Communication signal processing method for 5G fusion perception

Abstract

The invention relates to the field of wireless communication, in particular to a communication signal processing method for 5G fusion perception. The method comprises the following steps: acquiring a network topology, a reference signal list, a probe list and a resource budget parameter group, completing field mapping table registration, unifying time reference registration and session identifier registration, and forming a session identifier recording structure; collecting and receiving baseband sampling, extracting a pilot frequency observation abstract, performing time alignment, and generating an alignment data packet and a quality mark table; generating a perception observation abstract based on the alignment data packet and the quality tag table, extracting a candidate parameter set and assembling an evidence slot record set, and generating a prior constraint packet; and executing joint channel parameter estimation based on the evidence slot record set and the prior constraint packet, extracting uncertainty records, generating pilot frequency reconfiguration suggestions, and outputting detection strategy configuration. According to the method, a through link of acquisition, alignment, evidence assembly, estimation and configuration is realized, and the consistency of evidence organization and the traceability of detection strategy configuration are improved.

Description

Technical Field

[0001] This invention relates to the field of wireless communication, and more particularly to a fusion sensing communication signal processing method for 5G. Background Technology

[0002] In the field of wireless communication, existing schemes for 5G-oriented converged sensing communication signal processing methods typically rely on configuring reference signal lists and probe lists based on network topology. They then generate aligned data packets based on received baseband sampling and pilot observation summary extraction, and combine sensing observation summary generation with channel parameter estimation to form a detection strategy configuration. However, this approach suffers from limitations such as inconsistent field mapping table registration standards, inconsistencies between unified time reference registration and time alignment links, and difficulties in applying resource budget parameter constraints throughout the entire process. Existing methods often rely on single-path processing of received baseband sampling packets and subsequent generation of pilot and sensing observation summaries. Under resource budget parameter sets, insufficient correlation between aligned data packets and quality label tables, along with fragmented evidence organization, makes it difficult to achieve stable implementation of detection strategy configuration. For the joint processing of evidence slot record sets and prior constraint packages, existing technologies generally lack a unified assembly standard for evidence slot record sets between pilot observation summaries and sensing observation summaries. Furthermore, they lack traceable constraint carriers and consistent record links between joint channel parameter estimation, uncertainty record extraction, and pilot reconfiguration suggestion generation. This makes it difficult to form a consistent process of acquisition-alignment-assembly-estimation-suggestion-configuration in the application scenario of a 5G-oriented fusion sensing communication signal processing method. Consequently, the reference to uncertainty records and the generation path of pilot reconfiguration suggestions during the detection strategy configuration process are prone to standard drift and link breakage. Summary of the Invention

[0003] To address the aforementioned technical problems, this invention provides a 5G-oriented converged sensing communication signal processing method, comprising: S100: Obtain the network topology, reference signal list, probe list, and resource budget parameter group; perform field mapping table registration, unified time base registration, and session identifier registration processing to generate a session identifier record structure; wherein, the session identifier record structure includes a session identifier, a session start time record, a network topology snapshot identifier, a reference signal list version identifier, a probe list version identifier, a resource budget parameter group version identifier, a field mapping table version identifier, and a time base reference identifier; S200. Based on the session identifier record structure, perform received baseband sampling acquisition, pilot observation summary extraction and time alignment processing to generate aligned data packets and a quality tag table; wherein, the quality tag table includes a session identifier field, an alignment index field, a probe identifier field, a quality tag field, an anomaly source reference field and a supplementary sampling request reference field; S300. Based on the aligned data packet and quality label table, perform perception observation summary generation, candidate parameter set extraction and evidence slot record set assembly processing to generate evidence slot record set and prior constraint package; wherein, the prior constraint package includes session identifier field, time anchor point identifier field, prior rule set version identifier field, slot template version identifier field, prior confidence registration result field, constraint trigger field set and linkage review index field reference field; S400, based on the evidence slot record set and prior constraint package, performs joint channel parameter estimation, uncertainty record extraction and pilot reconfiguration suggestion generation processing to generate detection strategy configuration.

[0004] Furthermore, the process of registering the field mapping table includes: The session configuration generation process includes: the topology access submodule performing structure verification on the network topology; the reference signal registration submodule performing reference signal resource consistency verification on the reference signal list; the probe registration submodule performing probe capability caliber unification processing on the probe list; the budget registration submodule writing resource budget parameter groups into the budget record field; and the field mapping table registration submodule forming a field mapping table. The field mapping table includes field name specifications, field type specifications, unit specifications, timestamp caliber specifications, and default value caliber specifications.

[0005] Furthermore, the process of unified time base registration and session identifier registration includes: The unified time reference registration process includes: the time source access submodule selecting the clock source access path based on the time synchronization capability description of the probe list; the time synchronization verification submodule performing multi-probe timestamp consistency verification and reference signal observation window consistency verification; the drift monitoring submodule entering continuous monitoring state after passing the time synchronization verification; and the time reference record generation submodule generating a time reference record, which includes a time anchor point identifier, clock source type description, time synchronization status summary, time synchronization verification result summary, drift monitoring status summary, and timestamp caliber specification reference identifier. The session identifier registration process includes: a session identifier generation submodule calling the time anchor identifier in the time base record and the network domain identifier information in the network topology to generate a session identifier; a session identifier conflict verification submodule performing comparison verification with the historical session identifier archive domain and conflict window verification of concurrent sessions within the same network domain; a session identifier record structure generation submodule generating a session identifier record structure; and a session state binding submodule binding the session identifier record structure to the running state domain of the session configuration package.

[0006] Furthermore, the process of performing baseband sampling acquisition includes: The receiving baseband sampling acquisition and processing includes: a sampling triggering submodule forming sampling triggering rules based on the period and triggering conditions of the reference signal list; a sampling window orchestration submodule generating sampling window orchestration results from the network topology, reference signal list, and probe list; an RF front-end access submodule and a baseband sampling link access submodule performing link access processing; a sampling quality pre-inspection submodule performing sampling continuity checks, timestamp monotonicity checks, and field parsing availability checks; and a receiving baseband sampling packet encapsulation submodule encapsulating the receiving baseband sampling packet. The receiving baseband sampling packet includes a session identifier field, a time anchor identifier field, a network topology snapshot identifier field, a reference signal list version identifier field, a probe identifier field, a sampling window orchestration result field, a sampling granularity description field, a sampling data payload field, and a pre-inspection anomaly reference field.

[0007] Furthermore, the process of pilot observation summary extraction and time alignment includes: The pilot observation summary extraction and time alignment processing includes: an observation window positioning submodule locating sampling segments in the sampling window arrangement result field of the received baseband sampling packet based on the time-frequency resource location description of the reference signal list; a summary extraction submodule extracting pilot observation summaries from the located sampling segments; a time alignment submodule mapping the pilot observation summaries to a unified time reference to generate an alignment index; an alignment consistency verification submodule performing cross-probe alignment slot coverage consistency verification and alignment index continuity verification; and an alignment data packet encapsulation submodule encapsulating and generating alignment data packets. The alignment data packets include a session identifier field, a time anchor identifier field, an alignment index field, a pilot observation summary set field, an alignment deviation record field, and an alignment anomaly reference field.

[0008] Furthermore, the process of generating a summary of sensory observations includes: The sensing observation summary generation process includes: an input access unit performing packet header parsing and summary header parsing on the aligned data packet and pilot observation summary; an alignment index parsing unit mapping the alignment index field to a sensing time window index; a sensing resource orchestration unit generating a sensing resource orchestration result; a sensing feature extraction unit extracting a sensing feature summary according to the sensing resource orchestration result; a cross-source consistency judgment unit performing consistency judgment on the sensing observation summary and pilot observation summary; and a summary encapsulation unit encapsulating the sensing observation summary to obtain the sensing observation summary. The sensing observation summary includes a sensing time window index field, a sensing probe identifier field, a sensing feature summary field, a sensing confidence clue field, and an anomaly marker field.

[0009] Furthermore, the process of candidate parameter set extraction and evidence slot record set assembly includes: The candidate parameter set extraction and evidence slot record set assembly process includes: a candidate parameter extraction unit extracting candidate parameter clues from pilot observation summaries and perception observation summaries to generate a candidate parameter set; a candidate parameter deduplication unit performing deduplication and merging processing on the candidate parameter set; a slot template loading unit loading slot templates; a slot assembly unit mapping pilot observation summaries and perception observation summaries to evidence slots; a slot conflict merging unit performing conflict merging processing; a slot quality binding unit binding a quality mark table to the evidence slots; and an evidence slot record set encapsulation unit encapsulating and generating an evidence slot record set. The evidence slot record set includes a session identifier field, a time anchor identifier field, a slot template version identifier field, a slot number field, a slot type field, a set of slot input fields, a quality mark reference field, and a conflict summary field.

[0010] Furthermore, the process of performing joint channel parameter estimation includes: The joint channel parameter estimation process includes: an input access unit reading the prior constraint packet, the evidence slot record set, and the candidate parameter set and performing consistency comparison; a caliber alignment unit performing version alignment judgment on the slot template version identifier field and the prior rule set version identifier field; an evidence screening and gating unit performing gating screening processing on the evidence slot record set based on the prior confidence registration result field; a parameter entry arrangement unit performing estimation target arrangement processing on the candidate parameter set; a joint estimation core unit performing joint channel parameter estimation processing on the estimation batch unit, the evidence gating mark table, and the parameter-evidence binding record; an uncertainty generation unit generating uncertainty records; and a posterior encapsulation unit encapsulating the channel posterior packet, which includes a session identifier field, a time anchor identifier field, a parameter posterior record field, an uncertainty record field, an evidence adoption summary field, a conflict source explanation field, a delay aggregation reference field, and a version identifier field.

[0011] Furthermore, the process of uncertainty record extraction and pilot reconfiguration suggestion generation includes: The uncertainty record extraction and pilot reconfiguration suggestion generation process includes: a posterior access unit reading channel posterior packets and establishing an uncertainty index; an uncertainty analysis unit performing trigger interpretation processing on the uncertainty index to generate a reconfiguration trigger list; a reference signal entry mapping unit mapping parameter identifiers in the reconfiguration trigger list to a set of reference signal entry identifiers; a reconfiguration candidate generation unit generating a reconfiguration candidate set based on the reconfiguration mapping record; a budget constraint gating unit performing gating and pruning processing on the reconfiguration candidate set based on the resource budget parameter group; and a suggestion encapsulation unit encapsulating and generating a pilot reconfiguration suggestion. The pilot reconfiguration suggestion includes a session identifier field, a version identifier field, a reconfiguration action entry set field, an effective window summary field, a return field list summary field, a field pruning mark field, and a conflict resolution record field.

