A method for providing rendering resources of a meta-universe scene and related equipment

Abstract

This invention discloses a method and related equipment for providing rendering resources for a metaverse scene. When a user device moves to a new location in a virtual scene, the system first determines the set of distant rendering resources to be pre-fetched and the dynamic latency constraints based on the movement trajectory. It then checks whether the resources are cached on the user device's local machine and within its collaborative group. If not cached, the edge server, acting as the auctioneer, initiates a multi-dimensional auction request to one or more candidate service providers. The system receives bidding information from each candidate, including the provided latency, rendering quality, and price. Based on this information, the dynamic latency constraints, and a user experience quality value model, a social welfare score is calculated. A target service provider is selected based on the score, and the payment price is determined. Resources are then obtained, provided to the user device, and the actual effect is verified. The purpose of this invention is to establish a complementary collaboration mechanism between local, edge, and cloud environments, construct a multi-dimensional resource allocation mechanism, design an honesty incentive mechanism, achieve reliable collaborative rendering, and meet the needs of immersive virtual reality experiences.

Description

Technical Field

[0001] This invention belongs to the field of metaverse technology, specifically relating to a method for providing rendering resources for a metaverse scene and related equipment. Background Technology

[0002] According to recent market reports, the global metaverse market exceeded $105 billion in 2024 and is projected to surpass $936 billion by 2030. With the rise of the metaverse concept, the demand for immersive experiences in virtual reality applications is growing. To provide a satisfying user experience, systems must simultaneously achieve extremely low interaction latency and high-fidelity visual rendering quality.

[0003] Existing resource management and allocation technologies for metaverse collaborative rendering still have significant shortcomings in practical applications: 1) Efficiency issues with rendering modes: The workload of scene rendering in the metaverse is enormous, and a single rendering mode cannot simultaneously achieve low latency and high-fidelity rendering quality. Existing technologies struggle to establish efficient complementary collaboration mechanisms between local, edge, and cloud environments, failing to fully leverage the architectural advantages of each. 2) Conflicts and balance issues in multi-dimensional scene quality: Metaverse rendering requires finding a balance between price, latency, and rendering quality. However, existing allocation methods based on game theory or auctions often neglect one aspect while neglecting another, lacking a multi-dimensional resource allocation mechanism that can simultaneously coordinate these three conflicting dimensions. 3) Ignoring the selfishness and honesty incentives of edge nodes: In open edge network environments, edge servers are inherently rational and selfish. They may overstate costs, exaggerate service capabilities, or conceal their true status to maximize their own interests. Due to the lack of effective honesty incentive mechanisms, these methods cannot guarantee that service providers truthfully report their private information, causing resource allocation results to deviate from the goal of maximizing social welfare and failing to achieve reliable collaborative rendering. Summary of the Invention

[0004] To address the problems existing in the prior art, this invention provides a method and related equipment for providing rendering resources for a metaverse scene. Its purpose is to establish an efficient complementary collaboration mechanism between local, edge, and cloud environments, fully leveraging the advantages of these three architectures to solve rendering mode efficiency issues; to construct a multi-dimensional resource allocation mechanism that can simultaneously coordinate the three conflicting dimensions of price, latency, and rendering quality, resolving the conflict and balance issues of multi-dimensional scene quality; and to design an effective honesty incentive mechanism to ensure that service providers truthfully report private information, aligning resource allocation results towards maximizing social welfare, achieving reliable collaborative rendering, and thus meeting the demands of virtual reality applications for immersive experiences in terms of extremely low interactive latency and high-fidelity visual rendering quality.

[0005] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution: According to a first aspect of the present invention, a method for providing rendering resources for a metaverse scene is provided, applied to a system including a user device, an edge server, and a cloud server, the method comprising: In response to a user device moving to a new location in a virtual scene, a set of distant rendering resources to be prefetched is determined based on the user device's movement trajectory, and the dynamic latency constraint corresponding to each distant rendering resource in the set is determined. The system queries whether the distant rendering resource has been cached on the user equipment and within its corresponding collaboration group. The collaboration group consists of an edge server that provides access services to the user equipment and multiple user equipments within its coverage area. If not cached, the edge server acts as the auctioneer and initiates a multi-dimensional auction request for the vision rendering resource to one or more candidate service providers, including other edge servers and the cloud server. Receive bidding information submitted by each candidate service provider in response to the multidimensional auction request, wherein the bidding information includes at least the promised latency, the promised rendering quality, and the price quoted; Based on the promised latency, promised rendering quality, the dynamic latency constraints, and the preset user experience quality value model, calculate the social welfare score corresponding to each bidding information. The target service provider is selected from among the candidate service providers based on the social welfare score, and the service price to be paid to the target service provider is determined based on the preset payment rules. The system obtains the vision rendering resources from the target service provider and provides them to the user device, and verifies the actual latency and rendering quality provided.

