th sub-item of the evaluation dimension,f j (·) represents the sub - scoring function corresponding to the j sub - item of the said evaluation dimension, α j represents the weight coefficient corresponding to the j sub - item of the said evaluation dimension.

[0061] Among them, the total score of the radio resource service quality and the dimension score value of the said evaluation dimension are represented in a standardized dimension, which is convenient for unified comparison of cross - domain indicators.

[0062] 24) Use the total score of the radio resource service quality and the dimension score value of the said evaluation dimension as the said evaluation result.

[0063] It should be noted that the present invention also includes, based on the said evaluation module, according to the total score of the radio resource service quality and the dimension score value of the said dimension, outputting a score confidence interval and a risk label, which identify the deviation degree of the current service quality state of the radio resource relative to the set target value and the urgency of possible triggering of a scheduling response. For example, if the autonomous dimension score is significantly lower than the set threshold, a resource rescheduling recommendation signal will be automatically triggered to support the priority judgment of the scheduling control layer.

[0064] In one embodiment, the said evaluation result is transmitted to the scheduling module in the form of a structured QoRS evaluation report, and the content includes but is not limited to: the current total score of the radio resource service quality, the dimension score values of each evaluation dimension, the index deviation trend, the potential performance bottleneck, and the service level recommendation suitable for the service type, so as to provide a clear guidance for subsequent resource reconstruction and service guarantee.

[0065] The said scheduling module performs priority division and dynamic rescheduling on the radio resources based on the said evaluation result. First, construct a priority vector P = p 1 , p 2 , …, p n according to the dimension score values of each evaluation dimension. Among them n represents the number of users, and the radio resource scheduling priority of the i th user , QoRS (i) is the total score of the radio resource service quality of the i th user, Δ i represents the deviation degree between the optimal total score of the current QoRS and the optimal total score of the historical optimal QoRS. The larger Δ i is, the more obvious the degradation of the current resource state compared with the historical optimal state; when Δ i tends to 0, it indicates that the current service is close to or reaches the optimal.β ∈[0,1] is an adjustable weight coefficient, which is a reference weight for balancing the current performance and the historical evolution trend. Among them , QoRS (i) best is the optimal total score of the wireless resource service quality reached by the i -th user within the historical window, ε is a small positive number introduced to prevent the denominator from being zero. Then, all users are sorted using the above priority vector, and combined with the prediction of resource availability within the sliding window, a feasible scheduling solution set is constructed. For example, a dynamic scheduling algorithm based on weighted maximum matching and integer linear programming constraint optimization can be adopted. To fully reflect the guiding role of priority in the resource allocation process, priority is introduced as a scheduling weighting factor to participate in the construction of the dynamic scheduling optimization model. The dynamic scheduling result is obtained according to acquisition, where the resource allocation matrix A = a ij , a ij ∈ {0,1} indicates whether to allocate the resource j to the user i , w ij The weight of the resource j allocated to the user i . The constraint conditions ensure that each user can only occupy limited resources and each resource is not reallocated.

[0066] Step S3, obtain the scheduling execution result of the dynamic rescheduling, and correct the multi-task learning neural network based on the scheduling execution result and the multi-dimensional network service quality perception information.

[0067] Specifically, the dynamic rescheduling is executed in the edge control module, which specifically includes the following operations.

[0068] A) Spectrum migration: Switch the user to a band with less interference or higher idle degree through central scheduling or terminal coordination.

[0069] B) Path reconstruction: In the scenario of multi-hop transmission or link instability, update the forwarding path based on the shortest delay or highest stability index.

[0070] C) Bandwidth reallocation: Dynamically adjust the bandwidth allocation strategy according to the service level, and give priority to ensuring the key service segments of users with low QoRS scores.

[0071] The edge control module obtains the scheduling execution result, that is, the multi-dimensional network service quality perception information after scheduling, by executing the above dynamic rescheduling. Calculate the deviation index between the scheduling execution result and the multi-dimensional network service quality perception information , where QoS post (i) Indicates i The actual measurement value of multi-dimensional network service quality perception information, QoS pred (i) Indicates that before this round of scheduling i The predicted value of multi-dimensional network service quality perception information, δ i Indicates i The deviation index between the multi-dimensional network service quality perception information and the corresponding scheduling execution results.