[0012] Furthermore, the process of generating the detection strategy configuration includes: The detection strategy configuration generation process includes: a suggestion access unit reading pilot reconfiguration suggestions and generating a configuration version number; an action parsing unit performing action parsing processing on each reconfiguration action item set field and mapping it to detection strategy atomic actions; a strategy template loading unit loading the detection strategy template; a strategy orchestration unit writing the detection strategy atomic actions into the strategy action orchestration field; a strategy consistency verification unit performing consistency verification processing on the strategy action orchestration field and the strategy feedback field mapping field; a strategy delivery preparation unit generating the delivery payload; and a configuration encapsulation and archiving unit encapsulating and generating the detection strategy configuration. The detection strategy configuration includes a session identifier field, a configuration version number field, a detection strategy template version identifier field, a strategy action orchestration field, a strategy feedback field mapping field, an exception handling branch field, a rollback condition field, and a delivery payload reference field.

[0013] The key innovations of this invention include: (1) In S100, the network topology, reference signal list, probe list and resource budget parameter group are uniformly incorporated into the field mapping table registration, unified time base registration and session identifier registration processing link, and the session identifier record structure is used as the only entry carrier of S200 to realize the caliber constraint and traceability connection of subsequent baseband sampling acquisition, pilot observation summary extraction and time alignment processing.

[0014] (2) In S300, with the alignment data packet and quality label table as common input, a combined link of sensing observation summary generation, candidate parameter set extraction and evidence slot record set assembly is adopted to gather the key information of pilot observation summary and sensing observation summary in a structured manner of evidence slot record set, and simultaneously generate prior constraint packet for prior confidence registration of subsequent joint channel parameter estimation.

[0015] (3) In S400, joint channel parameter estimation is performed using the evidence slot record set and prior constraint package as constraint carriers. Based on the channel posterior package, uncertainty record extraction, pilot reconfiguration suggestion generation and detection strategy configuration generation are integrated to ensure that the pilot reconfiguration suggestion and detection strategy configuration are derived from the same joint estimation link and maintain the link consistency related to the resource budget parameter set.

[0016] The following are its main beneficial effects: (1) To address the link breakage problem caused by inconsistent field mapping table registration and inconsistent time reference registration and time alignment link in the existing scheme, the results of field mapping table registration and unified time reference registration are carried by the session identifier record structure and used as the input of S200. This enables the receiving baseband sampling acquisition, pilot observation summary extraction and time alignment processing to run in the same session identifier context, reducing the drift of the field mapping table when referencing across steps, and providing a verifiable link basis for the subsequent generation of alignment data packets and quality mark table.

[0017] (2) To address the problem of insufficient correlation between the alignment data packet and the quality mark table in the existing scheme and the discontinuity of the fusion link caused by the dispersed evidence organization method, the pilot observation summary and the sensing observation summary are aggregated into a unified evidence slot record set by assembling the evidence slot record set according to the candidate parameter set. The prior confidence in the evidence slot record set is registered with the prior constraint packet, so that the subsequent joint channel parameter estimation can call the alignment data packet and the quality mark table under the same evidence organization caliber, reducing the conflict and gap tracking difficulties in the fusion stage.

[0018] (3) To address the problem that existing technologies lack traceable constraints between joint channel parameter estimation, uncertainty record extraction and pilot reconfiguration suggestion generation, which can lead to link breaks in the detection strategy configuration generation path, this paper proposes to constrain joint channel parameter estimation with a priori constraint packets and generate channel posterior packets. Then, uncertainty records are extracted from the channel posterior packets and pilot reconfiguration suggestions and detection strategy configurations are derived. This ensures that the generation basis of pilot reconfiguration suggestions and the configuration source of detection strategy configurations are consistent, reducing ambiguity when the detection strategy configuration references uncertainty records. This is suitable for operation scenarios where resource budget parameter group constraints need to run through the entire process. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating a 5G-oriented converged sensing communication signal processing method provided in an embodiment of this application. Detailed Implementation

[0020] Example 1: Refer to Figure 1 This is a flowchart illustrating a 5G-oriented converged sensing communication signal processing method according to an embodiment of the present invention. The flowchart may include at least steps S100-S400: S100: Obtain network topology, reference signal list, probe list and resource budget parameter group, perform field mapping table registration, unified time base registration and session identifier registration processing, and generate session identifier record structure; S200: Based on the session identifier record structure, perform received baseband sampling acquisition, pilot observation summary extraction and time alignment processing to generate aligned data packets and quality tag tables; S300: Based on the aligned data packet and quality label table, perform perception observation summary generation, candidate parameter set extraction and evidence slot record set assembly processing to generate evidence slot record set and prior constraint package; S400, based on the evidence slot record set and prior constraint package, performs joint channel parameter estimation, uncertainty record extraction and pilot reconfiguration suggestion generation processing to generate detection strategy configuration.

[0021] Step S100 includes at least steps S110-S130: S110. Obtain the network topology, reference signal list, probe list and resource budget parameter group, perform field mapping table registration processing, and obtain the session configuration package; In a converged sensing communication signal processing scenario for fifth-generation (5G) mobile communication technology, the network topology is provided by a network-side management entity or orchestration control entity. Its content is defined as a structured description of the network element connection and bearer relationships related to the communication signal processing link, including at least: the logical connection relationship between access-side nodes and core network-side nodes, the adjacency relationship between adjacent cells or adjacent coverage areas, the bearer identifiers of backhaul links and fronthaul links, the role identifiers of nodes participating in converged sensing, and the reporting path identifiers related to signal processing. The reference signal list is defined as a structured description of a set of reference signal resources used for channel observation and synchronization observation within the network topology coverage area. The reference signal resources include at least a reference signal type identifier, a time-frequency resource location description, a period and trigger condition description, a transmitter association identifier, a receiver observation granularity description, and a resource occupation description. Among them, the reference signal type identifier is used to distinguish the observation source on which the pilot observation summary is relied upon in the subsequent generation stage; the time-frequency resource location description is used to ensure consistent positioning of the sampling window and the extraction window in the subsequent baseband sampling acquisition stage; the period and trigger condition description is used to support automatic updates during continuous session operation; and the transmitter association identifier and the receiver observation granularity description are used to constrain the observation responsibilities of each probe in the probe list. The probe list is defined as a structured description of a set of probes participating in communication signal processing and fusion sensing observation under the network topology. Each probe is an entity carrying a receiving link or sensing observation link, and its implementation includes base station-side receiving link probes, user-side receiving link probes, edge-side sensing access probes, or backhaul-side aggregation probes. Each probe entry includes at least a probe identifier, its owning node identifier, an installation or deployment location description, an RF front-end capability description, a baseband sampling link capability description, a time synchronization capability description, and a reporting channel description. The resource budget parameter set is defined as a set of resource boundaries available during session operation at the time slot or subframe level. It includes at least one or more of the following: computational resource budget, storage resource budget, reporting bandwidth budget, and power consumption budget. It also includes a budget activation window and budget adjustment trigger conditions to support subsequent steps in constraining the acquisition granularity, alignment granularity, and candidate parameter set size.

[0022] In its specific implementation, S110 is executed by the session configuration generation module, which consists of a topology access submodule, a reference signal registration submodule, a probe registration submodule, a budget registration submodule, and a field mapping table registration submodule. When the topology access submodule receives the network topology, it first performs a structural verification, including node identifier uniqueness verification, adjacency closure verification, and bearer identifier resolvability verification. When a node identifier conflict or adjacency break is detected, the topology access submodule writes the conflict record to the abnormal record field within the session and triggers a topology re-fetch action. Simultaneously, it writes the current topology fetch time to a metadata field that can be referenced in subsequent session identifier registration phases. When the reference signal registration submodule receives the reference signal list, it performs a reference signal resource consistency check. This check includes verifying whether the time-frequency resource location is compatible with the network topology's bearer configuration, whether the transmitter association identifier is resolvable in the network topology, and whether the observation granularity description meets the resource budget parameter group constraints. If the consistency check fails, the reference signal registration submodule marks the abnormal entry as a pending review entry and still includes it in the registration field of the reference signal list, writing it into the entry-level status field for reference in the quality marking process during subsequent alignment. When the probe registration submodule receives the probe list, it performs probe capability standardization processing. This processing includes merging capability fields reported by different equipment manufacturers into a unified field space, mapping the RF front-end capability description and baseband sampling link capability description to comparable capability levels, and mapping the time synchronization capability description to a unified time reference registration set of accessible clock source types. When a probe entry lacks a key capability field, the probe registration submodule writes the missing field into the gap record field and triggers a capability supplementation request. This request can be pushed by the network-side management entity or actively reported by the probe side. The budget registration submodule writes the resource budget parameter group into the budget record field, and registers the budget effective window and budget adjustment trigger conditions. The budget adjustment trigger conditions include at least one of link congestion trigger, packet loss increase trigger, or computing load increase trigger, so that the session has an automatic iteration trigger entry in the subsequent pilot reconfiguration suggestion generation stage.

[0023] The field mapping table registration submodule of S110 is used to generate the field mapping table. The field mapping table is defined as a set of data field alignment rules across network elements, probes, and links, and includes at least field name specifications, field type specifications, unit specifications, timestamp specifications, and default value specifications. During the field mapping table registration process, the submodule extracts key fields from the network topology, reference signal list, probe list, and resource budget parameter group to form field dictionary entries. Each field dictionary entry includes at least a field identifier, field source identifier, field parsing rules, and field validity constraints. Subsequently, the submodule performs conflict resolution processing on the field dictionary entries. Conflict resolution processing includes renaming synonymous and heteronymous fields, merging synonymous and heteronymous fields, and registering unit conversion rules. When an unresolvable field conflict occurs, the submodule writes the conflict into the conflict record field and marks the conflict entry as requiring manual review or secondary data collection and verification. Simultaneously, it writes the associated probe identifier of the conflict entry into the conflict index field that can be referenced in the subsequent quality mark table. After completing the field mapping table registration process, the session configuration generation module encapsulates the network topology, reference signal list, probe list, resource budget parameter group, and field mapping table into a session configuration package. This session configuration package is invoked as input in S120 and subsequently input as a "session configuration package" in S210 for receiving baseband sampling and acquisition processing. Simultaneously, the field mapping table is input as a "field mapping table" in S310 for sensing observation summary generation processing. To support version management strategies, the session configuration package includes a configuration version identifier and an effective window identifier. The configuration version identifier is automatically incremented by the configuration generation module with each change, and the effective window identifier is written by the budget registration submodule based on the budget effective window of the resource budget parameter group. When the reference signal list or probe list changes and triggers re-registration, the session configuration generation module stores the historical version in the version archive field and writes the version switch time and trigger reason record, ensuring a stable and auditable operating path for the invention during its evolution.

