Ray tracing digital twin channel intelligent agent construction method and device, and network equipment

Abstract

Embodiments of the present application provide a ray tracing digital twin channel agent construction method, device and network equipment, the method comprising: determining task knowledge and historical conversation information according to a received user task request; determining a task field according to the task request, the task knowledge and the historical conversation information; constructing a structured task according to the task field; selecting a target simulation tool from a simulation tool chain according to a target task type and output requirement of the structured task; inputting simulation input parameters in the structured task into the target simulation tool, simulating an electromagnetic propagation path, and obtaining a simulation result. The ray tracing digital twin channel agent construction method of the embodiments of the present application solves the problem that the simulation of the propagation path and the channel in the related art is seriously dependent on manual analysis and operation.

Description

Technical Field

[0001] This application relates to the field of wireless communication technology, specifically to a method, apparatus, and network device for constructing a ray-tracing digital twin channel agent. Background Technology

[0002] Existing ray tracing (RT) technology can complete propagation path search and channel result generation under given scenario and simulation parameters, but its task configuration, result interpretation, and continuous execution still mainly rely on professional personnel. For tasks such as area coverage analysis, weak coverage diagnosis, and antenna parameter adjustment, user requirements must be manually converted into detailed RT parameters, and simulation results must also be manually analyzed to determine subsequent operations, resulting in low automation. Furthermore, in multi-round tasks, historical configurations and results still need to be manually maintained, easily leading to repetitive operations and inconsistencies between previous and subsequent configurations. Therefore, existing RT systems lack an interactive processing mechanism that unifies task requests, physical simulation, result evaluation, and cross-round state management, making it difficult to meet the continuous task processing requirements of Digital Twin Channel (DTC). Summary of the Invention

[0003] At least one embodiment of this application provides a method, apparatus, and network device for constructing a ray-tracing digital twin channel agent, which addresses the problem that the simulation of propagation paths and channels in the prior art heavily relies on manual analysis and operation.

[0004] To solve the above-mentioned technical problems, this application is implemented as follows:

[0005] In a first aspect, embodiments of this application provide a method for constructing a ray-tracing digital twin channel agent, including:

[0006] Based on the received task request from the user, task knowledge and historical session information are determined; the task request is related to the simulation content of the electromagnetic propagation path.

[0007] Based on the task request, the task knowledge, and the historical session information, the task fields are determined; the task fields include at least one of the following: task type, propagation scenario, carrier frequency, transmitter configuration, receiving area or target receiving point, receiving grid, propagation mechanism and interaction order, antenna configuration, and output result type;

[0008] Construct a structured task based on the task fields;

[0009] Based on the target task type and output requirements of the structured task, a target simulation tool is selected from the simulation toolchain; the simulation toolchain stores simulation tools corresponding to the target task type.

[0010] The simulation input parameters in the structured task are input into the target simulation tool to simulate the electromagnetic propagation path and obtain the simulation results; the simulation input parameters are determined according to the task fields.

[0011] In some embodiments, the method further includes:

[0012] Based on the target task type and the simulation results, the task evaluation result is determined;

[0013] Based on the task evaluation results and the historical session information, the target operation is executed; the target operation includes the following: ending the current task, adjusting the propagation configuration, refining the receiver grid, selecting the target receiver point for multipath diagnosis, generating candidate antenna configurations, performing candidate configuration evaluation, and outputting a session summary.

[0014] In some embodiments, the method further includes:

[0015] Send the first data to the user;

[0016] Save the second data and update the third data to the session state;

[0017] The first data includes at least one of the following: the simulation results, the task evaluation results, the session summary, the propagation diagnosis, the candidate configuration comparison results, and the recommended configuration;

[0018] The second data includes at least one of the following: the user request, the target task type, the structured task, the simulation results, and the task evaluation results;

[0019] The third data includes: the propagation scenario corresponding to the task request and the RT configuration.

[0020] In some embodiments, task knowledge and historical session information are determined based on the received task request from the user, including:

[0021] Determine the first session identifier based on the task request;

[0022] Based on the first session identifier, determine the historical session information;

[0023] The task request includes at least one of the following: area coverage analysis, propagation configuration adjustment, receiver grid refinement, target point propagation diagnosis, antenna configuration adjustment, and historical data;

[0024] The historical session information includes at least one of the following: propagation scenarios used, RT configuration, historical simulation results, historical task evaluation results, and user feedback;

[0025] The task knowledge includes at least one of the following: task type, task fields, field constraints, RT terminology, tool description, and the correspondence between the task type and the simulation tool related to the task request.

[0026] In some embodiments, the task field is determined based on the task request, the task knowledge, and the historical session information, including at least one of the following:

[0027] If the task request carries a first field, the first field is determined to be the task field;

[0028] If the task request indication is not modified and field inheritance is allowed, the task field is determined based on the historical session information;

[0029] If the fields required for the task corresponding to the task request are insufficient, the task fields are determined based on the task knowledge.

[0030] Specifically, task fields are extracted from the task request; the task fields include: fields required by the task and / or fields that can be inherited.

[0031] In some embodiments, constructing a structured task based on the task field includes:

[0032] The task fields are validated for target attributes; the target attributes include at least one of the following: format, unit, value range, and relationship between fields;

[0033] If the task field passes the validation, the structured task is constructed based on the task field.

