An intelligent network slice adaptive orchestration method and system for 6G communication

Abstract

The application discloses a kind of intelligent network slice adaptive arrangement methods and systems for 6G communication, it is related to network slice technical field.For the same target slice in the same current period in the same target slice in the intersection area of vehicle-road cooperation, when control message transmission anomaly and target detection and tracking anomaly occur simultaneously, the existing method is difficult to determine the main local resource range of the current anomaly and the resource adjustment range is easy to enlarge problem, the application first unifies the two sides of the anomaly, then constructs the candidate resource domain covering wireless side resource and edge side resource, and compares the two sides of the anomaly associated results in the candidate resource domain, generates local anomaly type label, and then executes resource adjustment range mapping, extension, recovery and adaptive arrangement adjustment, to improve the stability of resource range determination and improve the consistency of range recovery.

Description

Technical Field

[0001] This invention relates to the field of network slicing technology, and more specifically, to an intelligent network slicing adaptive orchestration method and system for 6G communication. Background Technology

[0002] In the application of network slicing technology for 6G communication, different services have significantly different requirements for network resources. Typically, network slices require separate configuration of radio bearer resources, edge computing resources, and scheduling strategies to meet the service requirements of different services. With the further integration of communication and sensing capabilities, the same network slice may simultaneously carry sensing services such as target detection and tracking, in addition to control message transmission. Therefore, the operational status of the same slice is often affected by both radio-side resources and edge computing resources. Existing network slice operation monitoring can usually detect slice operational anomalies and collect operational status information of radio-side resources and edge computing resources separately. Then, based on the results of unilateral anomalies, resource occupancy, or preset empirical rules, the adjustment targets or scope are determined.

[0003] The existing technology has the following shortcomings:

[0004] On the one hand, within the vehicle-road cooperative intersection area, when traffic flow, pedestrian flow, or sudden events simultaneously increase the resource demands for both control message transmission and target detection and tracking tasks within a short period, the same target slice may experience both control message transmission anomalies and target detection and tracking anomalies within the same current period. In this case, the anomaly may primarily correspond to the wireless resource range, the edge computing resource range, or a local range where both factors affect the anomaly. While existing methods can obtain separate results for wireless-side and edge-side anomalies, along with corresponding resource status information, these results primarily rely on unilateral judgment, making it difficult to reliably determine which resource range the current anomaly primarily corresponds to under a unified judgment criterion. Consequently, existing methods lack a basis for reliably attributing bilateral anomalies to the same local resource range, making it difficult to accurately determine the scope of resource adjustments in this round.

[0005] On the other hand, when existing methods primarily rely on unilateral anomaly results or empirical rules to determine the adjustment scope, if the main impact range of the current anomaly is inaccurately determined, problems that originally occurred only locally can easily be amplified into larger-scale resource adjustments. This not only expands the scope of orchestration disturbances but also, after the anomaly is mitigated, the unstable basis for determining the initial scope leads to a lack of consistent logic in subsequent scope recovery. As a result, existing methods struggle to simultaneously ensure timely anomaly handling, controllable resource adjustment scope, and stable recovery after anomaly regression.

[0006] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide an intelligent network slicing adaptive orchestration method and system for 6G communication, in order to solve the problems mentioned in the background art. Summary of the Invention

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] An intelligent network slicing adaptive orchestration method for 6G communication includes:

[0009] S101, extract abnormal information of control message transmission and abnormal information of target detection and tracking in the current cycle of the target slice in the vehicle-road cooperative intersection area, and associate the two with the same target slice, the same current cycle and the same intersection area to form abnormal entry information;

[0010] S102, based on the anomaly entry information, jointly merge and pair wireless resource items and edge computing resource items that jointly serve the anomaly corresponding object to form a candidate resource domain;

[0011] S103, for each candidate resource domain, determine the correlation between control message transmission anomaly information and target detection and tracking anomaly information within that candidate resource domain, and generate local anomaly type markers based on the comparison results;

[0012] S104: Determine the initial resource adjustment range based on the local anomaly type marker, update the initial resource adjustment range, perform adaptive orchestration adjustment on the updated resource adjustment range, and shrink the resource adjustment range when the anomaly falls back.

[0013] In a preferred embodiment, when abnormal entry information is generated, the control-side abnormal range corresponding to the control message transmission abnormal information and the detection and tracking-side abnormal range corresponding to the target detection and tracking abnormal information are recorded; when only one-sided abnormal information is generated in the current period, the corresponding abnormal entry information is still established, and the vacancy status is registered on the other side.

[0014] In a preferred embodiment, when forming a candidate resource domain, the abnormal direct coverage segment is determined based on the abnormal information transmitted by the control message and the abnormal information of target detection and tracking. Starting from the abnormal direct coverage segment, the resource level adjacent to the task chain is extended along the task chain before and after it. Coverage segments with connected or overlapping spatial boundaries and continuous task chain positions are merged to obtain a local spatial range.

[0015] In a preferred embodiment, forming candidate resource domains includes:

[0016] Map the control object corresponding to the control message transmission anomaly information and the target object corresponding to the target detection and tracking anomaly information to the same object domain;

[0017] Wireless resource items that have the same service objects or whose service object overlap ratio reaches a preset object overlap threshold and are located in the same or adjacent task chain positions in the same control and sensing task closed loop are merged to form wireless candidate subdomains.

[0018] Edge computing resource items that have the same service objects or whose service object overlap ratio reaches a preset object overlap threshold and are located in the same or adjacent task chain positions in the same control and perception task closed loop are merged to form edge candidate subdomains.

[0019] Candidate resource domains are formed by pairing wireless candidate subdomains and edge candidate subdomains that jointly serve the mapped object and are located in the same continuous link of the control and perception task closed loop.

[0020] In a preferred embodiment, determining the association status within each candidate resource domain includes:

[0021] The control-side association result is determined based on the current candidate resource domain's capacity to support anomaly-related control objects, abnormal changes in control message transmission, and control anomaly references in adjacent candidate resource domains.

[0022] Based on the service level of the current candidate resource domain to the abnormal target object, the abnormal changes in target detection and tracking, and the abnormal reference of detection and tracking in adjacent candidate resource domains, the correlation results on the detection and tracking side are determined.

[0023] In a preferred embodiment, the local anomaly type labeling includes a control-side priority interest domain, a detection-tracking-side priority interest domain, a co-occurrence interest domain, and a low-interest state. High correlation is directly considered to meet a preset condition; medium correlation is considered to meet a preset condition when it is true for two consecutive current periods; low correlation corresponds to remaining in an observation state. When both the control-side correlation result and the detection-tracking-side correlation result meet the preset condition and are at the same correlation level, the corresponding candidate resource domain is labeled as a co-occurrence interest domain. When the control-side correlation result is higher than the detection-tracking-side correlation result and meets the preset condition, the corresponding candidate resource domain is labeled as a control-side priority interest domain. When the detection-tracking-side correlation result is higher than the control-side correlation result and meets the preset condition, the corresponding candidate resource domain is labeled as a detection-tracking-side priority interest domain. When neither the control-side correlation result nor the detection-tracking-side correlation result meets the preset condition, the corresponding candidate resource domain remains in a low-interest state.

[0024] In a preferred embodiment, determining the initial resource adjustment range includes:

[0025] The wireless candidate subdomains corresponding to the priority domains on the control side are determined as the initial resource adjustment range;

[0026] The edge candidate subdomains corresponding to the priority focus domains on the detection and tracking side are determined as the initial resource adjustment range;

[0027] The combined range of wireless candidate subdomains and edge candidate subdomains corresponding to the common interest domain is determined as the initial resource adjustment range;

[0028] For states of low concern, retain observation records and continue to assess them in subsequent cycles.

