Method for constructing virtual reality machine room positioning based on digital twin engine

Abstract

The invention provides a method for constructing virtual reality machine room positioning based on a digital twin engine, and the method comprises the steps: constructing a calculation power conversion model, deploying the calculation power conversion model on a virtual model operation platform, carrying out the calculation power conversion processing of a local algorithm task of an intelligent calculation module, judging whether the platform is in an idle calculation power resource sufficient state, and carrying out the calculation power conversion. Therefore, whether physical computing power resource mapping is carried out on the platform is determined, and the platform is ensured to have sufficient computing power resources to carry out machine room positioning calculation. Federated learning iteration processing is carried out on the improved hybrid weighting algorithm to obtain an optimal solution of the improved hybrid weighting algorithm; the virtual machine room scene and the real machine room scene are subjected to improved hybrid weighting algorithm processing with an optimal solution about UWB indoor positioning to obtain node coordinates of the machine room, so that a positioning result of the virtual reality machine room is determined, and an accurate and reliable virtual reality machine room positioning result is provided for digital twin engine deployment; and the machine room positioning requirements of digital twin engine deployment in different scenes are met.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence, and in particular to a method for constructing virtual reality computer room positioning based on a digital twin engine. Background Art

[0002] The most important deployment of the digital twin engine is the deployment of the computing interaction layer, that is, the deployment of the intelligent computing module and the interactive driving module. Among them, the intelligent computing module is divided into two parts: those that need to rely on the big data framework and those that do not. The part that needs to rely on the big data framework needs to deploy a dedicated server, such as Hadoop, and the part that does not rely on the big data framework (that is, the part that only relies on local algorithm calculations) only needs to be deployed on the virtual model running platform and can be implemented using C# and Unity. The deployment of the data twin engine has the problem of inaccurate regional positioning in both virtual and real scenes, which is mainly affected by factors such as computing power resources, hardware equipment performance, and data differentiation. In order to improve the accuracy and reliability of the deployment of the digital twin engine, it is necessary to accurately locate the computer room in the virtual scene and the real scene. Summary of the invention

[0003] The purpose of the present invention is to provide a method for constructing virtual reality computer room positioning based on a digital twin engine, construct a computing power quantization model and deploy it on a virtual model running platform, perform computing power quantization processing on the local algorithm tasks of the intelligent computing module, and judge whether the platform is in a state of sufficient idle computing power resources, so as to determine whether to perform physical computing power resource mapping on the platform, and ensure that the platform has sufficient computing power resources for computer room positioning calculation; the improved hybrid weighted algorithm is also subjected to federated learning iterative processing to obtain the optimal solution of the improved hybrid weighted algorithm; the virtual computer room scene and the real computer room scene are then subjected to the improved hybrid weighted algorithm with the optimal solution for UWB indoor positioning to obtain the node coordinates of the computer room, thereby determining the positioning result of the virtual reality computer room, providing accurate and reliable virtual reality computer room positioning results for the deployment of the digital twin engine, and meeting the computer room positioning requirements of the digital twin engine deployment in different scenarios.

[0004] The present invention is achieved through the following technical solutions:

[0005] The method for constructing virtual reality computer room positioning based on the digital twin engine includes:

[0006] Construct a computing power quantization model and deploy it on the virtual model operation platform; perform computing power quantization processing on the local algorithm tasks of the intelligent computing module of the virtual model operation platform, and determine whether the virtual model operation platform is in a state of sufficient idle computing power resources, so as to determine whether to perform physical computing power resource mapping on the virtual model operation platform;

[0007] The improved hybrid weighted algorithm is iteratively processed by federated learning to obtain the optimal solution of the improved hybrid weighted algorithm; the virtual computer room scene and the real computer room scene corresponding to the virtual model operation platform are processed by the improved hybrid weighted algorithm with the optimal solution for UWB indoor positioning to obtain the node coordinates of the virtual computer room and the real computer room, thereby obtaining the positioning results of the virtual computer room and the real computer room.

[0008] Optionally, a computing power quantization model is built and deployed on a virtual model operation platform, including:

[0009] Build a computing power quantification model based on the data layer, model layer, and computing interaction layer of the digital twin engine architecture;

[0010] Based on the platform address corresponding to the virtual model running platform, the computing power quantization model is deployed on the virtual model running platform.

[0011] Optionally, performing computational quantization processing on the local algorithm task of the intelligent computing module of the virtual model operation platform includes:

[0012] Obtaining a local algorithm task creation log of the intelligent computing module of the virtual model running platform, analyzing the local algorithm task creation log, and obtaining all local algorithm tasks that have been created by the intelligent computing module;

[0013] Based on the computing power quantification model, computing power quantification processing is performed on all local algorithm tasks that have been created to obtain the computing power resource demand information of each of the local algorithm tasks that have been created; wherein the computing power resource demand information includes the computing power resource types required by all local algorithm tasks that have been created and the demand for each type of computing power resources.