[0072] When the deviation index is greater than the preset threshold θ (Right now δ i > θ ), the evaluation module triggers the closed-loop correction process. Specifically, the evaluation module retrains or fine-tunes the multi-task learning neural network, adjusts the weight coefficient of the wireless resource priority, and strengthens the robust strategy. Among them, the multi-task learning neural network is retrained or fine-tuned based on the scheduling execution result to improve the accuracy of the next round of prediction. Re-correct the resource priority weight coefficient according to the deviation distribution β , optimize the scheduling logic. For high-frequency deviation indicators, introduce elastic resource configuration, fault redundancy protection or link compensation mechanism.

[0073] Therefore, the present invention significantly enhances the adaptability to complex environmental disturbances through the closed-loop self-evolution optimization process among the evaluation of the multi-task learning neural network, the resource reallocation strategy and the evaluation feedback path, and provides a strong guarantee for service continuity and user experience in wireless business scenarios such as 6G.

[0074] It should be noted that the multi-task learning neural network can adaptively adjust the policy parameters to adapt to different business sensitivity requirements. Among them, based on the identified business scenarios and their QoRS demand sensitivity, the scenario weight factors will be dynamically adjusted, and different priority weights will be assigned to the key indicators of the evaluation dimension. For example, in telemedicine scenarios, the weights of factors related to delay jitter are significantly improved; in industrial control, the weights of factors related to the control instruction issuance rate will increase. The weight adjustment is iteratively optimized through the feedback learning mechanism to ensure that the strategy has the ability to respond quickly to business changes. By continuously testing and evaluating the scoring results, the parameter combination is automatically adjusted to achieve the best overall performance. At the same time, the resource reallocation strategy and evaluation feedback form a closed loop, collect scheduling execution results and subsequent QoRS changes in real time, and correct the multi-task learning neural network predictions and policy decisions through error feedback to ensure that it can continuously adapt to environmental fluctuations and changes in business needs, and realize the dynamic evolution of strategies.

[0075] The protection scope of the method for comprehensively evaluating the quality of wireless resource service described in the embodiments of the present invention is not limited to the execution order of the steps listed in this embodiment. Any solution achieved by adding or subtracting steps of the prior art or replacing steps according to the principles of the present invention is included in the protection scope of the present invention.

[0076] The embodiments of the present invention further provide a system for comprehensively evaluating the quality of wireless resource service. The system for comprehensively evaluating the quality of wireless resource service can implement the method for comprehensively evaluating the quality of wireless resource service described in the present invention. However, the implementation devices of the system for comprehensively evaluating the quality of wireless resource service described in the present invention include, but are not limited to, the structure of the system for comprehensively evaluating the quality of wireless resource service listed in this embodiment. Any structural deformation and replacement of the prior art made according to the principles of the present invention are included in the protection scope of the present invention.

[0077] As Figure 3 shown, in one embodiment, the system for comprehensively evaluating the quality of wireless resource service of the present invention includes an acquisition module 31, an evaluation module 32, and a feedback module 33.

[0078] The acquisition module 31 is used to acquire multi-dimensional network service quality perception information provided by a user terminal for spectrum resources, environment, and computing power resources.

[0079] The evaluation module 32 is connected to the acquisition module 31 and is used to obtain an evaluation result of the quality of wireless resource service based on the multi-dimensional network service quality perception information according to a multi-task learning neural network, so that a scheduling module can perform priority division and dynamic rescheduling on wireless resources based on the evaluation result.

[0080] The feedback module 33 is connected to the evaluation module 32 and is used to obtain the scheduling execution result of the dynamic rescheduling and correct the multi-task learning neural network based on the scheduling execution result and the multi-dimensional network service quality perception information.

[0081] Among them, the structures and principles of the acquisition module 31, the evaluation module 32, and the feedback module 33 correspond one by one to the above method for comprehensively evaluating the quality of wireless resource service, so they will not be elaborated here.