[0029] In a preferred embodiment, when the local anomaly type labels of adjacent candidate resource domains are consistent, the service object set meets the preset overlap threshold, and the task chain positions are continuous, the adjacent candidate resource domains are merged to update the resource adjustment range.

[0030] In a preferred embodiment, when only one side of the current resource adjustment range is restored, or the corresponding abnormal association state remains established within a continuous preset period, or the adjacent candidate resource domain changes from low association to medium association or above, the resource adjustment range is expanded outward according to the resource adjacency relationship and the task chain continuity relationship. During expansion, strongly adjacent candidate resource domains that simultaneously satisfy the resource adjacency relationship and the task chain continuity relationship are preferentially selected. When the abnormal association result falls back to low association, the resource adjustment range is shrunk back to the corresponding core subdomain. When the abnormal association result on one side of the jointly acting attention domain remains low association within a continuous preset period while the other side still meets the preset establishment condition, the resource adjustment range is shrunk to the core subdomain corresponding to the side that still meets the preset establishment condition. When the resource adjustment range has been shrunk to the corresponding core subdomain and the corresponding abnormal association result remains low association within the subsequent preset period, the active resource adjustment ends and the observation record is retained.

[0031] An intelligent network slicing adaptive orchestration system for 6G communication includes an abnormal entry establishment unit, a candidate resource domain construction unit, an association judgment and marking unit, an adjustment range determination and orchestration unit, and a preset parameter storage area.

[0032] The abnormal entry establishment unit, the candidate resource domain construction unit, the association judgment and marking unit, and the adjustment range determination and arrangement unit are respectively used to execute S101, S102, S103 and S104;

[0033] The preset parameter storage area is used to store and provide the aforementioned units with periodic abnormal parameter groups, range determination parameter groups, resource domain construction parameter groups, association judgment parameter groups, and range update parameter groups.

[0034] The present invention discloses an intelligent network slicing adaptive orchestration method and system for 6G communication, and its effects and advantages are as follows:

[0035] This invention provides an intelligent network slicing adaptive orchestration method and system for 6G communication. It first unifies control-side and detection / tracking-side anomalies of the same target slice within the same current cycle, then jointly organizes radio-side and edge-side resources into candidate resource domains, and compares the correlation results of anomalies on both sides within these candidate resource domains. This allows previously scattered anomalies to be judged within the same local analysis unit under a unified judgment criterion, thereby improving the stability of the judgment of the resource range mainly corresponding to the anomaly in this round. Furthermore, this invention utilizes local anomaly type marking to drive the mapping, expansion, and reclamation of resource adjustment ranges, enabling the resource adjustment range to expand outward and reclamation inward as anomalies change, preventing local problems from being directly amplified into larger-scale resource adjustments. Therefore, based on existing anomaly monitoring results and resource status information, this invention can reduce the risk of orchestration disturbance expansion caused by empirical range determination and improve the consistency and controllability of range reclamation after anomaly mitigation. Attached Figure Description

[0036] Figure 1 This is a schematic diagram of the method flow of the present invention;

[0037] Figure 2 This is a schematic diagram of the system structure of the present invention;

[0038] Figure 3 This is a schematic diagram of step S102 of the present invention;

[0039] Figure 4 This is a schematic diagram of step S104 of the present invention. Detailed Implementation

[0040] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0041] This invention provides an intelligent network slicing adaptive orchestration method and system for 6G communication. It is applicable to scenarios where the same target slice within a vehicle-road cooperative intersection area experiences both control message transmission anomalies and target detection and tracking anomalies within the same current cycle. The method assesses the local resource range primarily corresponding to the anomalies in this cycle and performs adaptive orchestration adjustments accordingly. In such scenarios, the same target slice carries both control message transmission and target detection and tracking tasks; therefore, wireless resources and edge computing resources jointly participate in the service delivery of the same slice.

[0042] Existing network slicing operation monitoring can typically detect slice operation anomalies and collect operational status information for both wireless and edge computing resources. Adjustment targets or scope are then determined based on single-side anomaly results or empirical rules. While this method can obtain both-side anomaly results and resource status results, it often struggles to reliably determine, under a unified judgment criterion, whether the current anomaly primarily corresponds to the wireless resource range, the edge computing resource range, or a localized area affected by both.

[0043] Especially in vehicle-road cooperative intersection areas, changes in traffic flow, pedestrian flow, and sudden events can easily and simultaneously increase the resource demands of both control message transmission and target detection and tracking tasks within a short period. In such cases, anomalies may primarily correspond to the wireless resource range, the edge computing resource range, or a localized area affected by both. If the range primarily corresponding to the anomaly is not accurately determined, problems that originally occurred only in a localized area can easily be amplified into resource adjustments over a larger area, thereby increasing the scope of orchestration disturbances and affecting the recovery of the range after the anomaly is mitigated.

[0044] To this end, the present invention organizes the entire adaptive orchestration process around the dual-sided anomaly entry points formed within the same target slice and the same current period. First, control-side anomalies and detection-tracking-side anomalies are uniformly classified into the same target slice, the same current period, and the same intersection area. Then, wireless-side resources and edge-side resources are jointly organized into candidate resource domains. The dual-sided anomaly association results are compared side by side within each candidate resource domain to generate local anomaly type labels. Finally, resource adjustment range mapping, range expansion, range reclamation, and adaptive orchestration adjustment are performed based on the local anomaly type labels.

[0045] Based on the above design, this invention constructs a complete process for an intelligent network slicing adaptive orchestration method and system for 6G communication, consisting of steps S101 to S104. (Refer to...) Figure 1 , Figure 1 This is a schematic diagram of the method flow of the present invention, which includes:

[0046] Step S101 is used to unify the control message transmission anomalies and target detection and tracking anomalies under the same target slice, the same current period, and the same intersection area, forming a unified anomaly entry point for subsequent local analysis. Step S101 takes the slice anomaly input set X101 as input and outputs the anomaly entry result set R101. R101 is used to receive the unified anomaly content and its anomaly range from both sides, serving as a unified anomaly source for subsequent local resource domain construction and bilateral local anomaly association judgment.

[0047] Step S102 is used to jointly organize wireless-side resources and edge-side resources that are related to the abnormal object and are continuously connected in the control-sensing task closed loop into candidate resource domains based on the anomaly entry result set R101. Step S102 takes the anomaly entry result set R101 and the resource input set X102 as inputs and outputs the candidate resource domain set R102. R102 is used to record the candidate resource domains and their neighborhood relationships and task chain basis, providing a direct basis for subsequent steps to perform bilateral correlation comparison and range update on the same local heterogeneous analysis unit.

[0048] Step S103 is used to determine the degree of local correlation between control-side anomalies and detection / tracking-side anomalies within each candidate resource domain, and to generate local anomaly type labels based on the parallel comparison of the two-side correlation results. Step S103 takes the anomaly entry result set R101 and the candidate resource domain set R102 as inputs and outputs the anomaly correlation result set R103. R103 is used to record the two-side correlation results and their type labels, providing a direct basis for subsequent steps such as resource adjustment range mapping, neighborhood merging, and outward expansion.

[0049] Step S104 maps candidate resource domains to the initial resource adjustment range for this round based on local anomaly type labels. When single-domain adjustment is insufficient, similar anomalies are continuously distributed in adjacent domains, or the anomaly state does not decrease, the range is merged or expanded outwards. Adaptive orchestration adjustment is then performed within the updated range. Step S104 takes the anomaly association result set R103 as the main input and combines the ranges and neighboring domains recorded in the candidate resource domain set R102 to perform resource adjustment range mapping and updating, outputting the adjustment result set R104. R104 records the initial resource adjustment range, the updated resource adjustment range, and their adjustment results, forming a closed-loop range that expands outwards when anomalies rise and retracts when anomalies fall.