[0014] Optionally, determining whether the virtual model operation platform is in a state of sufficient idle computing resources includes:

[0015] Obtaining idle computing resource characteristic information of the virtual model operation platform; wherein the idle computing resource characteristic information includes the type of computing resources in an idle state in the virtual model operation platform and the available amount of each type of computing resources;

[0016] The idle computing power resource characteristic information is compared with the computing power resource demand information to determine whether the virtual model operation platform is in a state of sufficient idle computing power resources.

[0017] Optionally, comparing the idle computing resource characteristic information with the computing resource demand information to determine whether the virtual model operation platform is in a state of sufficient idle computing resources includes:

[0018] Compare the idle computing resource characteristic information with the computing resource demand information to obtain a comparison result;

[0019] When the comparison result shows that the available amounts of computing resources corresponding to the computing resource types in the idle state in the virtual model operation platform all meet the required amount of resources corresponding to the computing resource demand information, then the available amount value of the computing resource corresponding to the computing resource types in the idle state in the virtual model operation platform and the corresponding value of the required amount of resources in the computing resource demand information are retrieved;

[0020] Obtaining computing power margin indicator parameters using the available amount value of computing power resources corresponding to the computing power resource type in an idle state in the virtual model operation platform and the corresponding value of the required resource amount in the computing power resource demand information;

[0021] The computing power margin index parameter is obtained by the following formula:

[0022]

[0023] Where S represents the computing power margin indicator parameter; n represents the number of computing power resource types in idle state in the virtual model operation platform; K i K represents the available amount corresponding to the i-th idle computing resource type; yi K represents the required resource amount in the computing resource demand information corresponding to the i-th computing resource type in the idle state; b represents the standard deviation of the available amount corresponding to n idle computing resource types; e represents the dynamic adjustment coefficient, and the dynamic adjustment coefficient is obtained by the following formula:

[0024]

[0025] Where, e represents the dynamic adjustment coefficient; K max Indicates the maximum available amount of computing resources in idle state; K ymax Indicates the required resource quantity of the idle computing resource type corresponding to the maximum available quantity of the computing resource type in the idle state; K min Indicates the minimum available value of the computing resource type in idle state; K ymin Indicates the required resource amount of the idle computing resource type corresponding to the minimum available amount of the idle computing resource type;

[0026] Comparing the computing power margin indicator parameter with a preset indicator parameter threshold;

[0027] When the computing power margin indicator parameter is lower than the preset indicator parameter threshold, it is determined that there is computing power margin but the margin is insufficient, and an insufficient computing power margin warning is issued.

[0028] Optionally, determining whether to perform physical computing resource mapping on the virtual model operation platform includes:

[0029] If the virtual model operation platform is in a state with sufficient idle computing power resources, determining not to perform physical computing power resource mapping on the virtual model operation platform;

[0030] If the virtual model operation platform is not in a state of sufficient idle computing power resources, physical computing power resource mapping is performed on the first area part that relies on local algorithm calculation and the second area part that does not rely on local algorithm calculation between the intelligent computing model and the interactive driving model of the virtual model operation platform.

[0031] Optionally, performing federated learning iteration processing on the improved hybrid weighted algorithm to obtain an optimal solution of the improved hybrid weighted algorithm includes:

[0032] The improved Chan-Taylor hybrid weighted algorithm is subjected to federated learning iterative processing to obtain an optimal solution for algorithm parameters of the improved Chan-Taylor hybrid weighted algorithm.

[0033] Optionally, the virtual machine room scene and the real machine room scene corresponding to the virtual model running platform are processed by an improved hybrid weighted algorithm with an optimal solution for UWB indoor positioning to obtain the node coordinates of the virtual machine room and the real machine room, thereby obtaining the positioning results of the virtual machine room and the real machine room, including:

[0034] The virtual machine room scene and the real machine room scene corresponding to the virtual model operation platform are processed by the improved Chan-Taylor hybrid weighted algorithm with the optimal solution of the algorithm parameters for UWB indoor positioning to obtain the node coordinates of the virtual machine room and the real machine room;

[0035] Based on the node coordinates of the virtual machine room and the real machine room, the coverage ranges of the virtual machine room and the real machine room are delineated, thereby obtaining positioning results of the virtual machine room and the real machine room.