[0034] In some embodiments, the task evaluation result includes at least one of the following:

[0035] Regional coverage analysis tasks, target point multipath diagnosis, and antenna candidate configurations;

[0036] If the task evaluation result is an antenna candidate configuration, multiple candidate configurations are generated.

[0037] The structured task is reconstructed based on the multiple sets of candidate configurations to obtain the first structured task;

[0038] Based on the task type and output requirements of the first structured task, select the first simulation tool;

[0039] The simulation input parameters from the first structured task are input into the first simulation tool to obtain the simulation results.

[0040] Secondly, embodiments of this application provide a ray-tracing digital twin channel agent construction apparatus, comprising:

[0041] The first determining module is used to determine task knowledge and historical session information based on the received task request from the user; the task request includes at least one of the following: area coverage analysis, propagation configuration adjustment, receiving grid refinement, target point propagation diagnosis, antenna configuration adjustment, and historical data;

[0042] The second determining module is used to determine the task fields based on the task request, the task knowledge, and the historical session information; the task fields include at least one of the following: task type, propagation scenario, carrier frequency, transmitter configuration, receiving area or target receiving point, receiving grid, propagation mechanism and interaction order, antenna configuration, and output result type;

[0043] A construction module is used to construct a structured task based on the task fields;

[0044] The selection module is used to select a target simulation tool from the simulation toolchain according to the target task type and output requirements of the structured task; the simulation toolchain stores simulation tools corresponding to the target task type.

[0045] The simulation module is used to input the simulation input parameters from the structured task into the target simulation tool to simulate the electromagnetic propagation path and obtain simulation results; the simulation input parameters are determined according to the task fields.

[0046] Thirdly, embodiments of this application provide a network device, including: a processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the method described in the second aspect.

[0047] Fourthly, embodiments of this application provide a computer-readable storage medium storing a program that, when executed by a processor, implements the steps of the method described in the first aspect.

[0048] Fifthly, embodiments of this application provide a computer program product, including computer instructions that, when executed by a processor, implement the steps of the method described in the first aspect.

[0049] Compared with existing technologies, the ray-tracing digital twin channel agent construction method, apparatus, and network device provided in this application, by using the determined task knowledge and historical session information corresponding to the user's task request, can determine the task field based on the task request, the task knowledge, and the historical session information. This allows for the automatic construction of a structured RT task recognizable by the RT tool based on the task field, thereby selecting a target RT simulation tool and constructing a ray-tracing digital twin channel agent for the task corresponding to the task request using the target simulation tool. This solves the problem in related technologies where the simulation of propagation paths and channels heavily relies on manual analysis and operation. Attached Figure Description

[0050] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0051] Figure 1 This is a schematic diagram illustrating the steps of the ray-tracing digital twin channel agent construction method according to an embodiment of this application;

[0052] Figure 2 A schematic diagram illustrating the structure of the ray-tracing digital twin channel agent construction method according to embodiments of this application;

[0053] Figure 3 This is a schematic diagram of the structure of the ray-tracing digital twin channel agent construction device according to an embodiment of this application;

[0054] Figure 4 This is a schematic diagram of the structure of a network device according to an embodiment of this application. Detailed Implementation

[0055] The terms "first," "second," etc., used in this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such terms can be used interchangeably where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first" and "second" are generally of the same class, without limiting the number of objects; for example, the first object can be one or more. Furthermore, "or" in this application indicates at least one of the connected objects. For example, "A or B" covers three scenarios: Scenario 1: including A but not B; Scenario 2: including B but not A; Scenario 3: including both A and B. The character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0056] The term "instruction" in this application can be either a direct instruction (or explicit instruction) or an indirect instruction (or implicit instruction). A direct instruction can be understood as one in which the sender explicitly informs the receiver of specific information, the operation to be performed, or the requested result, etc.; an indirect instruction can be understood as one in which the receiver determines the corresponding information based on the instruction sent by the sender, or makes a judgment and determines the operation to be performed or the requested result, etc., based on the judgment result.

[0057] To enable those skilled in the art to better understand the embodiments of this application, the following description is provided first:

[0058] As 6G networks continue to evolve, the propagation environment, frequency range, antenna type, network deployment, and service requirements of communication systems are becoming increasingly complex and dynamic. Network digital twins, by establishing a corresponding representation of the physical network and its operational status in digital space, provide a virtual verification environment for coverage assessment, propagation analysis, and parameter configuration. Specifically, the Digital Transmission Control (DTC) is used to characterize the wireless propagation process and channel state in specific scenarios, providing scenario-related channel information for tasks such as regional coverage assessment, weak coverage location, propagation condition diagnosis, and network parameter adjustment.

[0059] RT (Rapid Transmission) is a deterministic wireless channel modeling method based on a three-dimensional propagation environment and electromagnetic propagation laws. Applying RT requires constructing a three-dimensional scene including environmental objects such as buildings, roads, and scatterers, and configuring input information such as material parameters, carrier frequency, transmitter, receiving area or receiving point, receiving grid, antenna parameters, and propagation mechanism. The RT solver searches for propagation paths such as direct, reflected, diffracted, penetrated, and scattered propagation paths based on geometric optics, geometric diffraction theory, and electromagnetic boundary conditions, and calculates the power, phase, and time delay of multipath components, thereby generating results such as received power, path loss, channel impulse response, propagation path map, and coverage heatmap.