[0050] For ease of understanding, the following embodiments are described under a unified system architecture, which can be modified equivalently according to actual needs. The system consists of an abnormal entry point establishment unit, a candidate resource domain construction unit, an association judgment and marking unit, an adjustment range determination and arrangement unit, and a preset parameter storage area. (Refer to...) Figure 2 , Figure 2 This is a schematic diagram of the system structure of the present invention.

[0051] Specifically, the system input includes at least the slice exception input set X101 and the resource input set X102. The preset parameter storage area is used to store and maintain the periodic exception parameter group, the range determination parameter group, the resource domain construction parameter group, the association judgment parameter group, and the range update parameter group, and provides a unified reading interface to each functional unit to keep the parameter standards consistent in the execution process of each step.

[0052] The abnormal entry establishment unit is used to execute method step S101. It receives the slice abnormal input set X101, reads the periodic abnormal parameter group from the preset parameter storage area, extracts the control message transmission abnormal information and target detection and tracking abnormal information, unifies the abnormalities on both sides into the same target slice, the same current period and the same intersection area, and outputs the abnormal entry result set R101.

[0053] The candidate resource domain construction unit is used to execute method step S102. It receives the abnormal entry result set R101 and the resource input set X102, reads the range determination parameter group and the resource domain construction parameter group from the preset parameter storage area, determines the local spatial range, forms the current period abnormal impact object set, merges and pairs the wireless resource items and edge computing resource items, and outputs the candidate resource domain set R102.

[0054] The association judgment and marking unit is used to execute method step S103. It receives the abnormal entry result set R101 and the candidate resource domain set R102, reads the association judgment parameter group from the preset parameter storage area, determines the control side association result and the detection and tracking side association result, and generates the abnormal association result set R103.

[0055] The adjustment range determination and orchestration unit is used to execute method step S104. It receives the abnormal association result set R103 and the candidate resource domain set R102, reads the range update parameter group from the preset parameter storage area, determines the initial resource adjustment range of this round, performs candidate resource domain merging or outward expansion, completes adaptive orchestration adjustment within the updated resource adjustment range, and outputs the adjustment result set R104.

[0056] The implementation process and operational effects of the method of the present invention will be described in detail below with reference to specific embodiments. It should be understood that the embodiments are only used to illustrate the technical solution of the present invention, and not to limit it. The relevant steps, parameters, and module divisions can be appropriately adjusted without changing the essence of the invention.

[0057] In one optional implementation, step S101 is executed by the anomaly entry establishment unit. Around the establishment of a unified anomaly entry, this step involves entry reading and target slice location, bilateral anomaly extraction, current period anomaly aggregation, and result writing back. This step establishes a unified anomaly entry shared by subsequent local resource domain construction and bilateral correlation comparison by uniformly registering control-side anomalies and detection / tracking-side anomalies under the same target slice, the same current period, and the same intersection area.

[0058] In the entry reading and target slice localization process, the target slice identifier, current intersection area identifier, current cycle identifier, control-side status record, detection and tracking-side status record, and service carrying relationship information are read from the slice anomaly input set X101. An anomaly entry analysis object supporting both control and sensing services within the current intersection area, centered around the same target slice, is established. X101 includes at least the slice identifier, intersection area identifier, time identifier, control-side status record, detection and tracking-side status record, and carrying mapping information between the status record and the target slice. The current cycle is used to unify the time scope of anomaly content on both sides. In this embodiment, the current cycle division parameters can be read from the cycle anomaly parameter group, and the current cycle is determined by combining the intersection control system cycle boundary, the sampling rhythm of the control chain and sensing chain, and the coverage requirements of the entire process of control message transmission, detection and tracking processing, and result use for control. Under the current implementation, the intersection control system cycle boundary is preferentially used as the basis for current cycle division; when there is a difference in the sampling cycles of the control chain and sensing chain, the smallest common time window that can completely cover the entire process of control message transmission, detection and tracking processing, and result use for control is used as the current cycle. The business carrying relationship information is used to establish the correspondence between status records and target slices, current intersection areas, service objects, and task chain positions, so as to support the unified management of subsequent bilateral abnormal content under the same slice, the same current period, and the same intersection area.

[0059] In the control-side anomaly extraction, control message transmission status records obtained after entry reading are used to extract control message transmission anomaly information related to the operation of the target slice control chain within the current intersection area. Under the current implementation, control message transmission status records are first filtered by target slice identifier, current intersection area identifier, controlled object identifier, and control message transmission chain position. Then, the current cycle record is compared with the baseline cycle record. The baseline cycle record can be obtained by statistically analyzing the same type of controlled object and chain position records from the most recent consecutive normal cycles. In this embodiment, control-side anomaly comparison parameters can be read from the cycle anomaly parameter group, and combined with the statistical results of the same type of controlled object and chain position within consecutive normal cycles and the control service stability requirements, the comparison criteria for transmission delay, message loss count, message blocking duration, and message interruption count can be determined. When the current cycle record has at least one of the aforementioned indicators significantly higher than the baseline cycle record, the corresponding record is identified as an anomaly record, and the control-side anomaly range is formed by the control message set, controlled object set, and control chain position set corresponding to the anomaly record.

[0060] In the anomaly extraction on the detection and tracking side, target detection and tracking anomaly information related to the operation of the target slice perception chain within the current intersection area is extracted from the target detection and tracking status records obtained after entry reading. Under the current implementation, target detection and tracking status records are first filtered by target slice identifier, current intersection area identifier, target object identifier, and detection and tracking processing location, and then the current cycle record is compared with the baseline cycle record. The baseline cycle record can be obtained by statistically analyzing the records of the same type of target object and the same type of processing location in the most recent consecutive normal cycles. In this embodiment, the anomaly comparison parameters on the detection and tracking side can be read from the cycle anomaly parameter group, and combined with the statistical results of the same type of target object and the same type of processing location within the consecutive normal cycles, as well as the stability requirements of the detection and tracking service, the comparison criteria for the missed detection rate, the number of tracking interruptions, the trajectory drift, the result generation delay, and the number of times the result is used for control are determined. When the current cycle record is significantly higher than the baseline cycle record in at least one of the aforementioned indicators, the corresponding record is identified as an anomaly record, and the target object set, processing location set, and result used for control location set corresponding to the anomaly record form the detection and tracking side anomaly range.

[0061] In the current cycle anomaly aggregation process, control message transmission anomalies and target detection and tracking anomalies are constrained to the same target slice, the same current cycle, and the same intersection area for unified registration, forming an anomaly entry result set R101. Under the current implementation, it is first checked whether control message transmission anomalies and target detection and tracking anomalies are both associated with the same target slice identifier and the same current intersection area identifier. Then, the anomalies on both sides are merged and registered according to the current cycle identifier. When the anomalies on both sides originate from the same target slice and fall into the same current cycle, they are merged into the same R101. When a single-sided anomaly is formed in the current cycle, the corresponding R101 is established synchronously, and a vacancy is registered on the other side. Through this aggregation process, subsequent steps can perform local resource domain construction and bilateral association comparison around the same anomaly source, while maintaining consistency in cross-link timing and attribution criteria.