[0036] Optionally, the virtual machine room scene and the real machine room scene corresponding to the virtual model running platform are processed by an improved Chan-Taylor hybrid weighted algorithm with an optimal solution of the algorithm parameters for UWB indoor positioning to obtain node coordinates of the virtual machine room and the real machine room, including:

[0037] Perform UWB indoor positioning on the virtual computer room scene and the real computer room scene, determine corresponding reference nodes and unknown nodes, and obtain the coordinates and TDOA values ​​of the reference nodes;

[0038] Performing Chan algorithm calculation on the coordinates and TDOA values ​​of the reference nodes to obtain the initial coordinates of the unknown nodes;

[0039] Based on the initial coordinates of the unknown node, Taylor series expansion calculation is performed to estimate the coordinates of the unknown node again;

[0040] Based on the pre-calculated weighting coefficients, the initial coordinates of the unknown node and the estimated coordinates are weighted to obtain the final coordinates of the unknown node, which are used as the node coordinates of the virtual computer room and the real computer room.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] The method for constructing virtual reality computer room positioning based on a digital twin engine provided in the present application constructs a computing power quantization model and deploys it on a virtual model running platform, performs computing power quantization processing on the local algorithm tasks of the intelligent computing module, and determines whether the platform is in a state of sufficient idle computing power resources, thereby determining whether to map the physical computing power resources of the platform to ensure that the platform has sufficient computing power resources for computer room positioning calculations; it also performs federated learning iterative processing on the improved hybrid weighted algorithm to obtain the optimal solution of the improved hybrid weighted algorithm; then, the improved hybrid weighted algorithm with the optimal solution for UWB indoor positioning is processed on the virtual computer room scene and the real computer room scene to obtain the node coordinates of the computer room, thereby determining the positioning result of the virtual computer room, providing accurate and reliable virtual reality computer room positioning results for the digital twin engine deployment, and meeting the computer room positioning requirements of the digital twin engine deployment in different scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0044] Figure 1 A flowchart of a method for constructing virtual reality computer room positioning based on a digital twin engine provided by the present invention. DETAILED DESCRIPTION

[0045] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are described in detail below in conjunction with the accompanying drawings. It is to be understood that the specific embodiments described herein are only used to explain the present application, rather than to limit the present application. It should also be noted that, for ease of description, only some structures related to the present application are shown in the accompanying drawings, rather than all structures. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0046] The terms "include" and "have" and any variations thereof in this application are intended to cover non-exclusive inclusions. For example, a process, method, method, product or device comprising a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products or devices.

[0047] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0048] See also Figure 1 As shown, an embodiment of the present application provides a method for constructing a virtual reality computer room positioning based on a digital twin engine. The method for constructing a virtual reality computer room positioning based on a digital twin engine includes:

[0049] Construct a computing power quantization model and deploy it on the virtual model operation platform; perform computing power quantization on the local algorithm tasks of the intelligent computing module of the virtual model operation platform, and determine whether the virtual model operation platform is in a state of sufficient idle computing power resources, so as to determine whether to perform physical computing power resource mapping on the virtual model operation platform;

[0050] The improved hybrid weighted algorithm is iteratively processed by federated learning to obtain the optimal solution of the improved hybrid weighted algorithm; the virtual computer room scene and the real computer room scene corresponding to the virtual model operation platform are processed with the improved hybrid weighted algorithm with the optimal solution for UWB indoor positioning to obtain the node coordinates of the virtual computer room and the real computer room, thereby obtaining the positioning results of the virtual computer room and the real computer room.

[0051] The beneficial effects of the above embodiments are as follows: the method for constructing virtual reality computer room positioning based on the digital twin engine constructs a computing power quantization model and deploys it on the virtual model operation platform, performs computing power quantization processing on the local algorithm tasks of the intelligent computing module, and determines whether the platform is in a state of sufficient idle computing power resources, thereby determining whether to map the physical computing power resources of the platform to ensure that the platform has sufficient computing power resources for computer room positioning calculations; it also performs federated learning iterative processing on the improved hybrid weighted algorithm to obtain the optimal solution of the improved hybrid weighted algorithm; then the virtual computer room scene and the real computer room scene are processed with the improved hybrid weighted algorithm with the optimal solution for UWB indoor positioning to obtain the node coordinates of the computer room, thereby determining the positioning result of the virtual reality computer room, providing accurate and reliable virtual reality computer room positioning results for the digital twin engine deployment, and meeting the computer room positioning requirements of the digital twin engine deployment in different scenarios.

[0052] In another embodiment, a computing power quantization model is constructed and deployed on a virtual model operation platform, including:

[0053] Build a computing power quantification model based on the data layer, model layer, and computing interaction layer of the digital twin engine architecture;

[0054] Based on the platform address corresponding to the virtual model operation platform, the computing power quantization model is deployed on the virtual model operation platform.