[0082] In several embodiments provided by the present invention, it should be understood that the disclosed system, apparatus, or method can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For example, the division of modules / units is only a logical functional division. In actual implementation, there may be other division methods. For example, multiple modules or units can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other can be through some interfaces. The indirect coupling or communication connection of devices or modules or units can be in electrical, mechanical, or other forms.

[0083] The modules / units described as separate components may or may not be physically separated. The components shown as modules / units may or may not be physical modules, that is, they can be located in one place or distributed to multiple network units. Some or all of the modules / units can be selected according to actual needs to achieve the purpose of the embodiments of the present invention. For example, in each embodiment of the present invention, the functional modules / units can be integrated in a processing module, or each module / unit can exist physically alone, or two or more modules / units can be integrated in one module / unit.

[0084] Those of ordinary skill in the art should further realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described according to functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0085] Embodiments of the present invention also provide a computer-readable storage medium. Those of ordinary skill in the art can understand that all or part of the steps in the above-mentioned method for comprehensively evaluating the quality of wireless resource services can be completed by instructing a processor through a program. The program can be stored in a computer-readable storage medium, and the storage medium is a non-transitory medium, such as random access memory, read-only memory, flash memory, hard disk, solid-state drive, magnetic tape, floppy disk, optical disc, and any combination thereof. The above storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or a data center integrating one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., digital video disc (DVD)), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0086] Embodiments of the present invention also provide a terminal. The terminal includes a processor and a memory.

[0087] The memory is used to store a computer program.

[0088] The memory includes various media that can store program codes, such as ROM, RAM, magnetic disk, USB flash drive, memory card, or optical disc.

[0089] The processor is connected to the memory and is configured to execute the computer program stored in the memory, so that the terminal executes the above-mentioned method for comprehensively evaluating the quality of wireless resource services.

[0090] Preferably, the processor can be a general-purpose processor, including a central processing unit (CPU for short), a network processor (NP for short), etc.; it can also be a digital signal processor (DSP for short), an application specific integrated circuit (ASIC for short), a field programmable gate array (FPGA for short), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0091] Such as Figure 4As shown, the evaluation terminal of the present invention is presented in the form of a general computing device. The components of the evaluation terminal may include, but are not limited to: one or more processors or processing units 41, a memory 42, and a bus 43 connecting different system components (including the memory 42 and the processing unit 41).

[0092] The bus 43 represents one or more of several types of bus architectures, including a memory bus or a memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of the various bus architectures. For example, these architectures include, but are not limited to, Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MAC) bus, Enhanced ISA bus, Video Electronics Standards Association (VESA) local bus, and Peripheral Component Interconnect (PCI) bus.

[0093] The evaluation terminal typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the evaluation terminal, including volatile and non-volatile media, removable and non-removable media.

[0094] The memory 42 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 421 and / or cache memory 422. The evaluation terminal may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, a storage system 423 can be used for reading and writing on non-removable, non-volatile magnetic media ( Figure 4 not shown, commonly referred to as a "hard disk drive"). Although Figure 4 not shown in the figure, a disk drive for reading and writing on a removable non-volatile disk (such as a "floppy disk"), and an optical disk drive for reading and writing on a removable non-volatile optical disk (such as a CD-ROM, DVD-ROM, or other optical media) can be provided. In these cases, each drive can be connected to the bus 43 through one or more data media interfaces. The memory 42 may include at least one program product having a set (e.g., at least one) of program modules that are configured to execute the functions of the various embodiments of the present invention.

[0095] A program / utility 424 having a set (at least one) of program modules 4241 can be stored, for example, in the memory 42. Such program modules 4241 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. The implementation of a network environment may be included in each or some combination of these examples. The program modules 4241 generally execute the functions and / or methods in the embodiments described in the present invention.

[0096] The evaluation terminal can also communicate with one or more external devices (such as a keyboard, a pointing device, a display, etc.), and can also communicate with one or more devices that enable a user to interact with the evaluation terminal, and / or communicate with any device (such as a network card, a modem, etc.) that enables the evaluation terminal to communicate with one or more other computing devices. Such communication can be carried out through the input / output (I / O) interface 44. Moreover, the evaluation terminal can also communicate with one or more networks (such as a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through the network adapter 45. As Figure 4 shown, the network adapter 45 communicates with other modules of the evaluation terminal through the bus 43. It should be understood that although not shown in the figure, other hardware and / or software modules can be used in combination with the evaluation terminal, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, etc.