[0060] Existing RT solvers are capable of performing relatively mature propagation path searches and electromagnetic calculations, but their use still largely relies on professionals for task configuration, result interpretation, and workflow organization. For tasks such as regional coverage analysis, propagation configuration refinement, weak coverage point diagnosis, and antenna parameter adjustment, user requirements typically need to be manually converted into specific RT input parameters. After simulation, coverage maps, multipath paths, and statistical indicators also require manual analysis and conversion into conclusions for network planning. In continuous task processing, the coverage results, propagation configurations, diagnostic points, and candidate parameters from the previous round need to be used as input for the next round of tasks. The lack of a unified mechanism for related state maintenance and parameter inheritance easily leads to problems such as duplicate configurations, missing fields, inconsistencies in parameters between rounds, and difficulty in tracing results.

[0061] Large Language Models (LLMs) possess the capabilities of natural language understanding and structured content generation, while Artificial Intelligence (AI) can invoke external tools based on task objectives and organize subsequent operations using the execution results. Introducing these capabilities into Real-Time (RT) task processing allows for the establishment of a task parsing and orchestration mechanism oriented towards RT professional fields, tool interfaces, and continuous task states, without replacing the physical RT solver. This enables user requests to be transformed into executable, structured RT tasks, and allows physical simulation results to generate coverage evaluations, multipath diagnostics, and candidate configuration evaluations. Furthermore, it allows for the saving and inheritance of scenarios, configurations, and historical results across multiple task rounds.

[0062] To address the problems in related technologies, such as the difficulty in directly translating user task requests into RT solver inputs, the inability of RT outputs to directly support task evaluation, and the reliance on manual maintenance for scenario configuration, simulation parameters, and historical results in continuous RT tasks, this application provides a method for constructing a ray-tracing digital twin channel agent. This method can build an interactive RT agent oriented towards DTC (Digital Tracing Control). The agent establishes a unified organizational mechanism among user requests, structured RT tasks, physical RT simulation results, and cross-round session states, thereby solving the technical problems of complex RT task configuration, reliance on manual result analysis, and lack of continuity and traceability in multi-round execution.

[0063] This application provides a method and apparatus for constructing a ray-tracing digital twin channel agent. The method and apparatus are based on the same concept, and since they address similar problems in principle, their implementations can be mutually referenced; repeated details will not be elaborated further.

[0064] like Figure 1 As shown in the figure, an embodiment of this application provides a method for constructing a ray-tracing digital twin channel agent, which includes the following steps:

[0065] Step 101: Based on the received task request from the user, determine the task knowledge and historical session information; the task request is related to the simulation content of the electromagnetic propagation path.

[0066] Step 102: Determine the task fields based on the task request, the task knowledge, and the historical session information; the task fields include at least one of the following: task type, propagation scenario, carrier frequency, transmitter configuration, receiving area or target receiving point, receiving grid, propagation mechanism and interaction order, antenna configuration, and output result type;

[0067] Step 103: Construct a structured task based on the task fields;

[0068] Step 104: Select a target simulation tool from the simulation toolchain according to the target task type and output requirements of the structured task; the simulation toolchain stores simulation tools corresponding to the target task type.

[0069] Step 105: Input the simulation input parameters from the structured task into the target simulation tool to simulate the electromagnetic propagation path and obtain the simulation results; the simulation input parameters are determined according to the task fields.

[0070] The ray-tracing digital twin channel agent construction method of this application, by using the determined task knowledge and historical session information corresponding to the user's task request, can determine the task field based on the task request, the task knowledge, and the historical session information. This allows for the automatic construction of a structured RT task recognizable by the RT tool, selection of a target RT simulation tool, and construction of a ray-tracing digital twin channel agent for the task corresponding to the task request using the target simulation tool. This solves the problem in related technologies where the simulation of propagation paths and channels heavily relies on manual analysis and operation.

[0071] Optionally, determining task knowledge and historical session information based on the received user task request includes:

[0072] Determine the first session identifier based on the task request;

[0073] Based on the first session identifier, determine the historical session information;

[0074] The task request includes at least one of the following: area coverage analysis, propagation configuration adjustment, receiver grid refinement, target point propagation diagnosis, antenna configuration adjustment, and historical data;

[0075] The historical session information includes at least one of the following: propagation scenarios used, RT configuration, historical simulation results, historical task evaluation results, and user feedback;

[0076] The task knowledge includes at least one of the following: task type, task fields, field constraints, RT terminology, tool description, and the correspondence between the task type and the simulation tool related to the task request.

[0077] It should be noted that the task request mentioned is a task request for an RT task.

[0078] In this embodiment of the application, after receiving the user's task request, the task knowledge is determined based on the task request;

[0079] Read historical session information based on the current session identifier;

[0080] From the task knowledge, obtain the RT task type, task fields, field constraints, RT terminology, tool description, and the correspondence between task types and tools related to the current request;

[0081] The parameters such as coordinates, frequency, distance, power, propagation mechanism name, and antenna parameter name in the task request are standardized to ensure consistency with the information in the task fields.