[0062] In the result write-back, the results of the extraction of anomalies from both sides and their unified classification results are written into the anomaly entry result set R101. R101 includes at least slice identifier, intersection area identifier, period identifier, anomaly information from both sides and their anomaly range. Among them, the anomaly range of the control side and the anomaly range of the detection and tracking side serve as the direct entry basis for the local spatial range contraction in step S102.

[0063] In one alternative implementation, refer to Figure 3Step S102 is executed by the candidate resource domain construction unit. Around the construction of the candidate resource domain, it performs entry point reading and local spatial range determination, local wireless resource item merging, local edge computing resource item merging, bilateral subdomain pairing, relationship basis generation, and result writing back. This step first shrinks the local spatial range of the current round based on the bilateral anomaly range in R101, then merges the wireless-side resource items and edge-side resource items falling into this range to form wireless candidate subdomains and edge candidate subdomains. Subsequently, it performs bilateral subdomain pairing around the current period's anomaly-affected object set and the control perception task closed-loop continuous constraints to form candidate resource domains. Furthermore, it generates task chain position basis and relationships between adjacent candidate resource domains before writing them into the candidate resource domain set R102. Thus, the candidate resource domain is constructed as a local heterogeneous analysis unit for subsequent bilateral local anomaly association judgment and resource adjustment range update.

[0064] In the entry point reading and local spatial range determination, the target slice identifier, current intersection area identifier, service coverage area, resource item record, resource item service object mapping information, resource item task chain location information, and resource adjacency relationship information are read from R101 and X102. X102 includes at least the resource item identifier, resource item coverage boundary, resource item service object set, resource item task chain location, and resource item adjacency relationship. The local spatial range is a local analysis area obtained based on the service coverage area of ​​the target slice in the current intersection area and the bilateral anomaly range contraction in R101, used to limit the resource entry range of the candidate resource domain construction in this round. In this embodiment, the local spatial range determination parameters can be read from the range determination parameter group, and the local spatial range can be determined by combining the anomaly action object coverage segment, the task chain adjacent resource level, and the continuous merging requirement of the coverage segment. Under the current implementation, coverage segments that intersect with either the control-side or the detection-tracking-side anomaly range are first selected as direct anomaly coverage segments. Then, one task chain adjacent resource level is extended forward and backward. Coverage segments with spatially connected or overlapping boundaries and continuous task chain positions are merged to form a local spatial range. A resource item is only included in the current round of candidate resource domain construction if it simultaneously meets three conditions: falling within the current local spatial range, having an object association with both sides of the anomaly range in R101, and having task chain continuity. If a resource item does not simultaneously meet all three conditions, it is not included in the current round of candidate resource domain construction, and the screening and merging preparation for other resource items within the local spatial range continues.

[0065] In the local radio resource item merging process, for radio resource records falling within the local spatial range, radio resource items participating in control message transmission and target detection and tracking data transmission are extracted as the current radio-side analysis objects. The service object mapping results and task chain positions corresponding to each radio resource item are also extracted. After reading, the radio resource items are merged according to service object consistency and task chain consistency. Service object consistency refers to two or more radio resource items corresponding to the same set of service objects, or the overlap ratio of service object sets reaching a preset object overlap threshold. Task chain consistency refers to two or more radio resource items being in the same control-sensing task closed loop, and their task chain positions being the same or adjacent, maintaining a continuous connection. Radio resource items satisfying service object consistency and task chain consistency are merged to form the same radio candidate subdomain. The formed radio candidate subdomain is used to inherit the local service range and task chain position of the current abnormal action object on the radio side, serving as the basis for subsequent radio-side pairing of bilateral subdomains.

[0066] In the local edge computing resource item merging process, for edge-side resource records falling within the local spatial range, edge computing resource items involved in target detection and tracking computation, result generation, and result use for control are extracted as the current edge-side analysis objects. The task scope, service object mapping results, and task chain position corresponding to each edge computing resource item are also extracted. Subsequently, merging is performed according to the service object consistency and task chain consistency rules consistent with the wireless side. When multiple edge computing resource items jointly serve the same batch of intersection-related control objects or the same batch of key target objects, and undertake the same or adjacent continuous processing links in the control perception task closed loop, these edge computing resource items are merged to form the same edge candidate subdomain. The edge computing resource items participating in this round of edge candidate subdomain construction form object associations around the current abnormal action object and maintain continuous connection in the current task closed loop. The formed edge candidate subdomain is used to take over the local processing scope and task chain position of the current abnormal action object on the edge side, serving as the basis for subsequent edge-side pairing of bilateral subdomains.

[0067] In the bilateral subdomain pairing, wireless candidate subdomains and edge candidate subdomains are used as pairing objects. Cross-side pairing is performed around the current period's set of objects affected by anomalies and the continuous connection relationship of the control and perception task closed loop to form the smallest heterogeneous analysis unit for subsequent bilateral local anomaly association judgment. The current period's set of objects affected by anomalies is the set of control objects and the set of target objects derived from the control-side anomaly range and the detection and tracking-side anomaly range in R101. In this embodiment, resource domain construction-related parameters can be read from the resource domain construction parameter group, and combined with the service object set overlap, task chain position continuity, object domain mapping relationship, direct service acceptance relationship, and control and perception task closed loop continuous link judgment requirements, the anomaly affected object set formation criteria and candidate resource domain pairing criteria are determined. Under the current implementation, the set of controlled objects in the control-side anomaly range and the set of target objects in the detection and tracking-side anomaly range are first mapped to the same object domain through service carrying relationships, and then an object intersection is formed. When the object intersection is formed, it is determined as the set of objects affected by the current period's anomaly. When the object intersection is not formed, the union of objects with direct service carrying relationships is determined as the set of objects affected by the current period's anomaly. Subsequently, the service object sets and task chain positions are extracted from the wireless candidate subdomains and edge candidate subdomains, respectively. Only when a wireless candidate subdomain and an edge candidate subdomain jointly serve the set of objects affected by the current period's anomaly, and both are located in the same continuous link of the control and sensing task closed loop, are the wireless candidate subdomain and the edge candidate subdomain paired to generate a candidate resource domain. When a wireless candidate subdomain and an edge candidate subdomain do not jointly serve the set of objects affected by the current period's anomaly, or are not located in the same continuous link of the control and sensing task closed loop, the wireless candidate subdomain and the edge candidate subdomain are not paired to generate a candidate resource domain, and pairing judgment is continued for other bilateral subdomain combinations. The candidate resource domain is formed by pairing wireless candidate subdomains and edge candidate subdomains that satisfy the constraints of the same abnormal impact object and the closed-loop continuity constraints of the same control and perception task. It serves as the basic processing object for subsequent bilateral local anomaly association judgment and resource adjustment range update.

[0068] In the relationship generation process, for each candidate resource domain, its wireless candidate subdomain identifier, edge candidate subdomain identifier, service object set, task chain position basis, and adjacent candidate resource domain relationships are recorded. The task chain position basis is used to characterize the receiving position of the two-sided subdomains in the control and perception task closed loop, and the adjacent candidate resource domain relationships are used to characterize the spatial adjacency and chain adjacency relationships between different candidate resource domains. Under the current implementation, spatial adjacency relationships are established based on the adjacency of resource coverage boundaries, and chain adjacency relationships are established based on the continuity of task chain positions. When a candidate resource domain simultaneously forms two types of adjacency relationships, it is marked as a strongly adjacent candidate resource domain for subsequent priority merging and priority expansion judgments. By recording the aforementioned relationship basis, the candidate resource domain not only receives the pairing results of the two-sided subdomains in the current period, but also simultaneously forms the neighborhood organization basis required for subsequent range updates.