[0024] S120. Extract the reference signal list and probe list from the session configuration package, register them using a unified time base, and generate a time base record. In S120, the unified time reference registration is performed by the time reference registration module, which consists of a time source access submodule, a time synchronization verification submodule, a drift monitoring submodule, and a time reference record generation submodule. The input source for S120 is the session configuration packet obtained in S110. After reading the session configuration packet, the time reference registration module first extracts a reference signal list and a probe list from it. The extraction process uses field parsing rules constrained by a field mapping table to avoid discrepancies in the parsing of fields from different sources. Subsequently, the time source access submodule selects the clock source access path based on the time synchronization capability description registered in the probe list. In one engineering embodiment, the network-side management entity sends time synchronization configuration to the probe side, and the probe side accesses the clock source from the network time synchronization link provided by the base station side and reports the time synchronization status to the time source access submodule; in another engineering embodiment, the edge-side aggregation probes access the local high-stability clock source and publish the timestamp caliber to each probe through the timestamp caliber specification in the session configuration package to form a unified time anchor point within the session.

[0025] The time synchronization verification submodule performs time synchronization verification after the clock source is connected. This verification includes multi-probe timestamp consistency verification and reference signal observation window consistency verification. Multi-probe timestamp consistency verification is performed by reading the time synchronization status reporting field and the local sampling timestamp field of each probe. When a probe's timestamp deviates from the session time anchor point beyond the allowable range, the time synchronization verification submodule marks the probe as a probe to be reviewed and writes the reason for review into the review field, simultaneously triggering a time synchronization retry for that probe. Reference signal observation window consistency verification is performed by reading the period and trigger condition descriptions registered in the reference signal list and combining them with the baseband sampling link capability descriptions registered in the probe list to construct an observation window reachability determination. When it is found that the probe's sampling link capability cannot cover the observation window in the reference signal list, the time synchronization verification submodule writes the mismatch relationship into the mismatch record field and associates this mismatch relationship with the status field of the reference signal entry, for subsequent quality marking processing in S230 to reference when generating the quality mark table. After the time synchronization verification is successful, the drift monitoring submodule enters the continuous monitoring state. The continuous monitoring state is driven by the budget effective window and budget adjustment trigger conditions in the session configuration package. When link congestion or increased packet loss is detected, causing the time synchronization state to be unstable, the drift monitoring submodule records the drift trend and triggers a time base update request to ensure that the session maintains a consistent time caliber in subsequent steps.

[0026] The time reference record generation submodule generates a time reference record after completing the unified time reference registration. The time reference record is defined as a structured carrier of the time caliber within a session, containing at least a time anchor identifier, clock source type description, time synchronization status summary, time synchronization verification result summary, drift monitoring status summary, and timestamp caliber specification reference identifier. The time reference record is written into the session running state field of the session configuration generation module at the end of S120 and is invoked as input to the subsequent S210. Specifically, in S210, the time reference record is used as the "time reference record" input to receive baseband sampling and acquisition processing, thereby ensuring that subsequent sampling timestamps and alignment timestamps follow the same caliber. Simultaneously, the time reference record is implicitly referenced in the time alignment processing of the subsequent S220, inheriting the timestamp caliber specification of the time reference record through the timestamp field of the aligned data packet. Furthermore, the extraction results of the reference signal list and probe list by S120 are retained in the runtime shadow domain of the session configuration package, so that the version of the reference signal list entries can be verified during the subsequent pilot reconfiguration suggestion generation stage when a policy write-back occurs, thereby maintaining the stability and traceability of version evolution during the continuous operation of the session.

[0027] S130. Perform session identifier registration processing on the session configuration package to generate a session identifier record structure; In S130, the session identifier registration process is executed by the session identifier registration module, which consists of a session identifier generation submodule, a session identifier conflict verification submodule, a session identifier record structure generation submodule, and a session state binding submodule. The input source for S130 is also the session configuration package obtained in S110, and it inherits the time base record generated in S120 on the runtime link. Specifically, after reading the session configuration package, the session identifier generation submodule calls the time anchor point identifier and time synchronization status summary in the time base record, and simultaneously calls the network domain identifier information in the network topology, and combines it with the session initiation trigger event information to generate the session identifier; wherein, the session initiation trigger event information is defined as the event carrier that triggers session creation, including a manual trigger identifier or an automatic trigger identifier, a trigger source identifier, and a trigger time record. When generating session identifiers, the session identifier generation submodule uses irreversible digest operation to combine the session startup time segment, network domain identifier segment, and random number segment. The random number segment is generated by the session identifier generation submodule from the local entropy source at startup and registered in the random source record field, thus enabling the session identifier to distinguish across sessions. At the same time, to avoid introducing unstable calibers in the session identifier generation process, the session identifier generation submodule writes the configuration version identifier and effective window identifier from the session configuration package into the metadata field that can be referenced by the session identifier record structure generation submodule, so that the session identifier has a version clue that can be associated when the configuration evolves.

[0028] The session identifier conflict verification submodule performs conflict verification processing after the session identifier is generated. This processing includes comparison with the historical session identifier archive and conflict window verification of concurrent sessions within the same network domain. When a conflict is detected, the conflict verification submodule triggers session identifier regeneration, writes the conflict cause into the conflict cause record, and simultaneously writes the network topology snapshot identifier related to the conflict window into the session identifier record structure for subsequent auditing and backtracking. The session identifier record structure generation submodule generates the session identifier record structure after successful conflict verification. This structure is defined as a structured carrier of session identity and traceable metadata, including at least the session identifier, session start time record, network topology snapshot identifier, reference signal list version identifier, probe list version identifier, resource budget parameter group version identifier, field mapping table version identifier, and time base reference identifier. The session state binding submodule binds the session identifier record structure to the runtime state field of the session configuration package, enabling subsequent steps to write the session identifier into the session field of the corresponding data structure when generating received baseband sampling packets, aligned data packets, evidence slot record sets, or prior constraint packets, thus forming a traceable link across steps.

[0029] At the end of S130, the session identifier record structure is output and used by subsequent steps. Specifically, in S210, when performing the received baseband sampling acquisition process, the session identifier record structure is used as input as a "session identifier record structure," and the session identifier is written into the session field of the received baseband sampling packet. Simultaneously, the session identifier record structure is inherited in the subsequent S220 when generating the alignment data packet, thus ensuring that the time alignment result of the alignment data packet can be associated with the same session identifier. Furthermore, the session identifier record structure and the aforementioned time reference record form a consistent reference relationship, both serving as baseline constraints for the continuous operation and automatic triggering of updates in the S200 phase. This allows the subsequent S300 perception observation summary generation process and the S400 joint channel parameter estimation process to perform field mapping, time alignment, and evidence assembly operations under the same session caliber.

[0030] In summary, the technical effects of this step are as follows: By serializing the registration of field mapping tables, unified time base registration, and session identifier registration, the session configuration package, time base record, and session identifier record structures form a consistent standard and support version archiving during the cross-probe and cross-link data access phase. This reduces the probability of field ambiguity and time standard drift in subsequent baseband sampling and time alignment processing, and provides a traceable anchor point for data structure association and audit backtracking in subsequent steps.

[0031] Step S200 includes at least steps S210-S230: S210. Obtain the session configuration packet, time base record and session identifier record structure, perform received baseband sampling acquisition processing, and obtain the received baseband sampling packet; S210 inherits the session configuration package generated by S110, the time reference record generated by S120, and the session identifier record structure generated by S130. The session configuration package serves as the configuration baseline for data acquisition and access in this stage, defining a unified standard for network topology, reference signal list, probe list, resource budget parameter group, and field mapping table. The time reference record serves as the sole basis for timestamp standards and time synchronization status in this stage, defining the intra-session time anchor identifier, clock source type description, time synchronization status summary, and drift monitoring status summary. The session identifier record structure serves as the sole carrier for data structure association and traceable registration in this stage, defining the session identifier, session start time record, network topology snapshot identifier, reference signal list version identifier, probe list version identifier, resource budget parameter group version identifier, and field mapping table version identifier. Specifically, S210 is executed by the receiving baseband sampling acquisition module, which consists of a sampling trigger submodule, a sampling window arrangement submodule, an RF front-end access submodule, a baseband sampling link access submodule, a sampling quality pre-inspection submodule, and a receiving baseband sampling packet encapsulation submodule. It is used to uniformly collect and encapsulate the baseband sampling data of the receiving link in the fusion sensing scenario, and write the session dimension identification information and time caliber information into the sampling product.

[0032] In the implementation process, the sampling triggering submodule first reads the reference signal list in the session configuration package and forms a sampling triggering rule based on the period and triggering conditions registered in the reference signal list. The sampling triggering rule includes at least one or more of the following: reference signal triggering, event triggering, and polling triggering. Reference signal triggering is defined as triggering sampling when an entry in the reference signal list reaches its registered transmission period or meets its registered triggering condition. Event triggering is defined as triggering sampling window re-arrangement and sampling granularity adjustment when the network-side management entity or edge-side control entity reports link congestion triggering, packet loss increase triggering, or computational load increase triggering. Polling triggering is defined as performing periodic sampling within the session effective window according to the budget effective window registered by the resource budget parameter group. When the sampling triggering submodule performs the trigger determination, it calls the time anchor point identifier and time synchronization status summary in the time reference record and writes the trigger determination into the trigger record field. When the time synchronization status summary shows that it is in the pending review state, the sampling triggering submodule downgrades the trigger action to the conservative sampling mode. The conservative sampling mode is defined as a sampling mode that reduces the number of sampling coverage windows or reduces the sampling duration, and writes the downgraded state into the sampling status field of the subsequently received baseband sampling packets.