[0097] As Figure 5 shown, in one embodiment, the wireless resource service quality full-dimensional evaluation system of the present invention includes the above-mentioned evaluation terminal 51, sensing module 52, scheduling module 53, and edge control module 54.

[0098] The sensing module 52 is connected to the evaluation terminal 51, and is used to collect multi-dimensional network service quality perception information provided by the user evaluation terminal for spectrum resources, environment, and computing power resources, and provide it to the evaluation terminal 51.

[0099] The scheduling module 53 is connected to the evaluation terminal 51, and is used to perform priority division and dynamic rescheduling on wireless resources based on the evaluation results provided by the evaluation terminal 51.

[0100] The edge control module 54 is connected to the evaluation terminal 51 and the scheduling module 53, and is used to generate the scheduling execution result of the dynamic rescheduling and provide it to the evaluation terminal.

[0101] Among them, the evaluation terminal 51, the sensing module 52, the scheduling module 53, and the edge control module 54 support distributed deployment and have the ability of cross-domain resource collaboration. The entire architecture has good module decoupling and scalability, and can adapt to the heterogeneity, diversity, and intelligent characteristics of future 6G networks, and is an important support platform for future network resource optimization management.

[0102] The above embodiments are only illustrative of the principles and effects of the present invention and are not intended to limit the present invention. Any person familiar with this technology can modify or change the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or changes made by those with ordinary knowledge in the technical field without departing from the spirit and technical ideas disclosed by the present invention should still be covered by the claims of the present invention.

Claims

1. A method for comprehensively evaluating the quality of service of radio resources, characterized in that: The method includes the following steps: Obtain the multi-dimensional network service quality perception information provided by the user terminal for spectrum resources, environment, and computing power resources; Based on the multi-dimensional network service quality perception information, obtain the evaluation result of the wireless resource service quality according to the multi-task learning neural network, so that the scheduling module performs priority division and dynamic rescheduling on the wireless resources based on the evaluation result; Obtain the scheduling execution result of the dynamic rescheduling, and correct the multi-task learning neural network based on the scheduling execution result and the multi-dimensional network service quality perception information.

2. The method for comprehensively evaluating the quality of wireless resource service according to claim 1, wherein: Obtaining the multi-dimensional network service quality perception information provided by the user terminal includes the following steps: Perform normalization processing on the spectrum resource perception information, environment perception information, and computing power resource perception information to obtain normalized perception information; Perform spatio-temporal alignment on the normalized perception information to obtain aligned perception information; Perform semantic encoding on the aligned perception information to obtain perception features; Perform vector compression on the perception features to obtain a multi-dimensional network service quality perception vector as the multi-dimensional network service quality perception information.

3. The method for comprehensively evaluating the quality of wireless resource service according to claim 1, wherein: The evaluation result includes the total score of the wireless resource service quality and the dimension score values of different evaluation dimensions; the evaluation dimensions include the performance, security, overhead, and autonomy of the wireless resources; The performance includes digital performance sub-items, algorithm performance sub-items, computing power performance sub-items, and connection performance sub-items; the security includes data security sub-items and algorithm security sub-items; The overhead includes storage overhead sub-items, computing overhead sub-items, transmission overhead sub-items, and energy consumption overhead sub-items; the autonomy includes data autonomy sub-items and algorithm autonomy sub-items.