[0069] In the result write-back, the local spatial range determination result, the bilateral subdomain merging and pairing result, and the relationship basis are written into the candidate resource domain set R102. R102 includes at least the candidate resource domain identifier, bilateral subdomain identifier, service object set, task chain position basis, and neighborhood relationship information, which serve as the direct basis for subsequent bilateral association judgment and resource adjustment range update.

[0070] In an optional implementation, step S103 is performed by the association judgment and labeling unit. This involves attributing bilateral local anomalies within the candidate resource domain, including entry point reading and candidate resource domain traversal, control-side association judgment, detection and tracking-side association judgment, local anomaly type classification, association reference information generation, and result writing back. This step compares the control-side association results and the detection and tracking-side association results side-by-side on the same candidate resource domain, generating local anomaly type labels that can directly drive subsequent resource adjustment range mapping.

[0071] In the entry reading and candidate resource domain traversal, bilateral anomaly content, bilateral subdomain information, service object set, task chain position basis, and adjacent candidate resource domain relationships are extracted from R101 and R102 according to the candidate resource domain. Each domain is used as the current analysis object, and an entry point for the local anomaly association judgment record is established for each candidate resource domain, so that the control-side association judgment and the detection and tracking-side association judgment are carried out around the same candidate resource domain.

[0072] In the control-side association determination, for the current candidate resource domain, the control-side association result is determined by comprehensively considering its carrying capacity for anomaly-related control objects, the control anomaly changes within the resource domain, and the control-side reference status provided by adjacent candidate resource domains. Specifically, the message proportion of relevant control objects in the intersection area is calculated based on the number of messages or the weighted message count to characterize the object association degree between the current candidate resource domain and the control chain anomaly; then, control message transmission anomaly information corresponding to the current candidate resource domain is extracted from R101 to identify its control anomaly changes in the current period; subsequently, a control-side neighborhood reference is formed by combining the relationship between adjacent candidate resource domains, prioritizing the control anomaly performance of adjacent candidate resource domains as a reference, and using the control anomaly degree of the same candidate resource domain in the previous period as a supplementary reference.

[0073] After the aforementioned extraction, three individual judgment results are formed based on message proportion, control anomaly changes, and control-side neighborhood reference. A message proportion reaching a high threshold is determined as high; reaching a medium threshold but not a high threshold is determined as medium; otherwise, it is determined as low. Control anomaly changes reaching a high threshold are determined as high; reaching a medium threshold but not a high threshold is determined as medium; otherwise, it is determined as low. In the control-side neighborhood reference, if an adjacent candidate resource domain has a continuous task chain position with the current candidate resource domain and exhibits prominent control-side anomaly performance, it is determined as high; if an adjacent candidate resource domain has an increased control-side anomaly performance, or if the control anomaly level of the same candidate resource domain in the previous period has not decreased, it is determined as medium; otherwise, it is determined as low. If at least two of the three individual judgment results are high, the control-side association result of the current candidate resource domain is determined as high association; if the high association condition is not met but at least two reach medium or above, it is determined as medium association; otherwise, it is determined as low association.

[0074] In the detection and tracking-side association determination, for the current candidate resource domain, the association result is determined by comprehensively considering its service level to the abnormal target objects, the changes in detection and tracking anomalies within the resource domain, and the detection and tracking-side reference status provided by adjacent candidate resource domains. Specifically, the proportion of key targets is calculated based on the number of target objects or the weighted number of targets to characterize the object association degree between the current candidate resource domain and the anomalies in the detection and tracking chain; then, the target detection and tracking anomaly information corresponding to the current candidate resource domain is extracted from R101 to identify its changes in detection and tracking anomalies in the current period; subsequently, a detection and tracking-side neighborhood reference is formed by combining the relationship between adjacent candidate resource domains, prioritizing the detection and tracking anomaly performance of adjacent candidate resource domains as a reference, and using the detection and tracking anomaly degree of the same candidate resource domain in the previous period as a supplementary reference.

[0075] After the aforementioned extraction, three individual judgment results are formed based on the proportion of key targets, the detection and tracking of abnormal changes, and the detection and tracking side neighborhood reference. A key target proportion reaching a high threshold is classified as high; reaching a medium threshold but not reaching a high threshold is classified as medium; otherwise, it is classified as low. Similarly, an abnormal change in detection and tracking reaching a high threshold is classified as high; reaching a medium threshold but not reaching a high threshold is classified as medium; otherwise, it is classified as low. In the detection and tracking side neighborhood reference, if an adjacent candidate resource domain has a continuous task chain position with the current candidate resource domain and exhibits prominent abnormal detection and tracking performance, it is classified as high; if an adjacent candidate resource domain has an increased abnormal detection and tracking performance, or if the abnormal detection and tracking performance of the same candidate resource domain in the previous period has not decreased, it is classified as medium; otherwise, it is classified as low. If at least two of the three individual judgment results are high, the detection and tracking side association result of the current candidate resource domain is classified as high association; if the high association condition is not met but at least two are medium or above, it is classified as medium association; otherwise, it is classified as low association.

[0076] In the classification of local anomaly types, for the current candidate resource domain, the correlation results from the control side and the correlation results from the detection and tracking side are compared side-by-side. Local anomaly type labels are generated based on the relative strength and validity of the correlation results from both sides to distinguish between control-dominated anomalies, detection and tracking-dominated anomalies, and anomalies caused by the combined effects of both sides. Preset validity conditions are used to determine whether a correlation result from one side is sufficient to trigger subsequent resource adjustments. High correlation is directly considered to have met the preset validity conditions; medium correlation is also considered to have met the preset validity conditions when it is valid for two consecutive current periods; low correlation corresponds to remaining in the observation state.

[0077] Among them, a medium correlation is established for two consecutive current periods, meaning that the corresponding correlation results of the current candidate resource domain are both medium correlations in the previous and current periods, and the service object set and task chain position are consistent or continuously connected. A low correlation means that the two-sided correlation results, change trends, and neighborhood reference information of the current candidate resource domain are retained and continue to be used as the correlation judgment object in subsequent periods, without directly triggering the active resource adjustment range mapping in the current period.

[0078] When both the control-side association result and the detection-tracking-side association result meet the preset conditions and are at the same association level, the current candidate resource domain is marked as a domain of common interest. When the control-side association result is higher than the detection-tracking-side association result and meets the preset conditions, the current candidate resource domain is marked as a control-side priority domain. When the detection-tracking-side association result is higher than the control-side association result and meets the preset conditions, the current candidate resource domain is marked as a detection-tracking-side priority domain. When neither side's association result meets the preset conditions, the current candidate resource domain is maintained in a low-interest state. The low-interest state is used to retain the observation records and continuous periodic tracking basis of the current candidate resource domain and serves as the basis for re-judging the local anomaly type in subsequent periods. The local anomaly type marking is used to drive the initial resource adjustment range mapping and subsequent range update in step S104, and the original bilateral association results are used to retain the bilateral association judgment basis for the corresponding candidate resource domain.

[0079] In the generation of associated reference information, for each candidate resource domain, its control-side association results, detection and tracking-side association results, local anomaly type markers, and bilateral anomaly reference states of adjacent candidate resource domains are recorded. Furthermore, the bilateral association change trend of the current candidate resource domain in continuous periods is recorded. When the current candidate resource domain rises from low association to medium association or above within the current period, or forms the same local anomaly type with adjacent candidate resource domains and exhibits a continuous distribution, it is registered as an extended reference state. This extended reference state is written into R103 as part of the neighborhood reference information for subsequent range expansion judgment.