[0082] The ray-tracing digital twin channel agent construction in this application embodiment, by obtaining task knowledge and historical session information from the task request, can obtain the corresponding task fields from the task knowledge and historical session information even when the task fields in the task request are incomplete, thereby improving the simulation success rate.

[0083] Optionally, based on the task request, the task knowledge, and the historical session information, a task field is determined, including at least one of the following:

[0084] If the task request carries a first field, the first field is determined to be the task field;

[0085] If the task request indication is not modified and field inheritance is allowed, the task field is determined based on the historical session information;

[0086] If the fields required for the task corresponding to the task request are insufficient, the task fields are determined based on the task knowledge.

[0087] Specifically, task fields are extracted from the task request; the task fields include: fields required by the task and / or fields that can be inherited.

[0088] It should be noted that the task request, the task knowledge, and the historical session information are identified through the LLM large model to determine the task type and task objective corresponding to the task request;

[0089] Based on the RT task type, determine the fields required for the current task, the fields that can be inherited, the RT tools to be called, and the expected output results.

[0090] Optionally, the method further includes:

[0091] When the task request contains multiple task objectives, the request is split into multiple executable subtasks;

[0092] Determine the execution order and dependencies between the subtasks;

[0093] The subtasks are executed according to the execution order and dependencies.

[0094] The construction of the ray-tracing digital twin channel agent in this application embodiment can automatically execute multiple task objectives without human intervention by splitting multiple task objectives into multiple sub-tasks with a specific execution order.

[0095] Optionally, a structured task is constructed based on the task fields, including:

[0096] The task fields are validated for target attributes; the target attributes include at least one of the following: format, unit, value range, and relationship between fields;

[0097] If the task field passes the validation, the structured task is constructed based on the task field.

[0098] In this embodiment of the application, for fields explicitly given in the task request, the field values ​​in the task request are used;

[0099] For fields that are not modified in the current request and are allowed to be inherited, read the corresponding field values ​​from the historical session information;

[0100] For fields required for the current task but not yet determined, complete them according to the preset rules in the task knowledge; if they cannot be completed, send a supplementary question to the user.

[0101] Optionally, the method further includes:

[0102] When there is a conflict between the task request and the historical session information regarding the same field, the field value explicitly given in the task request shall be adopted, or the field in the historical session information shall be updated after user confirmation.

[0103] Optionally, constructing the structured task based on the task fields includes:

[0104] The validated task fields are organized into structured RT tasks and converted into call parameters that can be recognized by RT tools.

[0105] The construction of the ray-tracing digital twin channel agent in this application embodiment, by constructing the structured task, enables the task to be transformed into a task that can be recognized by the RT tool, thereby realizing the automation of the construction of the ray-tracing digital twin channel agent.

[0106] Optionally, the simulation toolchain includes at least one of the following: a region coverage simulation tool, a target point multipath simulation tool, a candidate configuration batch simulation tool, or a result aggregation tool;

[0107] Based on the target task type and output requirements of the structured task, a simulation tool matching the task type of the structured task is selected from the simulation toolchain.

[0108] Optionally, the simulation input parameters from the structured task are input into the target simulation tool to simulate the electromagnetic propagation path and obtain simulation results, including:

[0109] Write the propagation scenario, carrier frequency, transmitter, receiving area or target receiving point, receiving grid, propagation mechanism, interaction order, and antenna configuration from the structured RT task into the calling interface of the selected RT tool.

[0110] Call the physical RT solver to perform propagation path search and electromagnetic calculations to obtain one or more of the following simulation results: coverage heatmap, received power, path loss, multipath propagation path, propagation delay, and candidate configuration.

[0111] The LLM central controller is used for task parsing, task generation, and tool scheduling; it does not directly generate received power, propagation path, or channel results.

[0112] The construction of the ray-tracing digital twin channel agent in this application embodiment improves the automation level of the ray-tracing digital twin channel agent construction by pre-building a simulation toolchain and automatically selecting the corresponding target simulation tool according to the target task type and output requirements of the structured task.

[0113] Optionally, the task evaluation result includes at least one of the following:

[0114] Regional coverage analysis tasks, target point multipath diagnosis, and antenna candidate configurations;

[0115] If the task evaluation result is an antenna candidate configuration, multiple candidate configurations are generated.

[0116] The structured task is reconstructed based on the multiple sets of candidate configurations to obtain the first structured task;

[0117] Based on the task type and output requirements of the first structured task, select the first simulation tool;

[0118] The simulation input parameters from the first structured task are input into the first simulation tool to obtain the simulation results.

[0119] Optionally, the method further includes:

[0120] Based on the target task type and the simulation results, the task evaluation result is determined;

[0121] Based on the task evaluation results and the historical session information, the target operation is executed; the target operation includes the following: ending the current task, adjusting the propagation configuration, refining the receiver grid, selecting the target receiver point for multipath diagnosis, generating candidate antenna configurations, performing candidate configuration evaluation, and outputting a session summary.

[0122] In this embodiment of the application, evaluation information corresponding to the task objective is extracted from the simulation results (RT output);

[0123] For regional coverage analysis tasks, extract one or more of the following: coverage heatmap, average received power, weak coverage ratio, weak coverage area and priority diagnostic point;

[0124] For target point multipath diagnosis tasks, extract one or more of the following: line-of-sight or non-line-of-sight propagation status, number of multipaths, total received power, main propagation path and propagation delay;

[0125] For the antenna candidate configuration evaluation task, extract one or more of the following for each candidate configuration: azimuth angle, downtilt angle, average received power, and weak coverage ratio.