[0055] The beneficial effects of the above embodiments are that network virtualization technology refers to supporting multiple virtual networks on a common physical network through abstraction, allocation, and isolation mechanisms. Each virtual network can use independent protocol systems and can reasonably configure the node resources and link resources in the entire network according to dynamically changing user needs. Each virtual network is a resource slice of the underlying network, which consists of virtual nodes (such as virtual routers) and virtual links. Network virtualization technology needs to use a virtual model operation platform to locate the virtual nodes and real nodes therein, so as to achieve accurate division and detection of virtual networks. In order to reasonably allocate the node resources within each virtual network during the network virtualization process, it is necessary to quantify the computing power resources of each node to provide a reliable and comprehensive basis for subsequent node resource allocation. For this reason, a computing power quantization model is constructed based on the data layer, model layer, and computing interaction layer of the digital twin engine architecture to ensure that the computing power quantization model can have accurate and efficient computing power quantization computing performance. In addition, based on the platform address corresponding to the virtual model operation platform, the computing power quantization model is deployed on the virtual model operation platform, which is convenient for the virtual model operation platform to call the computing power quantization model at any time and improve the real-time performance of the quantitative calculation of computing power resources.

[0056] In another embodiment, the local algorithm task of the intelligent computing module of the virtual model operation platform is subjected to computing power quantization processing, including:

[0057] Obtaining a local algorithm task creation log of the intelligent computing module of the virtual model running platform, analyzing the local algorithm task creation log, and obtaining all local algorithm tasks that have been created by the intelligent computing module;

[0058] Based on the computing power quantification model, computing power quantification is performed on all local algorithm tasks that have been created to obtain the computing power resource demand information of each of the local algorithm tasks that have been created; wherein the computing power resource demand information includes the computing power resource types required by all the local algorithm tasks that have been created and the demand for each type of computing power resources.

[0059] The beneficial effects of the above embodiments are that the virtual model operation platform includes an intelligent computing module and an interactive driving module; wherein the intelligent computing model is used to undertake the calculation of the corresponding algorithm tasks, and the interactive driving model is used to push and drive messages with the outside world. The calculation of the algorithm tasks by the intelligent computing model directly affects the availability of computing power resources of the virtual model operation platform. In order to accurately quantify the local algorithm tasks of the intelligent computing module, first obtain and analyze the local algorithm task creation log of the intelligent computing model to obtain all the local algorithm tasks that have been created by the intelligent computing module, and then use the computing power quantification model to perform computing power quantification on all the local algorithm tasks that have been created to obtain the computing power resource demand information of each of the local algorithm tasks that have been created, so as to provide reliable data basis for the subsequent judgment of whether the virtual model operation platform is in a state of sufficient idle computing power resources.

[0060] In another embodiment, determining whether the virtual model operation platform is in a state of sufficient idle computing resources includes:

[0061] Obtaining idle computing resource characteristic information of the virtual model operation platform; wherein the idle computing resource characteristic information includes the type of computing resources in an idle state in the virtual model operation platform and the available amount of each type of computing resources;

[0062] The idle computing power resource characteristic information is compared with the computing power resource demand information to determine whether the virtual model operation platform is in a state of sufficient idle computing power resources.

[0063] The beneficial effect of the above-mentioned embodiment is that whether the idle computing power resources of the virtual model operation platform are sufficient or not depends on the situation of the computing power resources in the virtual model operation platform itself that are in an idle state and the demand for computing power resources. To this end, the types of computing power resources in the virtual model operation platform that are in an idle state and the available amount of computing power resources of each type are obtained, so as to globally quantify the computing power resources in an idle state. Then compare the idle computing power resource feature information with the computing power resource demand information. If the available amount of any type of computing power resources in an idle state is less than the demand for the corresponding type of computing power resources, it is judged that the virtual model operation platform is not in a state of sufficient idle computing power resources; otherwise, it is judged that the virtual model operation platform is in a state of sufficient idle computing power resources, so as to facilitate the subsequent use of physical computing power resource mapping to improve the computing power resources of the virtual model operation platform in a targeted manner when the idle computing power resources are insufficient.

[0064] In another embodiment, comparing the idle computing resource characteristic information with the computing resource demand information to determine whether the virtual model operation platform is in a state of sufficient idle computing resources includes:

[0065] Compare the idle computing resource characteristic information with the computing resource demand information to obtain a comparison result;

[0066] When the comparison result shows that the available amounts of computing resources corresponding to the computing resource types in the idle state in the virtual model operation platform all meet the required amount of resources corresponding to the computing resource demand information, then the available amount value of the computing resource corresponding to the computing resource types in the idle state in the virtual model operation platform and the corresponding value of the required amount of resources in the computing resource demand information are retrieved;

[0067] Obtaining computing power margin indicator parameters using the available amount value of computing power resources corresponding to the computing power resource type in an idle state in the virtual model operation platform and the corresponding value of the required resource amount in the computing power resource demand information;

[0068] The computing power margin index parameter is obtained by the following formula:

[0069]

[0070] Where S represents the computing power margin indicator parameter; n represents the number of computing power resource types in idle state in the virtual model operation platform; K i K represents the available amount corresponding to the i-th idle computing resource type; yi K represents the required resource amount in the computing resource demand information corresponding to the i-th computing resource type in the idle state; brepresents the standard deviation of the available amount corresponding to n idle computing resource types; e represents the dynamic adjustment coefficient, and the dynamic adjustment coefficient is obtained by the following formula:

[0071]

[0072] Where, e represents the dynamic adjustment coefficient; K max Indicates the maximum available amount of computing resources in idle state; K ymax Indicates the required resource quantity of the idle computing resource type corresponding to the maximum available quantity of the computing resource type in the idle state; K min Indicates the minimum available value of the computing resource type in idle state; K ymin Indicates the required resource amount of the idle computing resource type corresponding to the minimum available amount of the idle computing resource type;

[0073] Comparing the computing power margin indicator parameter with a preset indicator parameter threshold;

[0074] When the computing power margin indicator parameter is lower than the preset indicator parameter threshold, it is determined that there is computing power margin but the margin is insufficient, and an insufficient computing power margin warning is issued.

[0075] The beneficial effect of the above embodiment is that by comparing the idle computing power resource feature information with the computing power resource demand information, it is possible to accurately determine whether the virtual model operation platform is in a state of sufficient idle computing power resources. This method avoids a simple "yes / no" judgment, but makes a detailed comparison based on the specific computing power resource type, available amount and required resource amount, thereby providing a more accurate and detailed computing power resource evaluation. By introducing the computing power margin index parameter S, the technical solution can quantitatively evaluate the difference or margin between the idle computing power resources and the demand. This quantitative evaluation helps to more intuitively understand the current surplus of computing power resources and provide data support for subsequent resource allocation, scheduling and optimization. The introduction of the dynamic adjustment coefficient e enables the calculation of the computing power margin index parameter to take into account the differences in the available amounts of different computing power resource types and the changes in demand. This dynamic adjustment mechanism enhances the flexibility and adaptability of the evaluation, and can better cope with the fluctuations and changes in computing power resources in actual operation. When the computing power margin index parameter is lower than the preset index parameter threshold, the technical solution can timely issue an early warning of insufficient computing power margin. This early warning mechanism helps to promptly discover and resolve computing resource shortages, and avoid model operation obstructions or performance degradation due to insufficient resources. Through the implementation of the above technical solutions, computing resources can be allocated and scheduled more effectively. Based on the understanding of the current computing resource status, resources can be optimized according to actual needs to improve resource utilization and model operation efficiency. Through accurate evaluation of computing resource status and the implementation of a timely early warning mechanism, this technical solution helps to improve the stability and reliability of the system. By avoiding problems such as computing resource shortages and overloads, the continuity and stability of model operation can be ensured, thereby improving the performance and reliability of the overall system.

[0076] In summary, the technical effects of the above technical solutions on performance indicators are mainly reflected in accurate evaluation, quantitative computing power margin, dynamic adjustment, early warning mechanism, resource optimization allocation and scheduling, and improvement of system stability and reliability. These effects work together on the virtual model operation platform to improve its overall performance and operation efficiency.

[0077] In another embodiment, determining whether to perform physical computing resource mapping on the virtual model operation platform includes:

[0078] If the virtual model operation platform is in a state of sufficient idle computing power resources, it is determined not to perform physical computing power resource mapping on the virtual model operation platform;

[0079] If the virtual model operation platform is not in a state of sufficient idle computing power resources, physical computing power resource mapping is performed on the first area part that relies on local algorithm calculation and the second area part that does not rely on local algorithm calculation between the intelligent computing model and the interactive driving model of the virtual model operation platform.

[0080] The beneficial effect of the above embodiment is that when the virtual model operation platform is in a state of sufficient idle computing power resources, it is determined that there is no need to perform physical computing power resource mapping on the virtual model operation platform, that is, there is no need to additionally map physical computing power resources from the outside. When the virtual model operation platform is not in a state of sufficient idle computing power resources, it is determined that physical computing power resource mapping is required for the virtual model operation platform. At this time, the first area part that depends on the local algorithm calculation and the second area part that does not depend on the local algorithm calculation between the intelligent computing model and the interactive driving model of the virtual model operation platform are mapped with physical computing power resources to ensure that the virtual model operation platform can obtain reliable and accurate physical computing power resource mapping in different types of areas. Specifically, according to the mapping process, components {ef}, {d, e}, {a, b}, {b, c} complete the second area part that does not depend on the local algorithm calculation in turn, but there are still two restricted parts {d, e}, {b, c} and a fixed component {c, d}; the mapping of the first area part that depends on the local algorithm calculation can be converted into a virtual node mapping and a mapping of a restricted link connected to the virtual node. Among all virtual nodes associated with the second region portion that does not rely on local algorithm calculation, the priority mapping The larger virtual node, where L C (n v ) represents the set of virtual links connected to the virtual node. v When, it is in the set of all available physical nodes S c Choose such that the mapping is consistent with n v The connected link is the physical node that consumes the least bandwidth resources. To map fixed components in a virtual network, since the virtual nodes at both ends of the fixed components have been mapped, the mapping of fixed components only requires the mapping of fixed links. The virtual nodes at both ends have been mapped to and Therefore, it is only necessary to map the fixed link to and Determine the shortest path.