[0033] The sampling window orchestration submodule performs sampling window orchestration processing after the trigger is established. The sampling window is defined as a combination of a time window and a frequency domain window used to extract baseband samples from the receiving link. Its orchestration is based on the network topology, reference signal list, and probe list from the session configuration package. Specifically, the sampling window orchestration submodule parses the receiving path identifier and adjacency relationship of adjacent coverage areas from the network topology, parses the time-frequency resource location description of reference signal resources from the reference signal list, parses the baseband sampling link capability description and reporting channel description of each probe from the probe list, and combines this with the computational resource budget and reporting bandwidth budget in the resource budget parameter group to generate the sampling window orchestration result. The sampling window orchestration result includes a sampling window identifier, probe identifier, sampling start and end time description, sampling frequency domain range description, and sampling granularity description, where the sampling granularity description is used to constrain the summary refinement in the subsequent pilot observation summary extraction stage. If, during the sampling window orchestration process, it is found that the baseband sampling link capability of a probe cannot cover the target time-frequency resource location description in the reference signal list, the sampling window orchestration submodule writes the mismatch relationship into the mismatch record field and marks the corresponding sampling window as a gap sampling window; the gap sampling window can be used as an input clue for the quality marking table in subsequent quality marking processing.

[0034] After the sampling window orchestration is completed, the RF front-end access submodule and the baseband sampling link access submodule perform link access processing. The RF front-end access submodule selects the corresponding RF front-end access configuration and sends out the gain, bandwidth, and filtering configurations based on the RF front-end capability descriptions registered in the probe list. The baseband sampling link access submodule configures the sampling buffer, sampling queue, and reporting channel based on the baseband sampling link capability descriptions registered in the probe list, and matches the sampling buffer capacity and queue depth with the storage resource budget in the resource budget parameter group. To support automated adjustments during continuous session operation, when an event triggers a change in the resource budget parameter group, the baseband sampling link access submodule performs dynamic reconfiguration. Dynamic reconfiguration includes expanding or shrinking the sampling buffer and rearranging the sampling queue priority, and records the reconfiguration reason and reconfiguration time in a reconfiguration record field that can be referenced by the receiving baseband sampling packet encapsulation submodule.

[0035] The sampling quality pre-inspection submodule performs sampling quality pre-inspection processing after the sampling data is written to the buffer. This processing includes at least a sampling continuity check, a timestamp monotonicity check, and a field parsing availability check. The sampling continuity check identifies whether there are missing or duplicate segments within the sampling window. The timestamp monotonicity check identifies whether the sampling timestamps conform to the timestamp caliber specifications of the time reference record. The field parsing availability check identifies whether the sampling data can be parsed according to the field type and unit specifications of the field mapping table. For detected abnormal segments, the sampling quality pre-inspection submodule writes the sampling window identifier, probe identifier, abnormality type, and abnormal time range of the abnormal segment into the pre-inspection abnormality record field, and writes the reference identifier of the pre-inspection abnormality record field into the pre-inspection field of the received baseband sampling packet. This enables S220 to avoid or mark abnormal segments at the summary layer when extracting pilot observation summaries.

[0036] The receiving baseband sampling packet encapsulation submodule performs encapsulation processing after the sampling quality pre-check is completed, resulting in a receiving baseband sampling packet. The receiving baseband sampling packet is defined as a structured data packet carrying the received link baseband sampling data and its session metadata, and includes at least: a session identifier field, a time anchor identifier field, a network topology snapshot identifier field, a reference signal list version identifier field, a probe identifier field, a sampling window orchestration result field, a sampling granularity description field, a sampling data payload field, and a pre-check anomaly reference field. The session identifier field is provided and written by the session identifier record structure, the time anchor identifier field is provided and written by the time reference record, and the sampling window orchestration result field is generated and written by the sampling window orchestration submodule, thus ensuring that the receiving baseband sampling packet maintains consistency and traceability during cross-step transmission. The received baseband sampling packet output in S210 is subsequently used as input in S220. Specifically, in S220, the received baseband sampling packet is used as input for pilot observation summary extraction and time alignment processing. Simultaneously, the pre-detection anomaly reference field and sampling window arrangement result field of the received baseband sampling packet are indirectly referenced in the subsequent quality marking process in S230, supporting the generation of the quality marking table and the location of anomaly sources. Furthermore, the received baseband sampling acquisition module synchronously writes the trigger record field, reconfiguration record field, and pre-detection anomaly record field of this sampling into the session running state field, enabling the session to reference historical sampling contexts when policy write-back occurs during the pilot reconfiguration suggestion generation stage, forming a stable connection across main steps.

[0037] S220. Extract pilot observation summaries from the received baseband sampling packets, perform time alignment processing, and generate aligned data packets; In S220, pilot observation summary extraction and time alignment processing are performed by the pilot observation alignment module. This module consists of a summary extraction submodule, an observation window positioning submodule, a time alignment submodule, an alignment consistency verification submodule, and an alignment data packet encapsulation submodule. The input source for S220 is the received baseband sampling packet obtained in S210. When the pilot observation alignment module receives the received baseband sampling packet, it first reads the session identifier field and the time anchor identifier field, and then retrieves the corresponding version of the reference signal list from the session running state domain based on the reference signal list version identifier field, thereby binding the extracted pilot observation summary to the version caliber configured in the session. The observation window positioning submodule locates the sampling segment corresponding to the reference signal in the sampling window arrangement result field of the received baseband sampling packet based on the time-frequency resource location description, period, and trigger condition description in the reference signal list. It also performs avoidance marking on the abnormal segments indicated by the gap sampling window and the pre-detection anomaly reference field. The avoidance marking is defined as the processing method of skipping, truncating, or marking abnormal segments at the summary layer. The specific method adopted can be constrained by the field validity constraints and resource budget parameter group in the session configuration packet.

[0038] After completing the observation window localization, the summary extraction submodule extracts pilot observation summaries from the located sampling segments. The pilot observation summary is defined as a structured compressed representation of pilot-related observation information, and its minimum set includes at least: observation reference signal entry identifier, observation probe identifier, observation timestamp summary, observation frequency range summary, observation energy or correlation summary, and observation vector summary related to subsequent channel parameter estimation. The observation timestamp summary inherits the timestamp specification from the time anchor identifier field; the observation frequency range summary inherits the frequency range description from the sampling window arrangement result field; the observation energy or correlation summary is used to preliminarily characterize observation availability; and the observation vector summary provides correlated observation input for subsequent sensing observation summary generation processing in S310 and prior confidence registration processing in S330. In terms of implementation, the summary extraction submodule can adopt two schemes: Scheme 1 is a window-aggregation-based summary extraction scheme, which segments and aggregates the located sampling fragments according to the sampling granularity description, outputs the energy or correlation summary of each segment, and forms an observation vector summary; Scheme 2 is an event boundary-based summary extraction scheme, which identifies the observation boundary according to the triggering condition description of the reference signal and outputs the observation vector summary within the boundary. Both schemes include an anomaly marker field in the summary. The anomaly marker field is generated by the avoidance marker and is used to indicate the source of the anomaly in the subsequent quality marking processing stage.

[0039] The time alignment submodule performs time alignment processing after the pilot observation summary is generated. This time alignment processing is defined as the alignment process of mapping pilot observation summaries from different probes and sampling windows to a unified time reference. Specifically, the time alignment submodule reads the time anchor identifier field and sampling timestamp field from the received baseband sampling packet, and generates an alignment deviation record by combining the time synchronization status summary and drift monitoring status summary registered in the time reference record. Subsequently, the time alignment submodule performs timestamp shifting and alignment index generation on each pilot observation summary. The alignment index is defined as the index identifier that assigns the observation to a discrete alignment slot on the session time axis, and its granularity is jointly limited by the measurement granularity and sampling granularity description registered in the session configuration packet. If the alignment deviation record indicates that a probe has a drift trend or is in a state awaiting verification, the time alignment submodule marks the pilot observation summary corresponding to that probe as a low-confidence observation summary and writes the drift information into a verification context field that can be referenced by the alignment consistency verification submodule, for subsequent quality labeling processing to use when calculating the quality label table.

[0040] The alignment consistency verification submodule performs consistency verification processing after time alignment is completed. This consistency verification process includes at least two parts: cross-probe alignment slot coverage consistency verification and alignment index continuity verification. Cross-probe alignment slot coverage consistency verification identifies whether there are missing probes or insufficient observations under the same alignment index; if so, it is written to the gap alignment record field. Alignment index continuity verification identifies whether there are jumps or rollbacks in the alignment index; if so, it is written to the alignment anomaly record field. The alignment data packet encapsulation submodule encapsulates and generates alignment data packets after consistency verification. The alignment data packet is defined as a structured data packet carrying pilot observation summaries and alignment indexes, and includes at least a session identifier field, a time anchor identifier field, an alignment index field, a pilot observation summary set field, an alignment deviation record field, and an alignment anomaly reference field. The alignment data packet is output at the end of S220 and used in subsequent steps. This alignment data packet serves as input to S230 for quality marking processing and as input to S310 for sensing observation summary generation processing. Simultaneously, the pilot observation summary is used as input to S310, specifically as the "pilot observation summary" input for sensing observation summary generation processing in S310. Further, the alignment data packet encapsulation submodule synchronously writes the alignment index field and alignment deviation record field into the session runtime state domain, enabling S420 to reference the historical alignment deviation context when extracting uncertainty records from the channel posterior packet, supporting the closed-loop connection of the pilot reconfiguration suggestion generation process.

[0041] S230. Perform quality marking processing on the aligned data packets to generate a quality marking table; In S230, the quality marking process is executed by the quality marking module, which consists of a quality rule loading submodule, a quality feature extraction submodule, a quality judgment submodule, a quality mark table generation submodule, and a quality log archiving submodule. The input source for S230 is the alignment data packet output from S220. After receiving the alignment data packet, the quality marking module first reads the session identifier field and alignment anomaly reference field from the alignment data packet, and then retrieves the field validity constraints and budget adjustment trigger conditions from the session configuration packet based on the session identifier field, using these as rule inputs for the quality rule loading submodule. The quality rules are defined as a set of rules for the availability judgment and gap handling of the alignment data packet. The minimum set includes at least: time alignment deviation tolerance, alignment index coverage tolerance, anomaly fragment avoidance, and gap re-collection triggering; wherein the gap re-collection triggering is associated with the budget adjustment triggering condition and is used to adjust the re-collection strategy and intensity when link congestion or packet loss increases. The quality rule loading submodule writes the loaded quality rules into the rule version record field and writes the identifier of the rule version record field into the context field of the subsequent quality tagging table generation submodule, so that the quality tagging process has version traceability capability.