[0080] In the result write-back, the bilateral association judgment results, local anomaly type markers, and their reference information are written into the anomaly association result set R103. R103 includes at least the candidate resource domain identifier, bilateral association results, local anomaly type markers, change trends, and neighborhood reference information, serving as the direct judgment basis for the execution range mapping, neighborhood merging, and outward expansion in step S104.

[0081] In one alternative implementation, refer to Figure 4 Step S104 is executed by the adjustment range determination and orchestration unit. It revolves around updating and orchestrating the resource adjustment range, including entry point reading and initial resource adjustment range determination, control-side priority domain processing, detection and tracking-side priority domain processing, joint action domain processing, merging and outward expansion of adjacent candidate resource domains, adaptive orchestration adjustment execution, shrinkage stop judgment, and result writing back. Based on local anomaly type markings, change trends, and neighborhood reference information, this step progressively completes initial range mapping, neighborhood merging, outward expansion, orchestration adjustment, and fallback shrinkage, enabling the resource adjustment range to adaptively update as the anomaly state evolves.

[0082] In the entry point reading and initial resource adjustment range determination, local anomaly type markers, bilateral correlation results, change trends, bilateral subdomain identifiers, service object sets, task chain location criteria, and neighborhood relationship information are read from R103 and R102. Candidate resource domains are then extracted sequentially as the current adjustment targets. The resource adjustment range is the resource coverage area for slice orchestration adjustments around a candidate resource domain within the current period. The initial resource adjustment range is the first round of adjustment range directly mapped from the current candidate resource domain based on the local anomaly type markers. The core subdomain is the wireless candidate subdomain or edge candidate subdomain within the current candidate resource domain that directly corresponds to the currently dominant anomaly side and is preferentially selected in this round of initial resource adjustment range mapping. Specifically, for control-side priority domains, the wireless candidate subdomain corresponding to the current candidate resource domain is determined as the core subdomain; for detection-tracking-side priority domains, the edge candidate subdomain corresponding to the current candidate resource domain is determined as the core subdomain; for jointly affected domains, the wireless candidate subdomain and the edge candidate subdomain corresponding to the current candidate resource domain are combined into the initial resource adjustment range, and their corresponding core subdomain identifiers are retained for subsequent contraction mapping when only one side remains valid. After reading, the initial resource adjustment range mapping is first performed on the current candidate resource domain according to the local anomaly type label. For control-side priority domains, the wireless candidate subdomain is used as the initial resource adjustment range. For detection-tracking-side priority domains, the edge candidate subdomain is used as the initial resource adjustment range. For jointly affected domains, the combined range of the wireless candidate subdomain and the edge candidate subdomain is used as the initial resource adjustment range. For low-attention states, the current candidate resource domain is not mapped to the active resource adjustment range; only the bilateral correlation results, change trends, and neighborhood reference information of the candidate resource domain are retained as observation objects for subsequent periods. After completing the initial resource adjustment range mapping, whether to increase the adjustment intensity, perform neighborhood merging, expand outward, or shrink back should be based on the corresponding side association result change trend and neighborhood reference information in the next current period or consecutive preset periods.

[0083] In the control-side priority domain processing, when a candidate resource domain is marked as a control-side priority domain, the corresponding radio candidate subdomain is used as the initial resource adjustment scope for this round. The set of radio resource items, control message carrying range, and service object set corresponding to this radio candidate subdomain are determined as the priority adjustment objects of the control chain for this round. Under the current implementation, lightweight radio-side adjustments are prioritized. If the control-side correlation result still meets the preset conditions after the lightweight adjustment, or if there is no decline within a consecutive preset period, the intensity of the radio-side adjustment is further increased. Radio-side adjustment actions may include resource quota reallocation, scheduling priority reset, bearer path switching, or load balancing.

[0084] In the priority domain processing on the detection and tracking side, when a candidate resource domain is marked as a priority domain on the detection and tracking side, the edge candidate subdomains corresponding to the current candidate resource domain are used as the initial resource adjustment scope for this round. The set of edge computing resource items, the scope of tasks they support, and the set of service objects corresponding to these edge candidate subdomains are determined as the priority adjustment objects for this round of the detection and tracking chain. Under the current implementation, lightweight adjustments on the edge side are performed first. If the correlation results on the detection and tracking side still meet the preset conditions after the lightweight adjustments, or if there is no decrease within a consecutive preset period, the intensity of the edge side adjustments is further increased. Edge side adjustment actions may include adjusting computing resource quotas, adjusting task concurrency, migrating task instances, resetting result generation priorities, or reconstructing the path for result control.

[0085] In the processing of jointly affected domains of interest, when a candidate resource domain is marked as a jointly affected domain of interest, the combined range of the wireless candidate subdomain and the edge candidate subdomain corresponding to the current candidate resource domain is used as the initial resource adjustment range for this round. The corresponding set of wireless resource items, set of edge computing resource items, control message carrying range, target object set, and task chain position are determined as the objects of this round of bilateral joint adjustment. Under the current implementation, bilateral collaborative lightweight adjustment is performed first. If the correlation result of either side still meets the preset conditions after the collaborative adjustment, or if neither side shows a decrease within a consecutive preset period, the intensity of bilateral collaborative adjustment is further increased. For jointly affected domains of interest, when only one side declines while the other side still meets the preset conditions, the combined range is maintained, and the subsequent expansion judgment focuses on the side that still meets the conditions and its adjacent candidate resource domains. When the side that declines continues to maintain low correlation within a subsequent preset period, while the other side still meets the preset conditions, the updated resource adjustment range is shrunk towards the core subdomain corresponding to the side that still meets the conditions. The dual-side coordinated adjustment actions may include synchronous adjustment of resource quotas on the wireless side and the edge side, switching of bearer paths, and remapping of task instances.

[0086] In the merging and outward expansion of adjacent candidate resource domains, for the current candidate resource domain whose initial resource adjustment range has been determined, the neighboring reference information, changing trends, and relationships of adjacent candidate resource domains are considered to determine whether the current resource adjustment range needs to be merged or expanded outward. Insufficient single-domain adjustment refers to a situation where, after performing resource adjustments on the current candidate resource domain alone in this round, the association results of the corresponding dominant anomaly side still meet the preset conditions for establishment at the end of the current cycle or the beginning of the next current cycle, or the current candidate resource domain and its strongly adjacent candidate resource domains still maintain the same local anomaly type and exhibit a continuous distribution. For jointly acting domains of interest, insufficient single-domain adjustment also includes situations where the association results of either side still meet the preset conditions for establishment, or where, although the two sides do not simultaneously maintain high association, they still jointly support the current continuous distribution of anomalies. Merging adjacent candidate resource domains refers to merging the resource adjustment ranges corresponding to these candidate resource domains into the same update range when multiple candidate resource domains form a continuous distribution of the same type of anomaly within a local spatial range. Under the current implementation, when the local anomaly type labels of adjacent candidate resource domains are consistent, the service object set meets the preset overlap threshold, and the task chain positions are continuous, the current candidate resource domain and the corresponding adjacent candidate resource domain are merged. Outward expansion refers to extending the current resource adjustment range outward to the adjacent candidate resource domain according to resource adjacency and task chain continuity when only one side of the current resource adjustment range has been restored, or the corresponding side's anomaly association state still meets the preset conditions within a continuous preset period, or the adjacent candidate resource domain changes from low association to medium association or higher. During expansion, strongly adjacent candidate resource domains that simultaneously satisfy resource adjacency and task chain continuity are preferentially selected; if no strongly adjacent candidate resource domain exists, an ordinary adjacent candidate resource domain is selected along the chain with the higher current anomaly. In this embodiment, range update-related parameters can be read from the range update parameter group, and combined with the continuous distribution state of local anomaly types, the adjacency relationship of candidate resource domains, the determination requirements for strongly adjacent candidate resource domains, and the continuous periodic association change requirements, the candidate resource domain merging caliber and outward expansion caliber are determined. By first performing local adjustments around the current candidate resource domain, and then gradually merging or expanding when similar anomalies are continuously distributed or the anomaly status continues to decline, it is possible to avoid performing overly broad adjustments on the entire slice service range at the beginning, while also covering the situation where local anomalies spread to adjacent domains.