[0006] In one possible implementation of the first aspect, determining a set of distant rendering resources to be prefetched based on the movement trajectory of the user equipment, and determining the dynamic latency constraint corresponding to each distant rendering resource in the set, includes: Determine multiple associated locations within a preset path distance range centered on the new location; The distant rendering resources corresponding to the multiple associated locations are determined as the set of distant rendering resources to be prefetched; Based on the path distance between the associated location of each distant rendering resource in the set and the new location, the dynamic latency constraint corresponding to each distant rendering resource is dynamically calculated.

[0007] In one possible implementation of the first aspect, the dynamic latency constraint corresponding to each distant rendering resource is dynamically calculated based on the path distance between the associated location of each distant rendering resource in the set and the new location, specifically as follows:

[0008] in, For dynamic delay constraints; The path distance between the associated location of each distant rendering resource and the new location; The average time required for a user to move a unit step in the virtual scene.

[0009] In one possible implementation of the first aspect, the user experience quality value model is used to quantify latency and rendering quality into user experience quality value, expressed as:

[0010] in, Value for user experience quality; and These are the user's preference weights for latency and rendering quality, respectively, and satisfy the following conditions: + =1; For time delay; For rendering quality; It is a non-increasing utility function with respect to time delay; This is a non-decreasing utility function with respect to rendering quality.

[0011] In one possible implementation of the first aspect, the calculation of the social welfare score corresponding to each bidding information based on the promised delivery latency, the promised rendering quality, the dynamic latency constraint, and a preset user experience quality value model includes: Determine whether the promised delivery delay for each bidding information meets the corresponding dynamic delay constraint; If satisfied, the promised latency and promised rendering quality are input into the user experience quality value model to calculate the corresponding user experience quality value. Subtracting the bid price from the user experience quality value yields the social welfare score corresponding to the bid information.

[0012] In one possible implementation of the first aspect, the selection of the target service provider from among the candidate service providers based on the social welfare score includes: The candidate service provider with the highest social welfare score is selected as the target service provider.

[0013] In one possible implementation of the first aspect, the preset payment rule is a VCG payment rule, and the determination of the service price to be paid to the target service provider based on the preset payment rule includes: The user experience quality value corresponding to the bidding information provided by the target service provider is denoted as the first value; Among all the social welfare scores corresponding to the bidding information of all other candidate service providers, the highest social welfare score is determined and recorded as the second score; Subtracting the second rating from the first value yields the service price payable to the target service provider.

[0014] In one possible implementation of the first aspect, the verification of the actual provided latency and rendering quality includes: After acquiring the distant rendering resources, the actual latency and the measured rendering quality are then measured. The actual latency is compared with the latency promised by the target service provider, and the measured rendering quality is compared with the rendering quality promised by the target service provider. If any of the promised standards are not met, the service price paid to the target service provider will be reduced according to the degree of breach.

[0015] According to a second aspect of the present invention, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for providing rendering resources for a metaverse scene.

[0016] According to a third aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing a computer program, which, when executed by a processor, implements the method for providing rendering resources for a metaverse scene.

[0017] According to a fourth aspect of the present invention, a computer program product is provided that, when executed by a processor, implements the method for providing rendering resources for a metaverse scene.

[0018] Compared with the prior art, the present invention has at least the following beneficial effects: This invention provides a method for providing rendering resources for a metaverse scene. By responding to user device movement, it dynamically determines the pre-fetching of distant rendering resources and differentiated latency constraints, prioritizing resource queries locally and within collaborative groups composed of edge servers and multiple users, thus constructing an efficient local-edge-cloud collaborative rendering mechanism. This approach allows for flexible pre-preparation of distant rendering resources and fully utilizes nearby caching, effectively alleviating latency pressure and bandwidth consumption in cloud-based remote rendering, thereby improving the overall resource utilization efficiency and rendering response speed of the system. When a cache miss occurs within the collaborative group, a dynamic resource acquisition mechanism based on multi-dimensional auction is introduced, incorporating conflicting dimensions of latency, rendering quality, and price into a unified decision-making framework. By calculating a social welfare score based on an experience quality value model and selecting a service provider accordingly, it can automatically find the optimal balance between service quality and cost while meeting users' personalized experience needs, overcoming the shortcomings of traditional methods in multi-dimensional indicator trade-offs. By determining the final payment price through preset payment rules and verifying the actual latency and quality after service delivery, a complete incentive and constraint mechanism is formed. This mechanism enables rational service providers to truthfully report their service capabilities and costs when bidding and to strive to ensure service quality when fulfilling contracts. This ensures the authenticity of the resource allocation process and the validity of the results, ultimately achieving reliable and efficient collaborative rendering. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the specific embodiments of the present invention, the drawings used in the description of the specific embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0020] Figure 1 A flowchart illustrating a method for providing rendering resources for a metaverse scene according to an embodiment of the present invention; Figure 2 A system block diagram illustrating a method for providing rendering resources for a metaverse scene according to an embodiment of the present invention; Figure 3 The above are simulation results of the average social welfare of the system under different numbers of users in this embodiment of the invention. Figure 4 The simulation results of the average SSIM value per frame under different numbers of users in this embodiment of the invention; Figure 5 The following are simulation results of end-to-end frame latency under different numbers of users in this embodiment of the invention; Figure 6 Simulation results of frame SSIM values ​​provided by the winner under different delay thresholds in embodiments of the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] This invention provides a method for providing rendering resources for a metaverse scene, applicable to a system including user devices, edge servers, and cloud servers. The aim is to solve the problems of rendering efficiency, multi-dimensional index balance, and node incentive honesty by coordinating user devices, edge servers, and cloud servers and introducing dynamic resource allocation based on a multi-dimensional auction and verification mechanism.