[0042] The quality feature extraction submodule then performs quality feature extraction processing on the alignment data packet. The quality features are defined as a structured set of features used to determine the availability of the alignment data packet, and include at least: alignment deviation record features, alignment index continuity features, alignment slot coverage features, pilot observation summary anomaly marker features, and alignment anomaly reference features. The alignment deviation record features are parsed from the alignment deviation record field in the alignment data packet; the alignment index continuity features are parsed from the jump and rollback conditions of the alignment index field; the alignment slot coverage features are parsed from the probe coverage conditions of the pilot observation summary set fields under the same alignment index; the pilot observation summary anomaly marker features are parsed from the anomaly marker field in the pilot observation summary set fields; and the alignment anomaly reference features are parsed from the alignment anomaly reference field. During the extraction process, the quality feature extraction submodule follows the field type and unit specifications of the field mapping table, and writes unparseable or missing feature fields into the feature gap record field; the feature gap record field is recorded in the subsequent quality log archiving.

[0043] The quality assessment submodule performs quality assessment processing after the quality features are extracted. This quality assessment processing is defined as the process of gating and generating labels for quality features based on quality rules. Specifically, the quality assessment submodule compares the time alignment deviation tolerance with the alignment deviation record features, the alignment index coverage tolerance with the alignment slot coverage features, and the abnormal segment avoidance tolerance with the pilot observation summary abnormal label features. When any comparison triggers a quality anomaly condition, the quality assessment submodule generates a quality anomaly label and associates the quality anomaly label with the corresponding alignment index and probe identifier. Furthermore, the quality determination submodule determines whether to trigger supplementary sampling based on the gap supplementary sampling trigger caliber and alignment anomaly reference characteristics. The supplementary sampling trigger is defined as the action of sending a supplementary sampling request to the receiving baseband sampling acquisition module. The supplementary sampling request includes at least the supplementary sampling alignment index range, the supplementary sampling probe identifier, and the supplementary sampling granularity description. When the supplementary sampling trigger is established and the resource budget parameter group allows it, the quality determination submodule writes the supplementary sampling request into the supplementary sampling request record field and triggers the update of the supplementary sampling event identifier in the session running state field, so that S210 can rearrange the sampling window based on the supplementary sampling event identifier in subsequent running cycles.

[0044] The quality tag table generation submodule generates a quality tag table after quality judgment is completed. The quality tag table is defined as a structured representation of the quality tagging results of the aligned data packet at the alignment index granularity and probe granularity. It includes at least a session identifier field, an alignment index field, a probe identifier field, a quality tag field, an anomaly source reference field, and a supplementary acquisition request reference field. The anomaly source reference field is used to reference the identifier of the alignment anomaly reference field, the pre-detection anomaly reference field, or the feature gap record field. The supplementary acquisition request reference field is used to reference the identifier of the supplementary acquisition request record field. The quality tag table is output at the end of S230 and used in subsequent steps. Specifically, in S320, the quality tag table is used as input to support the assembly of the evidence slot record set and the generation of the prior constraint package. Simultaneously, the quality tagging module archives the quality log into the quality log archiving field of the session running state domain. The quality log includes at least a quality rule version record, a quality feature extraction process record, a quality judgment process record, and a supplementary acquisition request process record, enabling the pilot and detection strategy linkage update phase in S400 to reference the quality history when generating pilot reconfiguration suggestions and detection strategy configurations. Furthermore, the quality label table and the alignment data packet form an input bridge from S200 to S300 in the cross-main step connection, enabling S310 to identify low-confidence alignment slots and perform consistent labeling and processing at the summary layer when generating the perception observation summary.

[0045] In summary, the technical effects of this step are as follows: By introducing the session configuration packet, time reference record, and session identifier record structure into the received baseband sampling acquisition, and performing unified time alignment and quality marking after pilot observation summary extraction, the received baseband sampling packet and alignment data packet have consistent caliber in both the session and time dimensions. At the same time, alignment anomalies and supplementary sampling triggers are written into the quality marking table and connected with the subsequent evidence slot record assembly, thereby establishing a traceable engineering processing path for sampling gaps and alignment drift in the signal processing link of fusion sensing.

[0046] Step S300 includes at least steps S310-S330: S310. Obtain the alignment data packet, pilot observation summary and field mapping table, perform sensing observation summary generation processing, and obtain the sensing observation summary; In the communication signal processing method of fusion sensing for fifth generation (5G) mobile communication technology of the present invention, S310 is located in the fusion sensing side summary generation stage after the completion of baseband sampling acquisition, pilot observation summary extraction, time alignment and quality marking. S310 is executed by the sensing observation summary generation module, which is composed of at least an input access unit, an alignment index parsing unit, a sensing resource orchestration unit, a sensing feature extraction unit, a cross-source consistency judgment unit, a summary encapsulation unit and an operation log unit. Specifically, the alignment data packet acquired in S310 comes from the output of S220. The alignment data packet includes at least a session identifier field, a time anchor identifier field, an alignment index field, a pilot observation summary set field, an alignment deviation record field, and an alignment anomaly reference field. The pilot observation summary acquired in S310 comes from the pilot observation summary generated by S220 before or simultaneously with the generation of the alignment data packet. The pilot observation summary is used to provide the observation boundary, observation window, and observation reliability clues of the communication measurement side to the perception side summary generation. The field mapping table acquired in S310 comes from the field mapping table registered in the session configuration package in S110. The field mapping table is used to limit the field name, field type, field unit, and field default value caliber in different probes, different access points, and different reporting channels, so that the field caliber of the perception observation summary is consistent with that of the subsequent evidence slot record set assembly during encapsulation.

[0047] During the input access phase, the input access unit performs packet header parsing on the alignment data packet, reads the session identifier field and the time anchor identifier field, and retrieves the session configuration packet digest associated with the session identifier field from the session running state domain to restore the context of the current sensing observation digest generation process. Simultaneously, the input access unit performs digest header parsing on the pilot observation digest, extracts the observation reference signal entry identifier, observation probe identifier, observation timestamp digest, and anomaly marker fields, and maps these fields to internal unified field names according to the field mapping table, writing them into the pilot-side digest index table. Further, the input access unit reads the alignment anomaly reference field from the alignment data packet and associates it with the alignment index field to form an alignment anomaly index table. This alignment anomaly index table serves as the trigger input for anomaly avoidance and degradation processing in subsequent sensing resource orchestration and cross-source consistency judgment. The minimum set of trigger inputs includes the alignment index, anomaly type digest, and anomaly source reference identifier. The anomaly source reference identifier is used to locate the anomaly to an archived record in the internal record field of the alignment data packet or the session running state domain, facilitating traceable assembly of the evidence slot record set in subsequent steps.

[0048] During the alignment index parsing phase, the alignment index parsing unit performs time window partitioning on the alignment index field, mapping the alignment index field to a perception time window index. The perception time window index is defined as a discrete index structure used by the fusion perception side to merge perception observation summaries, and its granularity is consistent with or an integer multiple of the alignment index granularity of the alignment data packet. When the resource budget parameter group registered in the session configuration packet summary indicates a state of computing resource shortage, the alignment index parsing unit adjusts the granularity of the perception time window index to a coarser granularity and writes the adjustment record into the summary generation operation log. The record includes a summary of the adjustment trigger conditions and a summary of the effective alignment index range. Subsequently, the alignment index parsing unit generates a time alignment deviation label based on the alignment deviation record field. The time alignment deviation label is used to identify the deviation status of the observation under a certain perception time window index in terms of the time anchor point. The time alignment deviation label is associated with the gating conditions of the subsequent cross-source consistency judgment unit. If the time alignment deviation label is in a pending review state, the perception feature extraction under the corresponding perception time window index enters the conservative feature set generation mode. The conservative feature set generation mode is defined as a mode that reduces feature dimensions or reduces cross-source fusion terms, and the reason for the downgrade is written into the anomaly marker field of the perception observation summary.

[0049] During the sensing resource orchestration phase, the sensing resource orchestration unit performs joint orchestration processing on the pilot-side summary index table and the alignment anomaly index table to generate a sensing resource orchestration result. The sensing resource orchestration result is defined as the structured orchestration data for the sensing observation input sources, input channels, and processing branches in an Integrated Sensing and Communication (ISAC) scenario under each sensing time window index. Its minimum set must include at least a set of sensing probe identifiers, a set of sensing input channel identifiers, a set of sensing data fragment references, and a sensing processing branch identifier. Specifically, the sensing resource orchestration unit first determines the available set of sensing probe identifiers based on the probe list and network topology snapshot identifiers in the session configuration packet summary, and then filters the set of sensing input channel identifiers based on the reported bandwidth budget in the resource budget parameter group. When the number of available channels after filtering is insufficient, the sensing resource orchestration unit writes this insufficient state into the operation log and marks the corresponding sensing time window index as a sparse observation window. Subsequently, the sensing resource orchestration unit performs channel rearrangement on the sensing time window index where the anomalies occur, based on the anomaly type summary recorded in the alignment anomaly index table. The channel rearrangement includes switching to the backup probe identifier, switching to the backup reporting channel, or delaying aggregation to the adjacent sensing time window index, and writes the channel rearrangement action into the rearrangement flag field in the sensing resource orchestration result.

[0050] In the sensing feature extraction stage, the sensing feature extraction unit performs sensing observation summary generation processing on each sensing time window index according to the sensing resource orchestration result. The sensing observation summary is defined as a structured compressed expression of the fused sensing side observation information. The structured compressed expression and the pilot observation summary can be associated under the session identifier and time anchor identifier caliber, and output a unified set of field names under the constraints of the field mapping table. Specifically, the sensing feature extraction unit dereferences the sensing data segment reference set, reads the corresponding sensing data segment and performs field mapping. The field mapping includes at least field name normalization, field type normalization and default value filling. Subsequently, feature extraction is performed within each sensing time window index. Feature extraction includes at least two of the following: time domain segment aggregation, frequency domain segment aggregation and cross-channel segment consistency extraction. Among them, time domain segment aggregation is used to generate a time-series stability summary, frequency domain segment aggregation is used to generate a spectral structure summary, and cross-channel segment consistency extraction is used to generate a cross-probe consistency summary. To facilitate subsequent extraction of candidate parameter sets and assembly of evidence slot record sets, the perception feature extraction unit writes each type of summary into a predefined field when generating perception observation summaries. The predefined field includes at least a perception time window index field, a perception probe identifier field, a perception feature summary field, a perception confidence clue field, and an anomaly marker field. The perception confidence clue field is used to carry the summary expression of sparse observation window markers, rearrangement marker fields, and time alignment deviation labels.