[0126] The evaluation information is organized into task evaluation results and returned to the LLM central controller;

[0127] The LLM central controller determines the subsequent operation (the target operation) based on the current task objective, task evaluation results, and historical session information.

[0128] It should be noted that when the target operation is candidate configuration evaluation, multiple sets of candidate configurations are generated according to a preset parameter range or user requirements, and the structured task is reconstructed according to the multiple sets of candidate configurations to obtain the first structured task.

[0129] Based on the task type and output requirements of the first structured task, select the first simulation tool;

[0130] The simulation input parameters from the first structured task are input into the first simulation tool to obtain the simulation results.

[0131] If the current task results do not meet the task requirements, rebuild the RT structured task; reselect the target simulation tool, and then re-perform the simulation.

[0132] Optionally, the method further includes:

[0133] Send the first data to the user;

[0134] Save the second data and update the third data to the session state;

[0135] The first data includes at least one of the following: the simulation results, the task evaluation results, the session summary, the propagation diagnosis, the candidate configuration comparison results, and the recommended configuration;

[0136] The second data includes at least one of the following: the user request, the target task type, the structured task, the simulation results, and the task evaluation results;

[0137] The third data includes: the propagation scenario corresponding to the task request and the RT configuration.

[0138] The ray-tracing digital twin channel agent construction method of this application, by saving the first data and the second data, can obtain the required data information as historical session information when encountering similar tasks, structured tasks, simulation results, propagation scenarios or RT configurations in the future.

[0139] In this embodiment of the application, the implementation of the ray-tracing digital twin channel intelligent agent construction method relies on an intelligent agent, and the intelligent agent possesses the capabilities included in the aforementioned ray-tracing digital twin channel intelligent agent construction method. For example... Figure 2 As shown, the intelligent agent includes multi-round user interaction, an LLM central controller, a RT simulation toolchain, a task evaluation module based on RT results, and a task knowledge and session memory module.

[0140] Specifically, the LLM central controller combines task requests, task knowledge, and historical session information to generate structured tasks that the RT solver can execute; the RT simulation toolchain completes 3D scene preparation, propagation path search, electromagnetic calculation, and result generation based on the structured tasks; the task evaluation module extracts evaluation information corresponding to the current task from the coverage heatmap, multipath propagation paths, and numerical results; and the task knowledge and session memory module saves RT task rules, tool descriptions, historical configurations, and execution results.

[0141] Therefore, this application establishes a continuous task processing procedure encompassing user task request acquisition, RT task parsing, structured task generation, physical RT calculation, result evaluation, subsequent task determination, and session state update. The LLM central controller does not directly generate received power, propagation path, or channel results; these physical results are calculated and returned by the RT simulation toolchain according to the structured task.

[0142] In one embodiment of this application, the interactive RT agent for DTC is applied to a campus RT scenario, and the method for constructing the ray-tracing digital twin channel agent is specifically as follows:

[0143] First, build the implementation environment and initialize the campus RT scenario:

[0144] This phase requires establishing the execution environment between task requests, structured RT tasks, and physical RT tools, and initializing the campus RT scenario and basic simulation parameters, specifically including:

[0145] 1) Establish a prototype system consisting of a user interface, a Python task handler, an LLM (Limited Language Management), and a RT (Real-Time) toolchain (e.g., the MATLAB RT toolchain). The user interface receives task requests, displays RT parameters, and presents task results; the Python task handler manages sessions, compiles tasks, invokes tools, and organizes results; the LLM combines task knowledge and historical session information to generate structured RT tasks; and the MATLAB RT toolchain handles scene loading, propagation path searching, electromagnetic calculations, and result visualization.

[0146] 2) Configure six types of task modes in the task knowledge: area coverage analysis, propagation configuration refinement, receiver grid refinement, target point multipath diagnosis, antenna candidate configuration evaluation, and session summary. Configure the required fields, inheritable fields, field constraints, corresponding tools, and expected outputs for each task mode.

[0147] 3) Establish a data interface between the Python task handler and the RT toolchain. The Python task handler converts structured RT tasks into calling parameters for area coverage simulation functions, target point multipath diagnostic functions, or candidate configuration batch simulation functions; the RT toolchain performs physical RT calculations and returns received power arrays, multipath propagation paths, statistical indicators, and graphical results, and the Python task handler organizes the returned content into corresponding task result objects.

[0148] 4) Obtain the building outlines and basic geographic information of the Beijing University of Posts and Telecommunications campus area (e.g., from open-source map data), and process the building geometry, building height, material properties and thickness information based on the building outlines and basic geographic information, and import the processed 3D scene into the RT toolchain.

[0149] 5) Set fixed transmitter coordinates, for example, transmitter coordinates are (0, −10, 50) m, receiver height is 1.5 m, carrier frequency is 3.5 GHz, and available power at the transmit antenna input is 1 W. The weak coverage threshold is denoted as Tcov. In this embodiment, Tcov = −110 dBm. Receiver points whose received power calculated by the RT solver is lower than Tcov are determined to be weak coverage points.