[0023] like Figure 2 As shown, the detailed task division for each layer of the system, which includes user devices, edge servers, and cloud servers, is explained below: At the cloud server layer, serving as the system's global data warehouse and computing power center, the cloud server leverages its powerful computing resources to perform offline pre-rendering tasks. Regarding pre-rendering and storage, the cloud server pre-renders the highest-quality background frames (background rendering resources) for all virtual scene grid points, which are the master frames. The master frames represent an ideal, distortion-free visual reference. The cloud server stores all pre-rendered master frames as a fallback data source in case of edge server cache misses, ensuring that the required content is always available.

[0024] At the edge server layer, multiple edge servers are deployed near base stations or access points. The system is structured to be geographically closer to user devices to provide low-latency services. Each edge server does not store all data; instead, it caches only frequently accessed background frames within its coverage area based on popularity. This reduces long-distance transmission latency and backbone network bandwidth consumption associated with repeatedly fetching data from cloud servers. Furthermore, each edge server and all user devices within its coverage area are grouped into a collaborative group. The edge server acts as the manager node of this collaborative group, managing the background frame resource pool within the group. When a user device requests a background frame that is available within the collaborative group, the background frame is directly transmitted to the user device. If the requested background frame is not available within the collaborative group, the primary edge server acts as a proxy, initiating resource requests to other nearby edge servers or cloud servers and transmitting the obtained background frame to the user device.

[0025] At the user device layer, mobile terminal devices such as VR headsets are included. The local user device is primarily responsible for handling tasks that are extremely sensitive to latency. It utilizes its GPU to render foreground interactive frames and foreground background frames in real time. The content of these frames typically changes in real time with user device operations and is not suitable for pre-rendering. Simultaneously, the user device also caches a small number of frequently used distant background frames. When a new distant background frame is needed, the user device first searches locally; if not found, it sends a request to its collaborating group. Finally, the local user device is responsible for fusing the received or locally cached distant background frames with the locally rendered foreground interactive frames and foreground background frames to synthesize the final VR panoramic image presented to the user.

[0026] Through the above architecture, the system offloads computationally intensive landscape rendering tasks from the user device to edge servers and cloud servers, while keeping latency-sensitive interactive tasks on the user device, thus achieving effective load balancing and end-to-end latency optimization.

[0027] like Figure 1 As shown, the method specifically includes the following steps: S1. In response to the user device moving to a new location in the virtual scene, based on the movement trajectory of the user device, determine the set of distant rendering resources to be prefetched, and determine the dynamic latency constraint corresponding to each distant rendering resource in the set.

[0028] In other words, the user device moves to a new grid position within the virtual scene of the metaverse. At this point, based on the user device's movement trajectory, the system determines a set of distant rendering resources (i.e., distant background frames) that need to be prefetched, and calculates a dynamic, position-dependent latency constraint for each distant rendering resource in the set. Simply put, this step is to prepare distant rendering resources in advance and set a reasonable delivery deadline.

[0029] S2. Query whether the far-view rendering resource has been cached in the user equipment local area and its corresponding collaboration group. The collaboration group consists of an edge server that provides access services to the user equipment and multiple user equipments within its coverage area.

[0030] In other words, the system first searches for the required landscape rendering resources in the user device's local cache. If not found, it searches within the user device's collaborative group. As mentioned above, the collaborative group consists of the edge server providing wireless access services to the user device, and all other user devices within the edge server's coverage area. The collaborative group shares the landscape rendering resources cached by all user devices within the group and on the edge server. If the search within the collaborative group yields a match, the resource is obtained directly.

[0031] S3. If not cached, the edge server, acting as the auctioneer, initiates a multi-dimensional auction request for the vision rendering resource to one or more candidate service providers, including other edge servers and the cloud server.

[0032] In other words, if a query reveals that the required rendering resources are not cached on the user's device or within the collaborating group (i.e., a cache miss), the rendering resource acquisition process is triggered. At this point, the primary edge server acts as a proxy for the user device (i.e., the auctioneer), broadcasting a multi-dimensional auction request to other potential candidate service providers in the network to compete for the rendering resources. Candidate service providers include other edge servers and cloud servers possessing the full dataset.

[0033] S4. Receive bidding information submitted by each candidate service provider in response to the multidimensional auction request, wherein the bidding information includes at least the promised latency, the promised rendering quality, and the price.