[0051] After the perceptual features are extracted, the cross-source consistency judgment unit performs consistency judgment processing on the perceptual observation summary and the pilot observation summary. In this invention, consistency judgment processing is embodied in the logical link constraint of fused perception. This logical link constraint does not introduce formulaic expressions; instead, it writes the consistency judgment result as a structured marker that can be called upon in subsequent evidence slot record set assembly. Specifically, based on the reference signal list registered in the session configuration packet summary, the cross-source consistency judgment unit aligns the observation window of the pilot observation summary with the perceptual time window index, forming a cross-source alignment relationship record. Subsequently, the cross-source consistency judgment unit performs consistency gating on the cross-source alignment relationship record. This consistency gating includes at least a time window consistency gating and anomaly marker consistency gating. The time window consistency gating is used to determine whether the observation timestamp summary of the pilot observation summary falls within the corresponding perceptual time window index range, and the anomaly marker consistency gating is used to determine whether the anomaly marker field of the pilot observation summary and the anomaly marker field of the perceptual observation summary exhibit the same direction of anomaly. When the gating passes, the cross-source consistency judgment unit writes the cross-source consistency flag field into the perception observation summary; when the gating fails, the cross-source consistency judgment unit writes the conflict flag field and writes the conflict cause summary into the operation log. The conflict cause summary includes at least the reference identifier of the aligned abnormal reference field or the sparse observation window flag.

[0052] After completing the consistency check, the summary encapsulation unit encapsulates the sensing observation summary into a unified output product and returns it to the process controller, obtaining the sensing observation summary output in this step. The sensing observation summary is named as an output field at the end of this step and is subsequently used as an input field in S320. Specifically, S320 extracts candidate parameter sets from the pilot observation summary and the sensing observation summary and assembles the evidence slot record set. Simultaneously, the sensing observation summary and the aligned data packet form an input bridge from S300 to S400 across main steps, enabling subsequent joint channel parameter estimation processing to reference sensing-side clues at the evidence slot record set level.

[0053] Understandably, in an engineering deployment scenario, the network topology corresponds to the multi-cell coverage topology of the vehicle-road cooperative road segment, the reference signal list registers reference signal entries related to vehicle uplink measurement, the probe list registers base station-side receiving link probes and edge-side sensing probes, and the sensing data fragment reference set corresponds to the fragmented observation data reported by the edge-side sensing probes; S310 completes field mapping, fragment aggregation and cross-source consistency judgment within each sensing time window index, and outputs the generated sensing observation summary to S320 as the input of the "sensing observation summary".

[0054] S320. Extract candidate parameter sets from pilot observation summaries and sensing observation summaries, assemble evidence slot record sets, and generate evidence slot record sets. In this invention, S320 is executed by the evidence slot assembly module, which comprises at least a candidate parameter extraction unit, a candidate parameter deduplication unit, a slot template loading unit, a slot assembly unit, a slot conflict merging unit, a slot quality binding unit, and an evidence slot record set encapsulation unit. Specifically, the input sources of S320 include the pilot observation summary generated in S220 and the sensing observation summary generated in S310. When accessing the two types of summaries, S320 first performs session caliber alignment, reads the session identifier field and time anchor identifier field from the two types of summaries, and performs a consistent mapping of the field names of the two according to the field mapping table, so that the candidate parameter extraction unit can identify parameter clues that can be used for subsequent joint channel parameter estimation processing within the same field space.

[0055] During the candidate parameter set extraction stage, the candidate parameter extraction unit performs parameter cue parsing on the pilot observation summary. This parameter cue parsing includes extracting candidate parameter cue that can be used to describe the channel state from the observation reference signal entry identifier, observation frequency domain range summary, observation energy or correlation summary, and observation vector summary. Simultaneously, the candidate parameter extraction unit performs perception cue parsing on the sensing observation summary. This perception cue parsing includes extracting candidate parameter cue that can be used to describe environmental coupling and observation reliability from the sensing time window index field, sensing feature summary field, cross-source consistency marker field, and conflict marker field. The candidate parameter set is defined as a set of structured parameter entries generated from the candidate parameter cue. Each parameter entry includes at least a parameter identifier, a parameter type identifier, an association alignment index, an associated reference signal entry identifier, an associated sensing time window index, and a cue source identifier. The parameter type identifier specifies that the parameter entry belongs to one of the following: channel parameter candidate entry, noise state candidate entry, interference state candidate entry, or observation bias candidate entry. To meet the description requirement of "minimum set" in this invention, it is understood that the minimum set of candidate parameters includes at least two types of candidate entries: channel parameter candidate entries and observation bias candidate entries. The channel parameter candidate entries are derived from the observation vector digest of the pilot observation digest, and the observation bias candidate entries are derived from the conflict flag field or cross-source consistency flag field of the sensing observation digest. The noise state candidate entries and interference state candidate entries are optional extended entries, which are loaded and written into the candidate parameter set when the resource budget parameter set allows.

[0056] After the candidate parameter set is initially generated, the candidate parameter deduplication unit performs deduplication and merging processing. This deduplication and merging process includes name-based merging based on parameter identifiers and nearest-neighbor merging based on association alignment indexes. Name-based merging is used to merge entries with different sources but the same parameter identifier and records multi-source clues. Nearest-neighbor merging is used to merge parameter entries within an adjacent alignment index range and records the merging range. The triggering condition for deduplication and merging processing is jointly determined by the clue source identifier written by the candidate parameter extraction unit and the resource budget parameter group in the session configuration packet summary. When the resource budget parameter group indicates that the reported bandwidth budget is tightening, nearest-neighbor merging is performed first, and the merging granularity is adjusted to a coarser granularity. After deduplication and merging are completed, the candidate parameter deduplication unit writes the candidate parameter set as a unified output product and assigns a version identifier. This version identifier is recorded in the session running state domain and used as the reference criterion for the candidate parameter set in subsequent S410.

[0057] During the evidence slot record set assembly stage, the slot template loading unit loads slot templates. A slot template is defined as a structured template that assembles pilot observation summaries and sensing observation summaries into evidence records that can be jointly estimated and invoked. It includes at least a slot number field, a slot type field, a set of slot input fields, and a set of slot caliber fields. The slot type field includes at least pilot evidence slots, sensing evidence slots, and cross-source consistency slots. Specifically, the slot template loading unit selects the corresponding version of the slot template based on the reference signal list version identifier and probe list version identifier registered in the session configuration packet summary, and writes the slot template version identifier into the packet header field of the subsequent evidence slot record set encapsulation unit, thereby giving the evidence slot record set version traceability characteristics. Subsequently, the slot assembly unit performs slot assembly processing on the pilot observation summary and the sensing observation summary. The slot assembly processing maps the pilot observation summary set fields to the slot input field set of the pilot evidence slot, maps the sensing feature summary fields to the slot input field set of the sensing evidence slot, and maps the cross-source alignment relationship records to the slot input field set of the cross-source consistency slot. During the mapping process, the slot assembly unit writes a session identifier field, a time anchor identifier field, an alignment index field, and a sensing time window index field to each assembly record, so that the pilot evidence slot, the sensing evidence slot, and the cross-source consistency slot can be uniformly gated in the subsequent prior confidence registration processing.

[0058] The slot conflict merging unit performs conflict merging processing after slot assembly is completed. For cases where conflict flag fields appear in cross-source consistency slots under the same alignment index, the conflict flag field is written to the conflict summary field, and the conflict cause summary and clue source identifier are written to the conflict record field. When the conflict flag field originates from a consistency gating failure in S310, the conflict merging unit further appends the reference identifier of the alignment anomaly reference field to the conflict record field, making the conflict source traceable to the anomaly record of the alignment data packet. The slot quality binding unit then performs quality binding processing. This quality binding process retrieves the quality flag table generated in S230 from the session running state field and associates it with the alignment index field, binding the quality flag field to the corresponding evidence slot record. When the quality flag field indicates low confidence or a gap, the slot quality binding unit writes a quality flag reference field into the evidence slot record and provides this reference field as a gating input for subsequent prior confidence registration processing. After completing assembly, conflict resolution, and quality binding, the evidence slot record set encapsulation unit encapsulates and generates an evidence slot record set. This evidence slot record set includes at least a session identifier field, a time anchor identifier field, a slot template version identifier field, a slot number field, a slot type field, a set of slot input fields, a quality flag reference field, and a conflict summary field. This evidence slot record set is named as an output field at the end of this step. The evidence slot record set is subsequently used as an input field in S330 and, in S410, is used as input for joint channel parameter estimation along with the prior constraint packet. Simultaneously, the candidate parameter set is also named as an output field at the end of this step and is subsequently used as an input field in S410. Therefore, S320 simultaneously serves as a bridge between the formation of the candidate parameter set and the formation of the evidence organization structure in the process chain, and forms a continuous transmission relationship with the perception observation summary output by S310.

[0059] S330. Perform prior confidence registration processing on the evidence slot record set to generate a prior constraint package; In this invention, S330 is executed by the prior confidence registration module, which consists of at least a prior rule loading unit, a slot validity judgment unit, a confidence generation unit, a prior consistency archiving unit, a prior constraint package encapsulation unit, and an audit record unit. Specifically, the input source of S330 is the evidence slot record set output by the aforementioned S320. The evidence slot record set already includes a quality mark reference field and a conflict summary field. After accessing the evidence slot record set, S330 first reads the session identifier field and the slot template version identifier field, and retrieves the session configuration package summary and quality rule version record field summary associated with the session identifier field from the session running state field, using them as the loading context of the prior rule loading unit. The prior confidence is defined as a structured registration result of the degree to which evidence slot records can be adopted in subsequent joint estimation. The prior confidence is not expressed by a formula, but is composed of discrete level fields, source interpretation fields, and constraint trigger fields. The prior constraint package is defined as an encapsulated data package of the prior confidence registration result of the evidence slot record set and the set of constraint fields that can be invoked in subsequent joint estimation.