[0081] In another embodiment, the improved hybrid weighted algorithm is subjected to federated learning iteration processing to obtain an optimal solution of the improved hybrid weighted algorithm, including:

[0082] The improved Chan-Taylor hybrid weighted algorithm is subjected to federated learning iterative processing to obtain the optimal solution of the algorithm parameters of the improved Chan-Taylor hybrid weighted algorithm.

[0083] The beneficial effects of the above embodiments are that in order to improve the accuracy and consistency of positioning coordinates in the virtual machine room scene and the real machine room scene, the improved Chan-Taylor hybrid weighted algorithm in the UWB indoor positioning technology is used for precise positioning calculation, thereby completing the precise layout of the virtual machine room scene. In actual operation, the improved Chan-Taylor hybrid weighted algorithm is subjected to federated learning iterative processing to obtain the optimal solution of the algorithm parameters of the improved Chan-Taylor hybrid weighted algorithm. Specifically, in the Nth training iteration of the federated learning iterative processing of the improved Chan-Taylor hybrid weighted algorithm, the user terminal performs local training based on the global model downloaded from the FL server using the locally collected images / videos. Then the terminal reports the intermediate training results (such as the gradient of the DNN) of the Nth iteration to the FL server through the 5G line channel. At the same time, the FL server sends the global model and training configuration of the (N+1)th intersection to the user terminal. While the FL server aggregates the gradients collected in the Nth iteration, the terminal can perform training for the (N1)th iteration. Federated Aggregation outputs an updated global model for the next iteration of training, and distributes the model to the terminal together with the updated training configuration. In the Nth training iteration, the user terminal performs local training based on the global model downloaded from the FL server using locally collected images / videos. The terminal then reports the intermediate training results of the Nth iteration (such as the gradient of the DNN) to the FL server through the 5G line channel. At the same time, the FL server sends the (N+1)th global model and training configuration to the terminal. While the FL server aggregates (Aggregation) the gradients collected in the Nth iteration, the terminal can perform training for the (N1)th iteration. Federated Aggregation outputs an updated global model for the next iteration of training, and distributes the model to the terminal together with the updated training configuration.

[0084] In another embodiment, the virtual machine room scene and the real machine room scene corresponding to the virtual model operation platform are processed by an improved hybrid weighted algorithm with an optimal solution for UWB indoor positioning to obtain the node coordinates of the virtual machine room and the real machine room, thereby obtaining the positioning results of the virtual machine room and the real machine room, including:

[0085] The virtual computer room scene and the real computer room scene corresponding to the virtual model operation platform are processed by the improved Chan-Taylor hybrid weighted algorithm with the optimal solution of the algorithm parameters for UWB indoor positioning to obtain the node coordinates of the virtual computer room and the real computer room;

[0086] Based on the node coordinates of the virtual machine room and the real machine room, the coverage ranges of the virtual machine room and the real machine room are delineated, thereby obtaining positioning results of the virtual machine room and the real machine room.

[0087] The beneficial effect of the above embodiment is that the virtual room scene and the real room scene corresponding to the virtual model operation platform are processed by the improved Chan-Taylor hybrid weighted algorithm with the optimal solution of the algorithm parameters for UWB indoor positioning, and the node coordinates of the virtual room and the real room are accurately calculated by combining the UWB indoor positioning technology and the improved Chan-Taylor hybrid weighted algorithm. Based on the node coordinates of the virtual room and the real room, the coverage range of the virtual room and the real room is delineated, so as to obtain the positioning results of the virtual room and the real room, and complete the accurate layout of the virtual room scene.

[0088] In another embodiment, the virtual machine room scene and the real machine room scene corresponding to the virtual model running platform are processed by an improved Chan-Taylor hybrid weighted algorithm with an optimal solution of the algorithm parameters for UWB indoor positioning to obtain the node coordinates of the virtual machine room and the real machine room, including:

[0089] Perform UWB indoor positioning on the virtual computer room scene and the real computer room scene, determine the corresponding reference nodes and unknown nodes, and obtain the coordinates and TDOA values ​​of the reference nodes;

[0090] The coordinates and TDOA value of the reference node are calculated using the Chan algorithm to obtain the initial coordinates of the unknown node;

[0091] Based on the initial coordinates of the unknown node, Taylor series expansion calculation is performed to estimate the coordinates of the unknown node again;

[0092] Based on the pre-calculated weighting coefficient, the initial coordinates of the unknown node and the estimated coordinates are weighted to obtain the final coordinates of the unknown node, which are used as the node coordinates of the virtual computer room and the real computer room.