[0034] Specifically, upon receiving the auction request, each candidate service provider assesses its current load, bandwidth, and computing resources, and formulates a bidding strategy based on the marginal cost of providing specific rendering quality and latency services. Each provider submits a multi-dimensional bidding document to the auctioneer. Each bidding document contains at least three commitments: the achievable latency of the provided rendering resource (i.e., the transmission latency from the provider to the main edge server), the image quality of the provided resource (measured by a structural similarity index, representing the similarity to the original high-quality main frame pre-rendered on the cloud server), and the price quoted for completing the service.

[0035] S5. Based on the promised latency, promised rendering quality, dynamic latency constraints, and a preset user experience quality value model, calculate the social welfare score corresponding to each bidding information.

[0036] Specifically, the main edge server, acting as the auctioneer, begins its evaluation after receiving all bids. The core of this evaluation is calculating the social welfare score for each bid. Specifically, for each valid bid, the main edge server first verifies whether its promised latency meets the dynamic latency constraints set for the rendering resource in step S1. If it does, the promised latency and rendering quality parameters are input into a preset user experience quality value model to calculate the value the service can bring to the user. Then, this value is subtracted from the bid price, and the resulting net value is the social welfare score for that bid.

[0037] S6. Select a target service provider from among the candidate service providers based on the social welfare score, and determine the service price to be paid to the target service provider based on the preset payment rules.

[0038] Specifically, the main edge server selects the target service provider from all candidate service providers based on a calculated social welfare score. The selection rule is to maximize social welfare, meaning the provider with the highest social welfare score is chosen as the winner, i.e., the target service provider. After determining the winner, payment is not made directly according to their bid, but rather the final service price is calculated based on preset payment rules, thereby incentivizing bidders to offer honest prices.

[0039] S7. Obtain the vision rendering resources from the target service provider and provide them to the user device, and verify the actual latency and rendering quality provided.

[0040] In other words, the target service provider, according to the agreement, transmits the required distant rendering resources to the main edge server, which then provides them to the user device. The user device is responsible for combining the received distant background frames with locally rendered, latency-sensitive foreground interaction frames and foreground background frames to form the final panoramic image. After service delivery, the system verifies the actual service quality: measuring the actual transmission latency and calculating the image quality of the delivered resources. The measured values ​​are compared with the bid commitments. If the commitment standards are not met, the final payment price will be deducted according to preset rules, thereby binding the service provider to fulfill its commitments.

[0041] In one possible implementation, the step of determining a set of distant rendering resources to be prefetched based on the user device's movement trajectory, and determining the dynamic latency constraint corresponding to each distant rendering resource in the set, is specifically implemented as follows: First, determine multiple associated locations within a preset path distance range centered on the new location, and then determine the distant rendering resources corresponding to the multiple associated locations as the set of distant rendering resources to be prefetched.

[0042] Specifically, the system abstracts and models the metaverse virtual space as a two-dimensional grid map. When user equipment u Move to a new grid location At that time, the system does not only prefetch the resources needed for the next location. To improve robustness and cope with the uncertainty of the user's movement direction, the system will... Centered on a target, determine all distances to its shortest path within a preset range. LThe grid positions within the frame are called associated positions. All the distant rendering resources corresponding to these associated positions constitute the set of distant rendering resources that need to be prefetched this time. Through a multi-frame collaborative prefetching strategy, urgent real-time requests are transformed into relaxed background tasks, allowing the system to complete transmission before the user arrives. This effectively addresses the uncertainty of the user's movement direction, ensuring a smooth visual experience regardless of which direction the user moves.

[0043] Secondly, based on the path distance between the associated location of each distant rendering resource in the set and the new location, the dynamic latency constraint corresponding to each distant rendering resource is dynamically calculated.

[0044] Specifically, for each distant rendering resource in the distant rendering resource set, its corresponding dynamic latency constraint is not a fixed value, but is dynamically calculated based on the distance between its associated location and the user device's current location. The calculation formula is:

[0045] in, For dynamic delay constraints; The path distance between the associated location of each distant rendering resource and the new location, i.e., from The shortest path number of steps required to reach the grid location corresponding to the distant rendering resource; The average time required for a user to move one unit step in the virtual scene is the average time required for a user to move one unit step on the grid map.

[0046] The meaning of this calculation formula is: for distance from the user For the distant view rendering resources, the user can tolerate the system not reaching that location at most (i.e., The delivery of this resource must be completed within a time step. The closer the distance ( The smaller the distance, the stricter the latency requirement; the greater the distance, the more lenient the latency requirement. This mechanism provides a flexible time window for resource scheduling. By utilizing the user device's movement trajectory and grid distance, the single real-time deadline is transformed into a flexible time window that dynamically changes with distance, thereby increasing the flexibility of resource scheduling while ensuring user experience.

[0047] In one possible implementation, the user experience quality value model is used to quantify latency and rendering quality into user experience quality value, expressed as:

[0048] in, Value for user experience quality; and These are the user's preference weights for latency and rendering quality, respectively, and satisfy the following conditions: + =1, users can adjust these two weights according to their own application scenarios. For example, competitive games prioritize latency, while sightseeing applications prioritize image quality. For time delay; For rendering quality; It is a non-increasing utility function with respect to time delay; This is a non-decreasing utility function with respect to rendering quality.