[0087] During adaptive orchestration adjustment, for the object range already determined as the initial or updated resource adjustment range, the orchestration adjustment strategy of the target slice is invoked, and this round of adaptive orchestration adjustment is performed on the wireless resource items and edge computing resource items within the corresponding range. Adaptive orchestration adjustment dynamically adjusts the target slice's wireless-side bearer relationships, edge-side task mapping relationships, bilateral scheduling priority relationships, and resource quota relationships around the abnormal clustering locations within the current resource adjustment range. After the adjustment is completed, the subsequent changes in the control-side association results and detection / tracking-side association results within the current resource adjustment range are further read to form a state summary after this round of adjustment, which serves as a direct basis for subsequent decisions on contraction, maintenance, or continued expansion.

[0088] In the shrinkage cessation judgment, for candidate resource domains that have already completed orchestration adjustments, it is determined whether the resource adjustment range after this round of updates needs to be maintained, shrunk, or stopped expanding. When the correlation results of both sides fall back to low correlation within a continuous period, the resource adjustment range is shrunk to the core subdomain corresponding to the current candidate resource domain, and outward expansion is stopped. For jointly affected interest domains, when both sides have retained the corresponding core subdomain identifiers before this round, if only one side maintains low correlation within a continuous preset period, while the other side maintains medium or high correlation, or still meets the preset conditions, the updated resource adjustment range is shrunk to the core subdomain corresponding to the side that still maintains the condition; if only one side falls back within the current period while the other side remains established, but has not yet met the shrinkage condition of continuous low correlation, the current range is maintained and the resource adjustment range of the corresponding side is retained. When the correlation result of either side continues to rise, or a new strongly adjacent candidate resource domain rises from low correlation to medium correlation or above, the expansion state is retained and the outward expansion judgment of the next current period begins. When the current resource adjustment scope has shrunk to the core subdomain and the corresponding abnormal correlation results remain at a low correlation within the subsequent preset period, the current round of proactive resource adjustment ends, and only the observation record is retained. The recovery status flag is used to record the results of shrinking, maintaining, continuing to expand, or stopping proactive adjustment. When the local abnormal correlation results fall back, the resource adjustment scope shrinks according to the current candidate resource domain and its core subdomain, forming a closed loop of scope expansion when abnormalities rise and recovery when abnormalities fall back.

[0089] In the result write-back, the resource adjustment range mapping results, neighborhood merging and outward expansion results, adaptive orchestration adjustment results, and recovery status markers are written into the adjustment result set R104. R104 includes at least the candidate resource domain identifier, the initial resource adjustment range, the updated resource adjustment range, the adjustment action type, the adjusted state summary, the recovery status marker, the expansion direction, and the shrinkage stop result, serving as the direct basis for further judging abnormal convergence states and resource adjustment range update directions in subsequent cycles.

[0090] In an optional implementation, taking target slice N1 within intersection area A as an example, the coordinated execution process of steps S101 to S104 is further explained. Target slice N1 simultaneously carries control message transmission services and target detection and tracking services within intersection area A. For ease of explanation, the three consecutive current cycles are denoted as cycle T1, cycle T2, and cycle T3.

[0091] Within period T1, target slice N1 provides operational support to controlled objects C1 and C2, as well as target objects O1, O2, and O3. At this time, the control message transmission latency and congestion duration increase in the eastern entrance chain segment, while the result generation latency and the number of failed attempts to use the results for control at the eastern entrance edge processing location also increase. Although these two types of anomalies occur in the control chain and the detection and tracking chain respectively, they both correspond to the operational support process of target slice N1 for the same batch of key traffic objects within intersection area A.

[0092] In step S101, the system extracts control-side anomaly information and detection-tracking-side anomaly information from the slice anomaly input set X101, and registers both types of anomalies under the same target slice N1, the same period T1, and the same intersection area A, obtaining the anomaly entry result set R101. The control-side anomaly range and the detection-tracking-side anomaly range recorded in R101 serve as the common basis for subsequent local range shrinkage and candidate resource domain construction.

[0093] In step S102, based on the bilateral anomaly range recorded in R101, the system performs local contraction on the service coverage of target slice N1 within intersection area A, obtaining a local spatial range near the east entrance. Within this local spatial range, radio resource items W1 and W2 jointly carry the control sensing task chains corresponding to controlled objects C1 and C2 and target objects O1 and O2, and are continuously connected in the task chain position, thus merging to form the same radio candidate subdomain. Edge resource items E1 and E2 jointly serve target objects O1 and O2, and are continuously connected in the detection processing, result generation, and result supply for control stages, thus merging to form the same edge candidate subdomain. Since this radio candidate subdomain and the edge candidate subdomain jointly serve the same batch of anomaly-related objects and are located in the same continuous task chain, the system pairs them to generate candidate resource domain D1. Another adjacent chain segment forms candidate resource domain D2, and D1, D2, and their adjacency relationships are written into the candidate resource domain set R102.

[0094] In step S103, the system performs association judgment on D1 and D2 respectively. For D1, the control messages it carries mainly correspond to control objects C1 and C2, and the control anomaly changes are significantly higher than those of the adjacent D2. Therefore, the control-side association result of D1 is determined to be high association. At the same time, the target objects O1 and O2 served by D1 directly correspond to the anomaly range on the detection and tracking side, but the anomaly changes on the detection and tracking side are lower than those on the control side. Therefore, the detection and tracking side association result of D1 is determined to be medium association. Based on this, the system marks D1 as a control-side priority area of ​​interest. For D2, the proportion of anomaly-related objects it carries is relatively low, and the anomaly changes on both the control side and the detection and tracking side are weaker than those on D1. Therefore, D2 remains in a low-interest state. The above judgment results are written into the anomaly association result set R103.

[0095] In step S104, the system determines the initial resource adjustment range for this round based on the local anomaly type markers in R103. Since D1 is marked as a priority domain for the control side, the radio candidate subdomain corresponding to D1 is taken as the initial resource adjustment range for this round, and lightweight radio-side adjustments are performed first. After the adjustment is completed, the initial range mapping result and the adjustment result are written into the adjustment result set R104.

[0096] Within cycle T2, the system executes steps S101 to S103 again. At this time, although the control-side anomalies in the region where D1 is located have decreased, they remain highly correlated. Simultaneously, the proportion of controlled object messages and changes in control anomalies for D2 have increased, and its control-side correlation result has risen to above medium correlation. Combining the correlation change trend recorded in R103 and the adjacency relationship between D1 and D2, the system determines that performing single-domain adjustments only around D1 is insufficient to cover the current anomaly range. Therefore, in step S104, D1 and D2 are merged, and the resource adjustment range is expanded from D1 to the combined range of D1 and D2, continuing orchestration adjustments within the updated range.