[0093] The beneficial effect of the above embodiment is that in the actual operation process, [x(k), y(k)] is defined as the coordinates of the unknown node calculated using the hth method. In the TDOA positioning using the Chan algorithm, the value that can be directly obtained is the difference in distance, so it is more convenient to use the difference in distance. Therefore, define The difference between the distance between the unknown node measured coordinates to anchor node i and the distance to anchor node 1 using the kth method, [x1, y1] and [x i ,y i ] are the coordinates of node i and node 1, that is,

[0094]

[0095] Define Δr as the square of the difference between the true value and the measured value:

[0096]

[0097] Define the weighting coefficient as η(k)

[0098]

[0099] Where η(k) is the number of reference base stations and n is the number of anchor nodes.

[0100] From the above algorithm, it can be seen that the smaller Δr is, the more accurate the estimated value is; therefore, the smaller the weighting coefficient η(k) is, the more accurate the calculated position coordinates are, that is, the measurement value obtained by this positioning method should account for a larger proportion and should be multiplied by the inverse of the weighting coefficient η(k). Then the final unknown node estimated coordinates can be calculated as:

[0101]

[0102] Through the above method, accurate and reliable virtual reality computer room positioning results can be obtained to meet the computer room positioning requirements of digital twin engine deployment in different scenarios.

[0103] In general, the method for constructing virtual reality computer room positioning based on the digital twin engine constructs a computing power quantization model and deploys it on the virtual model operation platform, performs computing power quantization processing on the local algorithm tasks of the intelligent computing module, and determines whether the platform is in a state of sufficient idle computing power resources, thereby determining whether to map the physical computing power resources of the platform to ensure that the platform has sufficient computing power resources for computer room positioning calculations; it also performs federated learning iterative processing on the improved hybrid weighted algorithm to obtain the optimal solution of the modified hybrid weighted algorithm; then the virtual computer room scene and the real computer room scene are processed with the improved hybrid weighted algorithm with the optimal solution for UWB indoor positioning to obtain the node coordinates of the computer room, thereby determining the positioning result of the virtual reality computer room, providing accurate and reliable virtual reality computer room positioning results for the digital twin engine deployment, and meeting the computer room positioning requirements of the digital twin engine deployment in different scenarios.

[0104] The above is only a specific implementation of the present invention, and any other improvements made based on the concept of the present invention are considered to be within the protection scope of the present invention.

Claims

1. A method for constructing virtual reality computer room positioning based on a digital twin engine, characterized in that: include: Build a computing power quantization model and deploy it on a virtual model operation platform; Performing computing power quantization processing on the local algorithm tasks of the intelligent computing module of the virtual model operation platform, and judging whether the virtual model operation platform is in a state of sufficient idle computing power resources, thereby determining whether to perform physical computing power resource mapping on the virtual model operation platform; Performing federated learning iterative processing on the improved hybrid weighted algorithm to obtain an optimal solution of the improved hybrid weighted algorithm; The virtual machine room scene and the real machine room scene corresponding to the virtual model operation platform are processed by an improved hybrid weighted algorithm with an optimal solution for UWB indoor positioning to obtain the node coordinates of the virtual machine room and the real machine room, thereby obtaining the positioning results of the virtual machine room and the real machine room.

2. The method for constructing virtual reality computer room positioning based on a digital twin engine according to claim 1, characterized in that: Build a computing power quantization model and deploy it on the virtual model operation platform, including: Build a computing power quantification model based on the data layer, model layer, and computing interaction layer of the digital twin engine architecture; Based on the platform address corresponding to the virtual model running platform, the computing power quantization model is deployed on the virtual model running platform.

3. The method for constructing virtual reality computer room positioning based on digital twin engine according to claim 2, characterized in that: Performing computational power quantization processing on the local algorithm tasks of the intelligent computing module of the virtual model operation platform includes: Obtaining a local algorithm task creation log of the intelligent computing module of the virtual model running platform, analyzing the local algorithm task creation log, and obtaining all local algorithm tasks that have been created by the intelligent computing module; Based on the computing power quantification model, computing power quantification processing is performed on all local algorithm tasks that have been created to obtain the computing power resource demand information of each of the local algorithm tasks that have been created; wherein the computing power resource demand information includes the computing power resource types required by all local algorithm tasks that have been created and the demand for each type of computing power resources.

4. The method for constructing virtual reality room positioning based on digital twin engine according to claim 3, characterized in that: Determining whether the virtual model operation platform is in a state of sufficient idle computing resources includes: Obtaining idle computing resource characteristic information of the virtual model operation platform; wherein the idle computing resource characteristic information includes the type of computing resources in an idle state in the virtual model operation platform and the available amount of each type of computing resources; The idle computing power resource characteristic information is compared with the computing power resource demand information to determine whether the virtual model operation platform is in a state of sufficient idle computing power resources.