[0049] In detail, the user experience quality in the metaverse primarily depends on end-to-end latency and visual quality (i.e., rendering quality). Therefore, a user experience quality value model is established to map objective network performance metrics (latency) and subjective visual quality to the same utility value, namely, the user experience quality value, as follows: Using structural similarity SSIM Metrics to measure the forward background frames delivered by edge servers Ideal main frame stored on cloud server The visual difference between them is denoted as In order to make objective The value is mapped to the user's subjective quality rating. This invention uses the Sigmoid function as a non-decreasing utility function for rendering quality. Specifically:

[0050] In the formula, Indicates the maximum quality score. The curve steepness coefficient controls the score as follows. SSIM Sensitivity to change for SSIM The midpoint value is used to calibrate the center position of the scoring curve. This simulates the non-linear perceptual characteristics of the human visual system to changes in image quality; that is, the perceptual changes are relatively gradual in the high-quality and low-quality ranges, while the changes are more drastic in the critical range.

[0051] End-to-end delay For the main edge server Initiate a request to the user equipment The total time to complete the final frame synthesis. It should be noted that the total latency consists of four parts, specifically:

[0052] In the formula, The latency of acquiring distant background frames for the main edge server. For the last hop wireless transmission latency, For decoding latency, This is the frame synthesis delay.

[0053] To ensure immersion, latency must meet the maximum tolerance limit. Define the delay relaxation time. This invention uses a non-decreasing concave function based on relaxation time to define a non-increasing utility function for time delay. ,Right now With relaxation time As the network size increases, user satisfaction will definitely rise, not fall. When the network is extremely congested and barely meets the deadline, if... Even a slight increase can significantly improve the user experience. When the network is already very fast, even with further speed improvements, users will barely perceive the difference, and the increase in satisfaction will be minimal. When the value is negative, it indicates a severe visual lag.

[0054] Combining the utility functions of the two dimensions mentioned above, a user device is constructed using a weighted summation method. Overall Experience Quality (QoE) Value Model ,Right now:

[0055] The model needs to satisfy two hard constraints simultaneously, namely and This is to ensure basic frame availability.

[0056] In one possible implementation, the social welfare score corresponding to each bidding information is calculated based on the promised delivery latency, the promised rendering quality, the dynamic latency constraint, and a preset user experience quality value model. The specific implementation process is as follows: Determine whether the promised delivery delay for each bidding information meets the corresponding dynamic delay constraint; If satisfied, the promised latency and promised rendering quality are input into the user experience quality value model to calculate the corresponding user experience quality value. Subtracting the bid price from the user experience quality value yields the social welfare score corresponding to the bid information.

[0057] More specifically, the main edge server received the bid. Then, the following steps are used to calculate its score to ensure that the social welfare of resource allocation is maximized: First, combining the known local last-hop wireless transmission, decoding, and synthesis latency, the bidding commitments will be... Convert to end-to-end total delay prediction . judge Does it meet the dynamic latency constraints corresponding to the distant rendering resource? If the conditions are not met, the bid will be deemed invalid.

[0058] Secondly, judge the promised rendering quality. Has the basic availability threshold been reached, for example? ≥ 0.9. If this is not met, the bid is also invalid.

[0059] Next, for valid bids that pass the above verification, the main edge server will assign its committed parameters ( , Substituting the values ​​into the user experience quality model, we can calculate the value of the service for a specific user. u The experiential value it brings.

[0060] Finally, from the calculated user experience value Subtract the bidder's requested price from the price. The social welfare score for this transaction is obtained as follows:

[0061] Due to the honest bidding strategy Equals the bidder's true marginal cost Therefore, the social welfare score directly reflects the net social welfare of the transaction, and comprehensively reflects the trade-off between improving user experience and incurring social costs (reflected in the bid price).

[0062] In one possible implementation, selecting the target service provider from among the candidate service providers based on the social welfare score includes: selecting the candidate service provider with the highest social welfare score as the target service provider.

[0063] In other words, after calculating the social welfare scores of all valid bids, the main edge server selects the target service provider based on the principle of maximizing social welfare. Specifically, the server with the highest social welfare score among all candidate service providers (including other edge servers and cloud servers) is chosen. The winner of the auction is the _____. This can be expressed as a formula:

[0064] It should be noted that if multiple bidders have the same social welfare score, their bids will be selected first. The lowest bidder will be chosen to further reduce costs. If the bids are the same, one of them can be selected randomly.

[0065] In one possible implementation, the preset payment rule is the VCG (Vickrey-Clarke-Groves) payment rule, and the service price to be paid to the target service provider is determined based on the preset payment rule, as follows: The user experience quality value corresponding to the bidding information provided by the target service provider is denoted as the first value; Among all the social welfare scores corresponding to the bidding information of all other candidate service providers, the highest social welfare score is determined and recorded as the second score; Subtracting the second rating from the first value yields the service price payable to the target service provider.