[0060] During the prior rule loading phase, the prior rule loading unit loads a prior rule set, which includes at least quality gating rules, conflict gating rules, cross-source consistency gating rules, and version consistency gating rules. Quality gating rules are used to mark evidence slot records as available, pending review, or not adopted based on the quality tag reference field. Conflict gating rules are used to mark cross-source consistency slots as conflicted based on the conflict summary field and write this information into the conflict source explanation field. Cross-source consistency gating rules are used to mark evidence slot records as consistent or weakly consistent based on the cross-source consistency tag field or its equivalent digest tag. Version consistency gating rules are used to determine whether evidence slot records are in the same version range based on the consistency between the slot template version identifier field and the version identifier in the session configuration package digest. After loading, the prior rule loading unit writes the prior rule set version identifier into the audit header field of the audit record unit and then writes this version identifier into the subsequent prior constraint package encapsulation unit, enabling the prior constraint package to have an auditable scope.

[0061] During the slot validity assessment phase, the slot validity assessment unit performs validity assessment processing on each record in the evidence slot record set. Validity assessment processing includes at least field integrity assessment, quality gating assessment, and conflict gating assessment. Field integrity assessment checks whether the slot input field set meets the minimum field set requirements of the slot template under the field mapping table. When a missing field or type mismatch occurs, the slot validity assessment unit marks the slot record as a field gap and writes it into the gap explanation field. Quality gating assessment reads the quality mark reference field and retrieves the corresponding entry from the quality mark table, writing the low confidence, gap, or supplementary collection trigger status summary into the quality explanation field. Conflict gating assessment reads the conflict summary field and, in conjunction with the conflict record field, writes the conflict source explanation field into the conflict explanation field. If both the pilot evidence slot and the perception evidence slot under the same alignment index are in a pending verification state, the slot validity judgment unit writes the alignment index into the linkage verification index field. The linkage verification index field is archived in the session running state field and can be used as a reference for the subsequent S420 pilot reconfiguration suggestion generation and processing, thereby forming a closed-loop connection with the subsequent detection strategy configuration across the main steps.

[0062] During the confidence generation phase, the confidence generation unit generates a priori confidence registration results based on the validity judgment results. Specifically, the confidence generation unit writes a priori confidence level field to each evidence slot record. The priori confidence level field includes at least one of high confidence, medium confidence, low confidence, and non-acceptance. Simultaneously, the confidence generation unit writes a source explanation field, which records the trigger basis summary for the priori confidence level field. The trigger basis summary is derived from at least the quality explanation field, conflict explanation field, or version consistency judgment result summary. Further, to support constraint invocation in subsequent joint channel parameter estimation processing, the confidence generation unit generates a constraint trigger field. The constraint trigger field indicates which slots are subject to which processing branch flags during joint estimation. The processing branch flags include at least one of weight reduction, elimination, delayed aggregation, and pending supplementary acquisition. Weight reduction and elimination are used as minimum set processing branch flags, while delayed aggregation and pending supplementary acquisition are used as optional extended processing branch flags. These are loaded and written when the resource budget parameter group allows and there are records in the linked review index field. After the confidence level is generated, the prior consistency archiving unit archives the prior confidence level registration result and the linked review index field together into the prior consistency archiving field. The archived content includes an alignment index range summary, a slot type summary, a prior confidence level statistical summary, and an anomaly source reference summary. The archived content can be directly retrieved through the session identifier field in subsequent audits and backtracking.

[0063] After completing confidence registration and archiving, the prior constraint package encapsulation unit encapsulates and generates a prior constraint package. The prior constraint package includes at least a session identifier field, a time anchor identifier field, a prior rule set version identifier field, a slot template version identifier field, a prior confidence registration result field, a constraint trigger field set, and a linkage review index field reference field. The prior constraint package is mentioned as an output field name at the end of this step. The prior constraint package is subsequently used as an input field in S410. Specifically, S410 obtains the prior constraint package, the evidence slot record set, and the candidate parameter set, performs joint channel parameter estimation processing, and obtains the channel posterior package. Simultaneously, S330 naturally inherits the evidence slot record set output from S320 on the main text link, and at the output end, jointly advances the prior constraint package and the evidence slot record set to stage S400. This allows the present invention to form an engineering-implementable continuous processing link that integrates perceptual evidence organization, quality gating, and prior constraint registration without introducing formulaic expressions.

[0064] In summary, the technical effects of this step are as follows: By introducing the quality tag reference field, conflict summary field, and version consistency caliber into the prior confidence registration of the evidence slot record set, and encapsulating the registration results into a prior constraint package for subsequent joint channel parameter estimation processing, the subsequent estimation process can complete traceable adoption, deweighting, or elimination of branches at the evidence level, thereby forming structured constraint inputs for cross-source conflicts and low-confidence slots.

[0065] Step S400 includes at least steps S410-S430: S410. Obtain the prior constraint packet, evidence slot record set and candidate parameter set, perform joint channel parameter estimation processing, and obtain the channel posterior packet. In the communication signal processing method of fusion sensing for fifth generation (5G) mobile communication technology of the present invention, S410 is located after the prior confidence registration process and belongs to the joint channel parameter estimation stage driven by fusion evidence. S410 is executed by the joint channel parameter estimation module, which consists of at least an input access unit, a caliber alignment unit, an evidence screening and gating unit, a parameter entry arrangement unit, a joint estimation core unit, an uncertainty generation unit, a posterior encapsulation unit, and an audit record unit. Specifically, the prior constraint package comes from the output of S330, and the prior constraint package includes at least a session identifier field, a time anchor point identifier field, a prior rule set version identifier field, a slot template version identifier field, a prior confidence registration result field, a constraint trigger field set, and a linkage review index field reference field; the evidence slot record set comes from the output of S320, and the evidence slot record set includes at least a session identifier field, a time anchor point identifier field, a slot template version identifier field, a slot number field, a slot type field, a slot input field set, a quality mark reference field, and a conflict summary field; the candidate parameter set also comes from the output of S320, and the candidate parameter set includes at least a parameter identifier, a parameter type identifier, an association alignment index, an association reference signal entry identifier, an association perception time window index, and a clue source identifier. During the input access phase, S410 first reads the session identifier field of the prior constraint packet and compares it with the session identifier fields of the evidence slot record set and the candidate parameter set. If an inconsistency is found, the caliber alignment unit marks the batch of inputs as cross-session anomalies and writes them into the anomaly record field of the audit record unit. At the same time, it triggers a search of the session running state field to confirm whether a session switching event exists. When a session switching event is confirmed, S410 temporarily stores the batch of inputs and waits for evidence consistent with its session identifier field to be supplemented before entering the joint estimation core unit. This enables the joint channel parameter estimation process to have the triggering conditions and anomaly closure loop for sustainable operation.

[0066] During the alignment phase, the alignment unit performs version alignment judgment on the slot template version identifier field and the prior rule set version identifier field, and retrieves the session configuration package summary, field mapping table summary, and reference signal list summary associated with the session identifier field from the session running state domain. The field mapping table summary is used to verify the field name and field type caliber of the slot input field set, and the reference signal list summary is used to verify the availability and validity period caliber of the associated reference signal entry identifier. Further, the alignment unit performs time window merging on the evidence slot record set based on the time anchor identifier field, merging pilot evidence slots, sensing evidence slots, and cross-source consistency slots within the same alignment index range into a unified estimation batch unit. The estimation batch unit serves as the smallest scheduling granularity for a joint estimation process in the joint estimation core unit. When the linkage verification index field reference field in the prior constraint package indicates that the estimation batch unit has an index to be verified, the alignment unit marks the estimation batch unit as a verification batch and writes it into the batch mark field, so that the subsequent evidence screening and gating units adopt more conservative gating branches. Understandably, the minimum set of the estimation batch units includes at least an alignment index range summary and a slot input field set reference. The alignment index range summary is used to limit the time window covered by this joint estimation, and the slot input field set reference is used to enable the joint estimation core unit to directly access pilot observation summary related fields and sensing observation summary related fields.

[0067] In the evidence screening and gating phase, the evidence screening and gating unit performs gating screening processing on each record in the evidence slot record set based on the prior confidence registration result field. Specifically, the evidence screening and gating unit extracts the prior confidence level field and constraint trigger field, which correspond one-to-one with the slot number field, from the prior confidence registration result field and writes them into the evidence gating mark table. The evidence gating mark table serves as an internal data structure for this step, used to apply differentiated adoption branches to different evidence slots in the joint estimation core unit. Further, when the constraint trigger field indicates removal, the evidence screening and gating unit removes the corresponding slot from the estimation batch unit and records a summary of the removal reason; when the constraint trigger field indicates deweighting or delayed aggregation, the evidence screening and gating unit retains the corresponding slot in the estimation batch unit and writes it into the processing branch mark field, enabling the joint estimation core unit to adopt a conservative update strategy for that slot in subsequent iterations. In addition to the prior constraint package, the triggering conditions for the gating and screening process can also be triggered by the quality mark reference field and the conflict summary field in the evidence slot record set: when the conflict summary field is in a conflict state and the quality mark reference field indicates low confidence, the evidence screening and gating unit marks the slot as a conflict low confidence state and prioritizes entering the delayed aggregation branch instead of directly entering the elimination branch, thereby reserving a traceable source of conflict evidence for subsequent pilot reconfiguration suggestions.

[0068] During the parameter entry arrangement phase, the parameter entry arrangement unit performs estimation target arrangement processing on the candidate parameter set. This estimation target arrangement processing maps the parameter entries in the candidate parameter set to parameter state entries that can be processed by the joint estimation core unit. Specifically, the parameter entry arrangement unit reads the parameter type identifier and divides the candidate parameter entries into channel parameter candidate entries, noise state candidate entries, interference state candidate entries, and observation bias candidate entries. Among these, channel parameter candidate entries and observation bias candidate entries are loaded first as the minimum set of entries for joint estimation. Noise state candidate entries and interference state candidate entries are loaded as extended entries when the resource budget parameter group allows it and the number of available slots in the evidence gating mark table meets the loading conditions. Subsequently, the parameter entry orchestration unit binds the parameter status entries to the estimation batch unit based on the associated alignment index and the associated perception time window index, forming a parameter-evidence binding record. This record includes at least a parameter identifier, a slot number reference, an alignment index range summary, and a clue source identifier. The clue source identifier distinguishes whether a parameter entry is primarily supported by pilot evidence slots, primarily by perception evidence slots, or by cross-source consistency slots, enabling the joint estimation core unit to perform hierarchical updates under evidence-driven conditions. Further, the parameter entry orchestration unit registers the parameter orchestration version identifier for this estimation in the session runtime state domain. This version identifier is associated with the version identifier of the candidate parameter set and is used to retrospectively check the input caliber of this estimation during subsequent uncertainty generation and pilot reconfiguration suggestion generation.