[0150] 6) Set the initial RT configuration to an omnidirectional transmitting antenna, with the maximum interaction order of reflection, diffraction and penetration being R=0, D=1 and P=0 respectively, and the initial receiving grid size being 40 m × 30 m.

[0151] 7) Write the propagation scenario, carrier frequency, transmitter configuration, receiving area, receiving grid, propagation configuration, antenna configuration, and weak coverage threshold into the current session state as the base fields for subsequent tasks to inherit.

[0152] Secondly, perform the initial region coverage analysis task:

[0153] During this phase, the system generates and executes regional coverage RT tasks based on the user's campus coverage analysis task request, specifically including:

[0154] 1) Obtain the campus area coverage analysis request submitted by the user and identify it as an area coverage analysis task.

[0155] 2) Read the campus scene, 3.5 GHz carrier frequency, transmitter configuration, receiver area, 40 m × 30 m receiver grid, R=0, D=1, P=0 propagation configuration and omnidirectional antenna configuration from the current session state.

[0156] 3) Generate a structured RT task for regional coverage analysis based on the above fields, and convert it into calling parameters for the regional coverage RT tool.

[0157] 4) Call the RT tool to perform regional propagation calculations, obtain the received power of each receiving grid point, and generate an initial coverage heatmap.

[0158] 5) Based on the -110 dBm weak coverage threshold, weak coverage areas were extracted, and the average received power was -86.78 dBm, with a weak coverage ratio of 42.3%.

[0159] 6) Write the initial coverage heatmap, average received power, weak coverage ratio, weak coverage area and corresponding RT configuration into the session state.

[0160] Next, refine the propagation configuration and receiving grid based on the coverage results:

[0161] In this phase, the system continuously performs propagation configuration refinement and receiver grid refinement based on the initial coverage results, enabling subsequent point diagnostics to inherit the updated RT configuration and refined coverage results.

[0162] The detailed propagation configuration includes:

[0163] 1) Obtain the user's task request for refining the propagation mechanism and identify it as a propagation configuration refinement task.

[0164] 2) Inherit the campus scene, carrier frequency, transmitter, receiving area, 40 m × 30 m receiving grid and omnidirectional antenna configuration, only update the propagation configuration from R=0, D=1, P=0 to R=2, D=1, P=1.

[0165] 3) Generate a structured RT task based on the updated propagation configuration, and call the region coverage RT tool to re-execute the region propagation calculation.

[0166] 4) The updated coverage heatmap was obtained, with an average received power of −84.47 dBm and a weak coverage ratio of 38.9%.

[0167] 5) Write the propagation configuration of R=2, D=1, P=1 and its corresponding coverage results into the session state.

[0168] Receiver grid refinement specifically includes:

[0169] 1) Obtain the user's request for receiving mesh refinement and identify it as a receiving mesh refinement task.

[0170] 2) Inherit the campus scene, 3.5 GHz carrier frequency, transmitter, receiver area, R=2, D=1, P=1 propagation configuration and omnidirectional antenna configuration, only updating the receiver grid size from 40 m×30 m to 5 m×5 m.

[0171] 3) Generate a structured RT task based on the updated receiver grid, and call the regional coverage RT tool to re-execute the coverage calculation.

[0172] 4) Based on the received power of each receiving point under the refined grid and the weak coverage threshold of −110dBm, weak coverage points are extracted, and the point with coordinates (−160.0, −380.1, 1.5)m is determined as the priority diagnostic point.

[0173] 5) Write the refined receive grid, refined coverage heatmap, weak coverage point set, and priority diagnostic points into the session state.

[0174] Next, perform the target point multipath diagnosis task:

[0175] During this phase, the system switches the area coverage task to a target point multipath diagnosis task to analyze the propagation conditions of priority diagnosis points, specifically including:

[0176] 1) Read the priority diagnostic point (−160.0, −380.1, 1.5) m from the session state, and select a reference point with coordinates (140.0, 99.9, 1.5) m from the well-covered area for comparison analysis.

[0177] 2) Inherit the campus scene, 3.5 GHz carrier frequency, transmitter configuration and R=2, D=1, P=1 propagation configuration, set the target receiver point field as the priority diagnostic point and reference point respectively, and generate a structured RT task for target point multipath diagnosis.

[0178] 3) Call the target point multipath RT tool to search for effective propagation paths from the transmitter to the target receiver, and calculate the power, path delay and total received power of each path.

[0179] 4) For the priority diagnostic point, the non-line-of-sight propagation state, two effective propagation paths, and a total received power of −157.66 dBm are obtained.

[0180] 5) For the reference point, the line-of-sight propagation status, 25 effective propagation paths, and -53.71 dBm total received power are obtained.

[0181] 6) Based on the line-of-sight status, number of effective propagation paths, and total received power of the two receiving points, a propagation diagnosis result is formed, and it is determined that the priority diagnosis point has local building obstruction and a small number of available propagation paths.

[0182] 7) Write the target receiver coordinates, point-level RT configuration, multipath propagation path, numerical results, and propagation diagnostic results into the session state.