[0066] In other words, in determining the target service provider Afterwards, pay to The price is not its quoted price. Instead, it is calculated according to VCG payment rules to ensure the mechanism's authentic incentive characteristics. The main edge server pays the target service provider. amount Specifically:

[0067] In the formula, For target service providers The total value generated by the services provided to users. If not selected The system can obtain the second highest social welfare score, that is, the winner's payment equals the marginal contribution he / she makes to the system, thus ensuring the honesty of the system.

[0068] In one possible implementation, the verification of the actual latency and rendering quality is as follows: After acquiring the distant rendering resources, the actual latency and the measured rendering quality are then measured. The actual latency is compared with the latency promised by the target service provider, and the measured rendering quality is compared with the rendering quality promised by the target service provider. If any of the promised standards are not met, the service price paid to the target service provider will be reduced according to the degree of breach.

[0069] In other words, after the target service provider j* delivers the rendering resources, the main edge server performs latency verification and rendering quality verification. It records the actual time from initiating the request to fully receiving the resource data packet, combines this with the local processing time to obtain the actual end-to-end latency, and compares this actual end-to-end latency with the latency constraints promised in the bidding. It also compares the actually received frame data with the original high-quality main frames stored on the cloud server to calculate the actual structural similarity index, i.e., the actual rendering quality.

[0070] The determination and penalties for breach of contract are as follows: If the actual end-to-end latency exceeds the latency constraint promised in the bid, it is considered a latency breach.

[0071] If the actual rendering quality is lower than the promised rendering quality or the base threshold, it is considered a quality breach.

[0072] If a breach occurs, the system will calculate the penalty value based on the deviation magnitude according to the preset tiered penalty mechanism. and from the originally scheduled payment amount The amount will be deducted directly from the auction proceeds. If a node defaults multiple times consecutively, the main edge server will suspend that node's bidding eligibility in subsequent auction cycles.

[0073] Based on the above implementation methods, the auction triggering and role definition triggering conditions are explained in detail. When the user device Move to new grid position And request to pre-fetch distant background frames. At that time, the main edge server First, the caches of all nodes within the collaboration group are queried. If a cache miss occurs within the group, meaning the frame is neither in the local cache nor in the cache of a neighboring edge server, an auction mechanism is immediately triggered. (Primary Edge Server) As auctioneer and buyer's agent, representing the user equipment Initiating a request; other nearby edge servers in the network And cloud servers with full data As a bidding party. Main edge server. Broadcast a frame request message to all potential bidders. The request explicitly includes the required background frame. The identifier and value function parameters are provided so that bidders can estimate the value that their services can bring to users; it is explicitly stated that a scoring rule based on maximizing social welfare and a payment rule based on VCG will be adopted.

[0074] Each bidder receiving the auction request acts as a seller. Each bidder assesses the quality of service it can provide based on its current load, bandwidth availability, and transcoding capabilities, and calculates the corresponding marginal cost. Bidding parties Submit a multi-dimensional bid Specifically:

[0075] In the formula: Representative from the bidding party To the main edge server The acquisition latency. This is an objective metric that the bidder can control and commit to, and does not include the user's private processing time. The price requested by the bidder for providing the service is the quoted price. Based on the honesty-based nature of the mechanism, rational bidders will truthfully declare their actual costs. .

[0076] Cost calculation is based on a cost model consisting of the following three parts:

[0077] In the formula, For fixed access costs, defined as This refers to the fixed computing resources required for bidders to respond to requests, retrieve cached data, and initiate transcoding. Quality-related transmission costs are positively correlated with the promised rendering quality and are defined as follows: In order to fulfill the promise The frame data size must be minimized. This cost is calculated based on the amount of data transmitted and the cost per bit. The product of. The latency-related bandwidth cost is negatively correlated with the promised latency, i.e. In order to guarantee the promised low latency Bidders must reserve sufficient transmission bandwidth. This cost includes reserved bandwidth and opportunity cost per unit bandwidth. The product of.

[0078] Main edge server Upon receiving the bid, the latency parameters provided by the bidder must be converted into complete metrics from a user experience perspective. That is, the main edge server will process the latency parameters reported by the bidder. By adding it to the user-side private information it possesses, the end-to-end prediction total latency used for the final score can be reconstructed. .

[0079] This invention addresses the performance bottlenecks of existing single-computing models by constructing a three-layer collaborative architecture of local-edge-cloud, achieving hierarchical and fine-grained distribution of computing tasks. By allowing local user devices to handle latency-sensitive foreground interactions and close-up backgrounds, while the cloud handles high-performance full pre-rendering, and the edge caches hotspot background frames to alleviate backbone network bandwidth pressure, this architecture fully leverages the computing and storage advantages of each layer. It effectively overcomes the network congestion bottlenecks of traditional single-cloud rendering and the resource constraints of single-edge computing, maximizing overall system throughput and resource utilization efficiency. To address the difficulty of balancing conflicting multi-dimensional metrics in existing resource allocation, this invention establishes a comprehensive QoE value model that incorporates latency and rendering quality, and integrates price factors through a multi-dimensional auction mechanism. This mechanism allows the system to find the optimal social welfare balance between pursuing high-quality rendering, low-latency transmission, and low cost, overcoming the shortcomings of existing single-dimensional allocation strategies that focus solely on price or latency, thus ensuring an immersive user experience while achieving efficient resource utilization. To address the lack of honesty caused by the selfishness of edge nodes in open network environments, this invention introduces the Vickrey-Clarke-Groves payment rule. From a mechanism design perspective, this rule ensures that the optimal strategy for bidders is to truthfully report their actual costs and service capabilities. This effectively solves the problem that edge nodes may exaggerate costs or service capabilities to maximize their own interests, thus guaranteeing that resource allocation results truly maximize social welfare.