5. The method for constructing virtual reality computer room positioning based on digital twin engine according to claim 3, characterized in that: Comparing the idle computing resource characteristic information with the computing resource demand information to determine whether the virtual model operation platform is in a state of sufficient idle computing resources includes: Compare the idle computing resource characteristic information with the computing resource demand information to obtain a comparison result; When the comparison result shows that the available amounts of computing resources corresponding to the computing resource types in the idle state in the virtual model operation platform all meet the required amount of resources corresponding to the computing resource demand information, then the available amount value of the computing resource corresponding to the computing resource types in the idle state in the virtual model operation platform and the corresponding value of the required amount of resources in the computing resource demand information are retrieved; Obtaining computing power margin indicator parameters using the available amount value of computing power resources corresponding to the computing power resource type in an idle state in the virtual model operation platform and the corresponding value of the required resource amount in the computing power resource demand information; The computing power margin index parameter is obtained by the following formula: Where S represents the computing power margin indicator parameter; n represents the number of computing power resource types in idle state in the virtual model operation platform; K i K represents the available amount corresponding to the i-th idle computing resource type; yi K represents the required resource amount in the computing resource demand information corresponding to the i-th computing resource type in the idle state; b represents the standard deviation of the available amount corresponding to n idle computing resource types; e represents the dynamic adjustment coefficient, and the dynamic adjustment coefficient is obtained by the following formula: Where, e represents the dynamic adjustment coefficient; K max Indicates the maximum available amount of computing resources in idle state; K ymax Indicates the required resource quantity of the idle computing resource type corresponding to the maximum available quantity of the computing resource type in the idle state; K min Indicates the minimum available value of the computing resource type in idle state; K ymin Indicates the required resource amount of the idle computing resource type corresponding to the minimum available amount of the idle computing resource type; Comparing the computing power margin indicator parameter with a preset indicator parameter threshold; When the computing power margin indicator parameter is lower than the preset indicator parameter threshold, it is determined that there is computing power margin but the margin is insufficient, and an insufficient computing power margin warning is issued.

6. The method for constructing virtual reality room positioning based on digital twin engine according to claim 4, characterized in that: Determining whether to perform physical computing resource mapping on the virtual model operation platform includes: If the virtual model operation platform is in a state with sufficient idle computing power resources, determining not to perform physical computing power resource mapping on the virtual model operation platform; If the virtual model operation platform is not in a state of sufficient idle computing power resources, physical computing power resource mapping is performed on the first area part that relies on local algorithm calculation and the second area part that does not rely on local algorithm calculation between the intelligent computing model and the interactive driving model of the virtual model operation platform.

7. The method for constructing virtual reality room positioning based on digital twin engine according to claim 1, characterized in that: The improved hybrid weighted algorithm is subjected to federated learning iteration processing to obtain an optimal solution of the improved hybrid weighted algorithm, including: The improved Chan-Taylor hybrid weighted algorithm is subjected to federated learning iterative processing to obtain an optimal solution for algorithm parameters of the improved Chan-Taylor hybrid weighted algorithm.

8. The method for constructing virtual reality room positioning based on digital twin engine according to claim 7, characterized in that: The virtual machine room scene and the real machine room scene corresponding to the virtual model operation platform are processed by an improved hybrid weighted algorithm with an optimal solution for UWB indoor positioning to obtain the node coordinates of the virtual machine room and the real machine room, thereby obtaining the positioning results of the virtual machine room and the real machine room, including: The virtual machine room scene and the real machine room scene corresponding to the virtual model operation platform are processed by the improved Chan-Taylor hybrid weighted algorithm with the optimal solution of the algorithm parameters for UWB indoor positioning to obtain the node coordinates of the virtual machine room and the real machine room; Based on the node coordinates of the virtual machine room and the real machine room, the coverage ranges of the virtual machine room and the real machine room are delineated, thereby obtaining positioning results of the virtual machine room and the real machine room.

9. The method for constructing virtual reality room positioning based on digital twin engine according to claim 8, characterized in that: The virtual machine room scene and the real machine room scene corresponding to the virtual model operation platform are processed by the improved Chan-Taylor hybrid weighted algorithm with the optimal solution of the algorithm parameters for UWB indoor positioning to obtain the node coordinates of the virtual machine room and the real machine room, including: Perform UWB indoor positioning on the virtual computer room scene and the real computer room scene, determine the corresponding reference nodes and unknown nodes, and obtain the coordinates and TDOA values ​​of the reference nodes; perform Chan algorithm calculation on the coordinates and TDOA values ​​of the reference nodes to obtain the initial coordinates of the unknown nodes; Based on the initial coordinates of the unknown node, Taylor series expansion calculation is performed to estimate the coordinates of the unknown node again; Based on the pre-calculated weighting coefficients, the initial coordinates of the unknown node and the estimated coordinates are weighted to obtain the final coordinates of the unknown node, which are used as the node coordinates of the virtual computer room and the real computer room.

Images