[0183] Then, candidate antenna configurations are generated and batch RT evaluation is performed:

[0184] In this phase, the system generates candidate configurations based on the antenna configuration adjustment task request, and evaluates the coverage effect of different candidate configurations through batch RT simulation, specifically including:

[0185] 1) Obtain the directional antenna configuration adjustment request submitted by the user and identify it as an antenna candidate configuration evaluation task.

[0186] 2) Inherit the campus scene, 3.5 GHz carrier frequency, 1W transmit antenna input available power, R=2, D=1, P=1 propagation configuration, receive area and receive grid, so that the RT fields are consistent among the candidate configurations except for the antenna azimuth angle and downtilt angle.

[0187] 3) Replace the transmitting antenna with a directional antenna with a beamwidth of 90°.

[0188] 4) Generate 33 sets of azimuth-downtilt candidate configurations within the preset parameter range, and assign a unique number to each set of candidate configurations.

[0189] 5) For each candidate configuration, write its azimuth and downtilt angle into the structured RT task, and call the regional coverage RT tool to perform coverage calculation.

[0190] 6) Calculate the average received power and weak coverage ratio for the RT output of each candidate configuration to form a candidate configuration evaluation record.

[0191] 7) Based on the evaluation principle of higher average received power and lower weak coverage ratio, 33 candidate configurations were compared.

[0192] 8) Configuration number 27 is determined to be the recommended configuration, with an azimuth angle of 92°, a downtilt angle of 8°, an average received power of −73.04dBm, and a weak coverage ratio of 14.83%.

[0193] 9) Write all candidate configurations, the corresponding coverage metrics for each candidate configuration, the recommended candidate number, and the recommended parameters into the session state.

[0194] Finally, organize the historical task results and output a session summary:

[0195] In this phase, the system organizes the results of the aforementioned tasks according to the session state and outputs a user-facing session summary, specifically including:

[0196] 1) Read historical RT fields such as campus scene, carrier frequency, transmitter, receiving area, propagation configuration, receiving grid and antenna configuration based on the session identifier.

[0197] 2) Read the initial coverage results, propagation configuration refinement results, receive grid refinement results, target point multipath diagnosis results, and antenna candidate configuration evaluation results.

[0198] 3) Organize historical task records according to the execution order of “initial area coverage analysis - propagation configuration refinement - receiver grid refinement - target point multipath diagnosis - antenna candidate configuration evaluation”.

[0199] 4) Attach the corresponding user's task request, structured RT task, actual call parameters, raw RT output and task evaluation results to each task result.

[0200] 5) Output task summary, coverage map, multipath map, candidate configuration evaluation trajectory and recommended parameters through user interface.

[0201] 6) Continue to save the organized session state; when the user makes a new scenario analysis or parameter adjustment request, read the session state and re-enter the overall implementation process.

[0202] The various methods of the embodiments of this application have been described above. Apparatus for implementing the above methods will now be provided.

[0203] like Figure 3 As shown in the figure, this application embodiment also provides a ray-tracing digital twin channel intelligent agent construction device 300, including:

[0204] The first determining module 301 is used to determine task knowledge and historical session information based on the received task request from the user; the task request includes at least one of the following: area coverage analysis, propagation configuration adjustment, receiving grid refinement, target point propagation diagnosis, antenna configuration adjustment, and historical data;

[0205] The second determining module 302 is used to determine the task field based on the task request, the task knowledge, and the historical session information; the task field includes at least one of the following: task type, propagation scenario, carrier frequency, transmitter configuration, receiving area or target receiving point, receiving grid, propagation mechanism and interaction order, antenna configuration, and output result type;

[0206] Module 303 is used to construct a structured task based on the task fields;

[0207] Module 304 selects a target simulation tool from the simulation toolchain based on the target task type and output requirements of the structured task; the simulation toolchain stores simulation tools corresponding to the target task type.

[0208] The simulation module 305 is used to input the simulation input parameters from the structured task into the target simulation tool to simulate the electromagnetic propagation path and obtain simulation results; the simulation input parameters are determined according to the task fields.

[0209] The ray-tracing digital twin channel agent construction device of this application, by using the determined task knowledge and historical session information corresponding to the user's task request, can determine the task field based on the task request, the task knowledge, and the historical session information. This allows for the automatic construction of a structured RT task recognizable by the RT tool, selection of a target RT simulation tool, and construction of a ray-tracing digital twin channel agent for the task corresponding to the task request using the target simulation tool. This solves the problem in related technologies where the simulation of propagation paths and channels heavily relies on manual analysis and operation.

[0210] This application also provides a network device, such as... Figure 4 As shown, it includes a transceiver 410, a processor 400, a memory 420, and a program or instructions stored in the memory 420 and executable on the processor 400; when the processor 400 executes the program or instructions, it implements the various processes of the above-described embodiment of the ray-tracing digital twin channel agent construction method and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0211] The transceiver 410 is used to receive and send data under the control of the processor 400.

[0212] Among them, Figure 4In this context, the bus architecture may include any number of interconnected buses and bridges, specifically linking various circuits together, represented by one or more processors (processor 400) and memory (memory 420). The bus architecture may also link together various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. The bus interface provides an interface. The transceiver 410 may be multiple elements, including transmitters and receivers, providing a unit for communicating with various other devices over a transmission medium. The processor 400 is responsible for managing the bus architecture and general processing, and the memory 420 may store data used by the processor 400 during operation.