4. The method for comprehensively evaluating the quality of wireless resource service according to claim 3, characterized in that: Based on the multi-dimensional network service quality perception information, obtaining the evaluation result of the wireless resource service quality according to the multi-task learning neural network includes the following steps: Input the multi-dimensional network service quality perception information into the trained multi-task learning neural network; Obtain the total QoS score of the wireless resource service output by the multi-task learning neural network , where N represents the total number of sub-items of each evaluation dimension, x i represents the multi-dimensional network QoS perception information corresponding to the sub-items of the i th evaluation dimension, f i (·) represents the sub-scoring function corresponding to the sub-items of the i th evaluation dimension, α i represents the weight coefficient corresponding to the sub-items of the i th evaluation dimension and ∑ α i = 1; Obtain the dimension score value of any evaluation dimension , where M represents the total number of sub-items corresponding to the evaluation dimension, x j represents the multi-dimensional network service quality perception information corresponding to the j th sub-item of the evaluation dimension, f j (·) represents the sub-scoring function corresponding to the j th sub-item of the evaluation dimension, α j represents the weight coefficient corresponding to the j th sub-item of the evaluation dimension; Use the total score of the wireless resource service quality and the dimension score values of the evaluation dimensions as the evaluation result.

5. The method for comprehensively evaluating the quality of wireless resource service according to claim 1, characterized in that: The dynamic scheduling module performing priority division and dynamic rescheduling on the wireless resources based on the evaluation result includes the following steps: Obtain the i wireless resource scheduling priority of the th user, where QoRS (i) is the total QoS score of the wireless resources of the i th user, β ∈[0,1] is an adjustable weight coefficient; , QoRS (i) best is the i th user's optimal total QoS score of the wireless resources within the historical window, ε is a small positive number introduced to prevent the denominator from being zero; Based on the wireless resource scheduling priority p i and the schedulable resource set, the system obtains the dynamic scheduling result , where A = a ij , a ij ∈ {0, 1} indicates whether to allocate the resource j to the user i , w ij Resource j allocated to the user i 's weight.

6. The method for comprehensively evaluating the quality of wireless resource service according to claim 1, wherein: Correcting the multi-task learning neural network based on the scheduling execution result and the multi-dimensional network service quality perception information includes the following steps: Calculate the deviation index between the scheduling execution result and the multi-dimensional network service quality perception information , where QoS post (i) represents the actual measured value of the i -th multi-dimensional network service quality perception information, and QoS pred (i) represents the predicted value of the i -th multi-dimensional network service quality perception information before this round of scheduling, δ i represents the deviation index between the i -th multi-dimensional network service quality perception information and the corresponding scheduling execution result; When the deviation index is greater than the preset threshold, retrain or fine-tune the multi-task learning neural network and adjust the weight coefficient of the wireless resource priority.

7. A full-dimensional evaluation system for the quality of wireless resource services, characterized in that: The system includes an acquisition module, an evaluation module, and a feedback module; The acquisition module is used to obtain the multi-dimensional network service quality perception information provided by the user terminal for spectrum resources, environment, and computing power resources; The evaluation module is used to obtain the evaluation result of the wireless resource service quality according to the multi-task learning neural network based on the multi-dimensional network service quality perception information, so that the scheduling module performs priority division and dynamic rescheduling on the wireless resources based on the evaluation result; The feedback module is used to obtain the scheduling execution result of the dynamic rescheduling, and correct the multi-task learning neural network based on the scheduling execution result and the multi-dimensional network service quality perception information.

8. A storage medium, on which a computer program is stored, characterized in that, When executed by a processor, the program implements the method for comprehensive dimension evaluation of wireless resource service quality described in any one of claims 1 to 6.

9. An evaluation terminal, characterized in that, It includes: a processor and a memory; the memory is used to store a computer program; the processor is used to execute the computer program stored in the memory, so that the evaluation terminal executes the method for comprehensive dimension evaluation of wireless resource service quality described in any one of claims 1 to 6.

10. A full-dimensional evaluation system for the quality of wireless resource services, characterized in that: It includes the evaluation terminal, a perception module, a scheduling module, and an edge control module described in claim 9; the perception module is used to collect multi-dimensional network service quality perception information provided by the user evaluation terminal for spectrum resources, environment, and computing power resources, and provide it to the evaluation terminal; the scheduling module is used to perform priority division and dynamic rescheduling of wireless resources based on the evaluation results provided by the evaluation terminal; the edge control module is used to generate the scheduling execution result of the dynamic rescheduling and provide it to the evaluation terminal.

Images