[0080] To verify the actual effect of the method described in this invention, this embodiment conducted simulation verification in a typical metaverse rendering scenario.

[0081] The simulation environment constructs a three-tier collaborative rendering system comprising one cloud server, ten edge servers, and multiple groups of mobile user devices. In each time slot, the number of user devices requesting distant background frames increases from the set... Caches are randomly selected from the available data and randomly distributed within the coverage area of ​​the edge servers. The cache capacity of each edge server is determined independently from... The local cache capacity of each user device is selected from the options provided. Uniform sampling is performed between the samples. The physical space is discretized into... A two-dimensional grid is constructed, with cloud servers, edge servers, and user devices randomly distributed within it. To ensure a fair comparison, three traditional schemes are selected as baselines: 1) Centralized Cloud Fetch (CCF), which always directly retrieves missing background frames from the cloud; 2) Proximity-Based Selection (Prox-S), which selects the physically closest edge node as the content provider, and falls back to the cloud if it is unavailable; 3) Single-Dimensional Auction (SDA), which selects a provider solely based on the lowest price, ignoring latency and quality dimensions.

[0082] like Figure 3 As the number of user devices increases, the mechanism proposed in this invention consistently maintains the highest social welfare. In contrast, the performance of all baseline methods declines with increasing load due to increased costs caused by congestion and cache misses. Specifically, the CCF scheme, being entirely cloud-dependent, experiences a sharp drop in social welfare with increasing load; while the SDA scheme, focusing solely on price and neglecting experience quality, results in the lowest social welfare. These results demonstrate that this invention, through a multi-dimensional auction mechanism, effectively balances user experience and cost, achieving optimal allocation of system resources.

[0083] like Figure 4 and Figure 5 This invention comprehensively demonstrates its performance advantages across two core dimensions of Quality of Image (QOE). Specifically, existing technologies typically excel in only one dimension, making it difficult to achieve a balance across multiple dimensions. While the CCF solution maintains a near-perfect SSIM (Signal Instance Quality) of 1.0 by directly acquiring the main frame from the cloud, its end-to-end latency increases dramatically with the number of users. This high latency severely disrupts the user's immersion and negates the advantages brought by image quality. The SDA solution, focusing solely on price and neglecting quality optimization, delivers the worst SSIM among all solutions, failing to meet the demands of high-quality rendering. Although the Prox-S solution utilizes edge resources, its lack of comprehensive consideration of congestion and quality leads to increased latency due to queuing as the load increases, resulting in a low and unstable SSIM. In contrast, the method of this invention achieves the optimal balance across two key metrics: in terms of visual quality, it consistently maintains a high and stable level, ensuring that the delivered image quality exceeds the user's perception threshold. In terms of latency performance, it consistently maintains the lowest end-to-end latency even under high load, significantly outperforming the baseline. These results demonstrate... This invention successfully resolves the conflict between latency and image quality by introducing a multi-dimensional auction mechanism, significantly reducing transmission latency while ensuring visual quality, thereby achieving a comprehensive QoE superior to all baseline methods.

[0084] like Figure 6 Showing the winning bidder SSIM The changing trend with maximum tolerable delay. When time budget... When the value is small, the concave utility function A steep latency penalty was imposed, forcing the mechanism to prioritize low-latency offers that might sacrifice some image quality. With As latency increases, the marginal utility diminishes, and the decision-making focus shifts to visual quality, leading to a preference for higher-quality visuals. SSIM The quotes provided demonstrate that the present invention can intelligently and dynamically adjust between latency and image quality based on the urgency of the task.

[0085] In another embodiment of the present invention, a computer device is provided, comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in the computer storage medium to achieve a corresponding method flow or corresponding function. The processor described in this embodiment of the present invention can be used in the operation of a method for providing rendering resources for a metaverse scene.

[0086] In another embodiment of the present invention, a storage medium is provided, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device used to store programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the computer device and extended storage media supported by the computer device. The computer-readable storage medium provides storage space that stores the terminal's operating system. Furthermore, the storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be Random Access Memory (RAM) or non-volatile memory, such as at least one disk storage device. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the corresponding steps of the method for providing rendering resources for a metaverse scene in the above embodiments.

[0087] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, optical storage, etc.) containing computer-usable program code.

[0088] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0089] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0090] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0091] This invention also provides a computer program product, which is used to execute any of the above-described methods for providing rendering resources for a metaverse scene. Since the computer program product provided by this invention and the above-described method for providing rendering resources for a metaverse scene belong to the same inventive concept, the computer program product provided by this invention has all the advantages of the above-described method for providing rendering resources for a metaverse scene. Therefore, the beneficial effects of the computer program product provided by this invention will not be elaborated further here.