[0213] This application also provides a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the various processes of the above-described ray-tracing digital twin channel intelligent agent construction method embodiment, achieving the same technical effects. To avoid repetition, it will not be described again here. The computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc.

[0214] This application also provides a computer program product, including computer instructions. When the computer instructions are executed by a processor, they implement the various processes of the above-described ray-tracing digital twin channel agent construction method embodiment and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0215] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0216] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0217] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A method for constructing a ray-tracing digital twin channel intelligent agent, characterized in that, include: Based on the received task request from the user, determine the task knowledge and historical session information; The task request is related to the simulation content of the electromagnetic propagation path; Based on the task request, the task knowledge, and the historical session information, the task fields are determined; the task fields include at least one of the following: task type, propagation scenario, carrier frequency, transmitter configuration, receiving area or target receiving point, receiving grid, propagation mechanism and interaction order, antenna configuration, and output result type; Construct a structured task based on the task fields; Based on the target task type and output requirements of the structured task, a target simulation tool is selected from the simulation toolchain; the simulation toolchain stores simulation tools corresponding to the target task type. The simulation input parameters in the structured task are input into the target simulation tool to simulate the electromagnetic propagation path and obtain the simulation results; the simulation input parameters are determined according to the task fields.

2. The method according to claim 1, characterized in that, The method further includes: Based on the target task type and the simulation results, the task evaluation result is determined; Based on the task evaluation results and the historical session information, the target operation is executed; the target operation includes the following: ending the current task, adjusting the propagation configuration, refining the receiver grid, selecting the target receiver point for multipath diagnosis, generating candidate antenna configurations, performing candidate configuration evaluation, and outputting a session summary.

3. The method according to claim 2, characterized in that, The method further includes: Send the first data to the user; Save the second data and update the third data to the session state; The first data includes at least one of the following: the simulation results, the task evaluation results, the session summary, the propagation diagnosis, the candidate configuration comparison results, and the recommended configuration; The second data includes at least one of the following: the user request, the target task type, the structured task, the simulation results, and the task evaluation results; The third data includes: the propagation scenario corresponding to the task request and the RT configuration.

4. The method according to claim 1, characterized in that, Based on the received task request from the user, determine the task knowledge and historical session information, including: Determine the first session identifier based on the task request; Based on the first session identifier, determine the historical session information; The task request includes at least one of the following: area coverage analysis, propagation configuration adjustment, receiver grid refinement, target point propagation diagnosis, antenna configuration adjustment, and historical data; The historical session information includes at least one of the following: propagation scenarios used, RT configuration, historical simulation results, historical task evaluation results, and user feedback; The task knowledge includes at least one of the following: task type, task fields, field constraints, RT terminology, tool description, and the correspondence between the task type and the simulation tool related to the task request.

5. The method according to claim 1, characterized in that, Based on the task request, the task knowledge, and the historical session information, the task fields are determined, including at least one of the following: If the task request carries a first field, the first field is determined to be the task field; If the task request indication is not modified and field inheritance is allowed, the task field is determined based on the historical session information; If the fields required for the task corresponding to the task request are insufficient, the task fields are determined based on the task knowledge. Specifically, task fields are extracted from the task request; the task fields include: fields required by the task and / or fields that can be inherited.

6. The method according to claim 1, characterized in that, Based on the task fields, construct a structured task, including: The task fields are validated for target attributes; the target attributes include at least one of the following: format, unit, value range, and relationship between fields; If the task field passes the validation, the structured task is constructed based on the task field.

7. The method according to claim 2, characterized in that, The task evaluation result includes at least one of the following: Regional coverage analysis tasks, target point multipath diagnosis, and antenna candidate configurations; If the task evaluation result is an antenna candidate configuration, multiple candidate configurations are generated. The structured task is reconstructed based on the multiple sets of candidate configurations to obtain the first structured task; Based on the task type and output requirements of the first structured task, select the first simulation tool; The simulation input parameters from the first structured task are input into the first simulation tool to obtain the simulation results.

8. A device for constructing a ray-tracing digital twin channel intelligent agent, characterized in that, include: The first determining module is used to determine task knowledge and historical session information based on the received task request from the user. The task request includes at least one of the following: area coverage analysis, propagation configuration adjustment, receiver grid refinement, target point propagation diagnosis, antenna configuration adjustment, and historical data; The second determining module is used to determine the task fields based on the task request, the task knowledge, and the historical session information; the task fields include at least one of the following: task type, propagation scenario, carrier frequency, transmitter configuration, receiving area or target receiving point, receiving grid, propagation mechanism and interaction order, antenna configuration, and output result type; A construction module is used to construct a structured task based on the task fields; The selection module is used to select a target simulation tool from the simulation toolchain according to the target task type and output requirements of the structured task; the simulation toolchain stores simulation tools corresponding to the target task type. The simulation module is used to input the simulation input parameters from the structured task into the target simulation tool to simulate the electromagnetic propagation path and obtain simulation results; the simulation input parameters are determined according to the task fields.

9. A network device, characterized in that, include: Transceiver, processor, memory, and programs or instructions stored in the memory and executable on the processor; When the processor executes the program or instructions, it implements the steps of the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the method as described in any one of claims 1 to 7.

Images