[0097] Within cycle T3, the system continues to execute steps S101 to S103. At this point, both the control-side anomalies and the detection / tracking-side anomalies recorded in R101 have significantly decreased. The control-side correlation result of D1 decreases, and the control-side correlation result of D2 also falls back to low correlation. The detection / tracking-side anomalies no longer show a sustained increase. Based on this, the system determines that the adjustment range after the previous round of expansion has covered the main anomaly locations and achieved mitigation. Therefore, in the contraction stop judgment in step S104, the updated resource adjustment range is contracted from the joint range of D1 and D2 back to the core subdomain corresponding to D1, and further outward expansion is stopped. Simultaneously, the recovery status for this round is registered in R104.

[0098] As can be seen from the above single-scenario closed-loop process, this solution first uniformly registers the two types of anomalies formed on the same target slice within the same current cycle. Then, it organizes candidate resource domains within a local area and judges the prominence of control-side anomalies and detection / tracking-side anomalies separately within the same candidate resource domain, thereby determining the scope of resource adjustment in this round. As the anomaly state changes, the resource adjustment scope can continue to be merged, expanded, and reclaimed, thus keeping the adjustment actions within a range that matches the anomaly location.

[0099] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0100] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and inventive constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0101] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0102] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0103] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A smart network slicing adaptive orchestration method for 6G communication, characterized in that, include: S101, extract abnormal information of control message transmission and abnormal information of target detection and tracking in the current cycle of the target slice in the vehicle-road cooperative intersection area, and associate the two with the same target slice, the same current cycle and the same intersection area to form abnormal entry information; S102, based on the anomaly entry information, jointly merge and pair wireless resource items and edge computing resource items that jointly serve the anomaly corresponding object to form a candidate resource domain; S103, for each candidate resource domain, determine the correlation between control message transmission anomaly information and target detection and tracking anomaly information within that candidate resource domain, and generate local anomaly type markers based on the comparison results; Local anomaly type labeling includes control-side priority interest domains, detection-tracking-side priority interest domains, joint-effect interest domains, and low-interest status. High correlation is directly considered to meet the preset conditions; medium correlation is considered to meet the preset conditions when it is true for two consecutive current periods; low correlation corresponds to remaining in the observation state. When both control-side and detection-tracking-side correlation results meet the preset conditions and are at the same correlation level, the corresponding candidate resource domain is labeled as a joint-effect interest domain. When the control-side correlation result is higher than the detection-tracking-side correlation result and meets the preset conditions, the corresponding candidate resource domain is labeled as a control-side priority interest domain. When the detection-tracking-side correlation result is higher than the control-side correlation result and meets the preset conditions, the corresponding candidate resource domain is labeled as a detection-tracking-side priority interest domain. When neither the control-side nor the detection-tracking-side correlation result meets the preset conditions, the corresponding candidate resource domain remains in a low-interest state. S104: Determine the initial resource adjustment range based on the local anomaly type marker, update the initial resource adjustment range, perform adaptive orchestration adjustment on the updated resource adjustment range, and shrink the resource adjustment range when the anomaly falls back.

2. The intelligent network slicing adaptive orchestration method for 6G communication according to claim 1, characterized in that, When abnormal entry information is generated, the abnormal range on the control side corresponding to the abnormal control message transmission information and the abnormal range on the detection and tracking side corresponding to the abnormal target detection and tracking information are recorded. When only one-sided abnormal information is generated in the current cycle, the corresponding abnormal entry information is still established, and the vacancy status is registered on the other side.

3. The intelligent network slicing adaptive orchestration method for 6G communication according to claim 1, characterized in that, When forming a candidate resource domain, the abnormal direct coverage segment is determined based on the abnormal information transmitted by the control message and the abnormal information of target detection and tracking. Starting from the abnormal direct coverage segment, the resource level of the adjacent task chain is extended along the task chain before and after it. Coverage segments with connected or overlapping spatial boundaries and continuous task chain positions are merged to obtain the local spatial range.

4. The intelligent network slicing adaptive orchestration method for 6G communication according to claim 3, characterized in that, When forming candidate resource domains, the following are included: Map the control object corresponding to the control message transmission anomaly information and the target object corresponding to the target detection and tracking anomaly information to the same object domain; Wireless resource items that have the same service objects or whose service object overlap ratio reaches a preset object overlap threshold and are located in the same or adjacent task chain positions in the same control and sensing task closed loop are merged to form wireless candidate subdomains. Edge computing resource items that have the same service objects or whose service object overlap ratio reaches a preset object overlap threshold and are located in the same or adjacent task chain positions in the same control and perception task closed loop are merged to form edge candidate subdomains. Candidate resource domains are formed by pairing wireless candidate subdomains and edge candidate subdomains that jointly serve the mapped object and are located in the same continuous link of the control and perception task closed loop.

5. The intelligent network slicing adaptive orchestration method for 6G communication according to claim 1, characterized in that, When determining the association status within each candidate resource domain, the following is included: The control-side association result is determined based on the current candidate resource domain's capacity to support anomaly-related control objects, abnormal changes in control message transmission, and control anomaly references in adjacent candidate resource domains. Based on the service level of the current candidate resource domain to the abnormal target object, the abnormal changes in target detection and tracking, and the abnormal reference of detection and tracking in adjacent candidate resource domains, the correlation results on the detection and tracking side are determined.

6. The intelligent network slicing adaptive orchestration method for 6G communication according to claim 1, characterized in that, When determining the initial resource adjustment range, the following should be included: The wireless candidate subdomains corresponding to the priority domains on the control side are determined as the initial resource adjustment range; The edge candidate subdomains corresponding to the priority focus domains on the detection and tracking side are determined as the initial resource adjustment range; The combined range of wireless candidate subdomains and edge candidate subdomains corresponding to the common interest domain is determined as the initial resource adjustment range; For states of low concern, retain observation records and continue to assess them in subsequent cycles.

7. The intelligent network slicing adaptive orchestration method for 6G communication according to claim 6, characterized in that, When adjacent candidate resource domains have the same local anomaly type label, the service object set meets the preset overlap threshold, and the task chain position is continuous, the adjacent candidate resource domains are merged to update the resource adjustment range.

8. The intelligent network slicing adaptive orchestration method for 6G communication according to claim 7, characterized in that, When only one side of the current resource adjustment range is restored, or the corresponding abnormal association status remains valid for a continuous preset period, or the adjacent candidate resource domain changes from low association to medium association or above, the resource adjustment range is expanded outward according to the resource adjacency relationship and the task chain continuity relationship. During expansion, strong adjacent candidate resource domains that simultaneously satisfy the resource adjacency relationship and the task chain continuity relationship are selected first. When the abnormal association result falls back to low association, the resource adjustment range is shrunk back to the corresponding core subdomain. When the abnormal association result on one side of the jointly acting attention domain remains low association for a continuous preset period while the other side still meets the preset conditions, the resource adjustment range is shrunk to the core subdomain corresponding to the side that still meets the preset conditions. When the resource adjustment range has been shrunk to the corresponding core subdomain and the corresponding abnormal association result remains low association for a subsequent preset period, the active resource adjustment ends and the observation record is retained.

9. A smart network slicing adaptive orchestration system for 6G communication, used to implement the smart network slicing adaptive orchestration method for 6G communication as described in any one of claims 1 to 8, characterized in that, It includes an abnormal entry point establishment unit, a candidate resource domain construction unit, an association judgment and marking unit, an adjustment range determination and arrangement unit, and a preset parameter storage area; The abnormal entry establishment unit, the candidate resource domain construction unit, the association judgment and marking unit, and the adjustment range determination and arrangement unit are respectively used to execute S101, S102, S103 and S104; The preset parameter storage area is used to store and provide the aforementioned units with periodic abnormal parameter groups, range determination parameter groups, resource domain construction parameter groups, association judgment parameter groups, and range update parameter groups.

Images