[0092] In this invention, the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to a specific feature, structure, material, or characteristic described in connection with that embodiment or example, which is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0093] Finally, it should be noted that the above-described embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and not to limit them. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments within the scope of the technology disclosed in the present invention, or make equivalent substitutions for some of the technical features; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered within the scope of protection of the present invention.

Claims

1. A method for providing rendering resources for a metaverse scene, characterized in that, Applied to systems including user equipment, edge servers, and cloud servers, the method includes: In response to a user device moving to a new location in a virtual scene, a set of distant rendering resources to be prefetched is determined based on the user device's movement trajectory, and the dynamic latency constraint corresponding to each distant rendering resource in the set is determined. The system queries whether the distant rendering resource has been cached on the user equipment and within its corresponding collaboration group. The collaboration group consists of an edge server that provides access services to the user equipment and multiple user equipments within its coverage area. If not cached, the edge server acts as the auctioneer and initiates a multi-dimensional auction request for the vision rendering resource to one or more candidate service providers, including other edge servers and the cloud server. Receive bidding information submitted by each candidate service provider in response to the multidimensional auction request, wherein the bidding information includes at least the promised latency, the promised rendering quality, and the price quoted; Based on the promised latency, promised rendering quality, the dynamic latency constraints, and the preset user experience quality value model, calculate the social welfare score corresponding to each bidding information. The target service provider is selected from among the candidate service providers based on the social welfare score, and the service price to be paid to the target service provider is determined based on the preset payment rules. The system obtains the vision rendering resources from the target service provider and provides them to the user device, and verifies the actual latency and rendering quality provided.

2. The method for providing rendering resources for a metaverse scene according to claim 1, characterized in that, The step of determining a set of distant rendering resources to be prefetched based on the user equipment's movement trajectory, and determining the dynamic latency constraint corresponding to each distant rendering resource in the set, includes: Determine multiple associated locations within a preset path distance range centered on the new location; The distant rendering resources corresponding to the multiple associated locations are determined as the set of distant rendering resources to be prefetched; Based on the path distance between the associated location of each distant rendering resource in the set and the new location, the dynamic latency constraint corresponding to each distant rendering resource is dynamically calculated.

3. The method for providing rendering resources for a metaverse scene according to claim 2, characterized in that, The dynamic latency constraint for each distant rendering resource is dynamically calculated based on the path distance between the associated location of each distant rendering resource in the set and the new location. Specifically: in, For dynamic delay constraints; The path distance between the associated location of each distant rendering resource and the new location; The average time required for a user to move a unit step in the virtual scene.

4. The method for providing rendering resources for a metaverse scene according to claim 1, characterized in that, The user experience quality value model is used to quantify latency and rendering quality into user experience quality value, expressed as: in, Value for user experience quality; and These are the user's preference weights for latency and rendering quality, respectively, and satisfy the following conditions: + =1; For time delay; For rendering quality; Let be a non-increasing utility function with respect to time delay; This is a non-decreasing utility function with respect to rendering quality.

5. The method for providing rendering resources for a metaverse scene according to claim 4, characterized in that, The social welfare score corresponding to each bidding information is calculated based on the promised latency, promised rendering quality, dynamic latency constraints, and a preset user experience quality value model, including: Determine whether the promised delivery delay for each bidding information meets the corresponding dynamic delay constraint; If satisfied, the promised latency and promised rendering quality are input into the user experience quality value model to calculate the corresponding user experience quality value. Subtracting the bid price from the user experience quality value yields the social welfare score corresponding to the bid information.

6. The method for providing rendering resources for a metaverse scene according to claim 5, characterized in that, The selection of a target service provider from among the candidate service providers based on the social welfare score includes: The candidate service provider with the highest social welfare score is selected as the target service provider.

7. The method for providing rendering resources for a metaverse scene according to claim 6, characterized in that, The preset payment rule is the VCG payment rule, and the determination of the service price to be paid to the target service provider based on the preset payment rule includes: The user experience quality value corresponding to the bidding information provided by the target service provider is denoted as the first value; Among all the social welfare scores corresponding to the bidding information of all other candidate service providers, the highest social welfare score is determined and recorded as the second score; Subtracting the second rating from the first value yields the service price payable to the target service provider.

8. A method for providing rendering resources for a metaverse scene according to any one of claims 1 to 7, characterized in that, The verification of the actual latency and rendering quality includes: After acquiring the distant rendering resources, the actual latency and the measured rendering quality are then measured. The actual latency is compared with the latency promised by the target service provider, and the measured rendering quality is compared with the rendering quality promised by the target service provider. If any of the promised standards are not met, the service price paid to the target service provider will be reduced according to the degree of breach.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements a method for providing rendering resources for a metaverse scene as described in any one of claims 1 to 8.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements a method for providing rendering resources for a metaverse scene as described in any one of claims 1 to 8.

Images