[0069] In the joint estimation core unit stage, the joint estimation core unit performs joint channel parameter estimation processing on the estimation batch unit, the evidence gating mark table, and the parameter-evidence binding record to obtain the channel posterior packet. The joint channel parameter estimation processing in this invention is manifested as an iterative update process driven by multi-source evidence. This iterative update process is not expressed by formulas, but rather achieves an achievable engineering operation path through discrete update rounds and update records. Specifically, the joint estimation core unit first establishes an estimation state container, which is used to hold parameter state entries, evidence slot input field set references, and update round records. Subsequently, in the initial round, the joint estimation core unit writes initial confidence clues to the parameter state entries based on the prior confidence level field. These initial confidence clues at least include high confidence, medium confidence, and low confidence levels, and are written together with the parameter-evidence binding record into the estimation state container. Furthermore, in each update round, the joint estimation core unit schedules the pilot evidence slot input field set and the perception evidence slot input field set according to the aligned index range summary, and selects the update path based on the processing branch marker field: when the processing branch marker field is adoption or deweighting, the joint estimation core unit updates the observation summary and parameter status entries in the corresponding slot input field set; when the processing branch marker field is delayed aggregation, the joint estimation core unit temporarily stores the slot input field set in the delayed aggregation cache and performs joint updates with the slot input field set of the adjacent aligned index range summary in subsequent rounds or subsequent estimation batches; when the processing branch marker field is removal, the joint estimation core unit skips the slot and only retains its removal reason summary for uncertainty generation. When performing associated updates, the joint estimation core unit simultaneously reads the cross-source consistency flag field and conflict record field of the cross-source consistency slot. For estimation batch units that have conflict states, a conflict suppression update strategy is adopted. The conflict suppression update strategy is manifested in reducing cross-source common update items and writing the conflict source explanation field into the update round record so that the subsequent pilot reconfiguration suggestion can identify whether the conflict source belongs to time alignment deviation, probe channel abnormality or reference signal entry mismatch.

[0070] In a practical deployment scenario, the network topology corresponds to a multi-cell coverage topology on the urban road side. The reference signal list registers various observable reference signal entries, and the probe list registers base station-side receiving link probes and roadside sensing probes. The pilot evidence slots in the evidence slot record set carry the pilot observation summary aggregation field from S220, the sensing evidence slots carry the sensing feature summary field from S310, and the cross-source consistency slots carry the cross-source alignment relationship record formed by merging S310 and S320. The joint estimation core unit performs several update rounds within each alignment index range summary, mapping the clues of the pilot evidence slots and sensing evidence slots to the update record of the parameter state entry. For the conflict state batch, a delayed aggregation buffer is used to wait for the evidence to be collected in the next alignment index range summary before updating. This ensures that the channel posterior packet contains both parameter posterior records and conflict source explanations and delayed aggregation references.

[0071] The uncertainty generation unit generates an uncertainty record after the joint estimation core unit completes the iterative update. In this invention, the uncertainty record is defined as a structured expression of the reliability and volatility of the parameter posterior record. The uncertainty record does not use formulas but instead employs a combination of a level field, a range summary field, a source explanation field, and a trigger suggestion field. Specifically, the uncertainty generation unit generates an uncertainty level field for each parameter identifier and associates and archives it with the corresponding alignment index range summary, evidence adoption summary, conflict summary, and delayed aggregation reference. When the uncertainty level field of a parameter identifier is in a high uncertainty state and its clue source identifier mainly comes from pilot evidence slots, the uncertainty generation unit writes a pilot reconfiguration related trigger flag in the trigger suggestion field. When its clue source identifier mainly comes from sensing evidence slots, the uncertainty generation unit writes a sensing-side supplementary acquisition related trigger flag in the trigger suggestion field, but this flag is only used as an input clue for subsequent S420 in this step and does not directly generate a suggestion. The a posteriori encapsulation unit then encapsulates the channel a posteriori packet, which includes at least a session identifier field, a time anchor identifier field, a parameter a posteriori record field, an uncertainty record field, an evidence acceptance summary field, a conflict source explanation field, a delay aggregation reference field, and a version identifier field. The channel a posteriori packet is named as an output field at the end of this step. This channel a posteriori packet is then used as an input field in subsequent step S420, specifically by extracting the uncertainty record from the channel a posteriori packet and performing pilot reconfiguration suggestion generation. After encapsulation, the audit recording unit writes the key operational records of this joint estimation process into the audit domain. These key operational records include at least a priori rule set version identifier, slot template version identifier, parameter orchestration version identifier, update round summary, and abnormal branch summary, thus providing the joint channel parameter estimation process with an auditable operational path supported by a version management strategy.

[0072] S420. Extract uncertainty records from the channel posterior packet, perform pilot reconfiguration suggestion generation processing, and generate pilot reconfiguration suggestions. In this invention, S420 is located after the joint channel parameter estimation process and belongs to the pilot resource adaptive orchestration stage for the next round of observation and detection. S420 is executed by the pilot reconfiguration suggestion generation module, which consists of at least a posteriori access unit, uncertainty analysis unit, reference signal entry mapping unit, reconfiguration candidate generation unit, budget constraint gating unit, suggestion encapsulation unit, and suggestion archiving unit. Specifically, the channel posteriori packet obtained by S420 comes from the output of S410, and the channel posteriori packet at least includes an uncertainty record field and a parameter posteriori record field. During the posteriori access stage, S420 reads the session identifier field and the version identifier field, and retrieves the reference signal list, probe list, and resource budget parameter group associated with the session identifier field from the session running state domain, so as to follow the predetermined criteria and budget boundaries during the pilot reconfiguration suggestion generation process. Furthermore, S420 extracts the uncertainty record field from the channel a posteriori packet and establishes an uncertainty index. The uncertainty index includes at least a parameter identifier, an uncertainty level field, an alignment index range summary, a source explanation field, and a trigger suggestion field. The uncertainty index serves as the core input of the reconfiguration candidate generation unit.

[0073] During the uncertainty analysis phase, the uncertainty analysis unit performs trigger interpretation processing on the uncertainty index. This trigger interpretation processing includes threshold interpretation based on the uncertainty level field and scenario interpretation based on the source explanation field. Threshold interpretation determines which parameter identifiers are in a state requiring pilot reconfiguration, while scenario interpretation identifies the main sources of uncertainty as time alignment deviation, conflict suppression updates, evidence sparsity, or insufficient reference signal entry coverage. Specifically, when the trigger suggestion field contains pilot reconfiguration-related trigger markers, the uncertainty analysis unit writes the corresponding parameter identifier into the reconfiguration trigger list and records its alignment index range summary and source explanation field. When the trigger suggestion field does not contain pilot reconfiguration-related trigger markers but the uncertainty level field is in a high uncertainty state, the uncertainty analysis unit further backtracks to the evidence adoption summary field in the parameter posterior record field. If the evidence adoption summary field shows insufficient pilot evidence slot adoption or frequent delayed aggregation references, it is still written into the reconfiguration trigger list and marked as an evidence sparsity trigger type. The reconfiguration trigger list exists as an internal data structure in this step, and its minimum set contains at least parameter identifiers, trigger type identifiers, and alignment index range summaries.

[0074] During the reference signal entry mapping phase, the reference signal entry mapping unit reads the reference signal list and maps the parameter identifiers in the reconfiguration trigger list to the reference signal entry identifier set. Specifically, the reference signal entry mapping unit traces back its associated reference signal entry identifiers based on the orchestration version identifier of the candidate parameter set in the session runtime state domain, and performs an expiration check on the reference signal entry identifier set: if a reference signal entry identifier is unavailable or has been gated disabled in the session configuration packet digest, the reference signal entry mapping unit replaces it with a spare reference signal entry identifier of the same type and records the replacement reason summary. Further, during the mapping process, the reference signal entry mapping unit simultaneously reads the probe list to determine which probes can observe and report each reference signal entry, and writes the set of available probe identifiers into the reconfiguration mapping record; the reconfiguration mapping record at least includes the reference signal entry identifier, the set of available probe identifiers, the trigger type identifier, and the alignment index range summary, which is used by the reconfiguration candidate generation unit to generate executable reconfiguration actions.

[0075] During the reconfiguration candidate generation stage, the reconfiguration candidate generation unit generates a reconfiguration candidate set based on the reconfiguration mapping record. The reconfiguration candidate set is defined as a set of executable pilot reconfiguration action entries, and each pilot reconfiguration action entry includes at least a reconfiguration action identifier, a reference signal entry identifier, a probe identifier reference, a reconfiguration effective window summary, and a return field list summary. Specifically, the reconfiguration candidate generation unit combines the uncertainty level field and alignment index range summary in the channel posterior packet to determine the reconfiguration effective window summary, and selects the reconfiguration action category based on the trigger type identifier: when the trigger type identifier indicates insufficient coverage of reference signal entries, the reconfiguration candidate generation unit generates pilot density enhancement action entries, which are manifested by increasing the configuration frequency of a certain type of reference signal entry within several subsequent alignment index range summaries and associating it with the available probe identifier set; when the trigger type identifier indicates that conflict suppression updates lead to an increase in uncertainty, the reconfiguration candidate generation unit generates pilot diversity enhancement action entries, which are manifested by introducing backup reference signal entry identifiers within the same alignment index range summary and observing probe combinations associated with cross-source consistency slots; when the trigger type identifier indicates sparse evidence, the reconfiguration candidate generation unit generates pilot focusing action entries, which are manifested by selecting probes with high observation stability from the available probe identifier set and encrypting their observation windows, while limiting the returned field list summary to the minimum set of fields required to support subsequent S220 pilot observation summary extraction and S230 quality marking processing. Understandably, the minimum set of fields includes at least a reference signal entry identifier, an observation timestamp summary, an observation quality cue summary, and an alignment index reference, so that subsequent steps can access the reconfigured observation data without expanding the caliber.

[0076] During the budget constraint gating phase, the budget constraint gating unit performs gating and pruning processing on the reconfiguration candidate set based on the resource budget parameter group. Specifically, the budget constraint gating unit reads the reported bandwidth budget and computing resource budget from the resource budget parameter group and prioritizes pruning of pilot density enhancement action items: when the reported bandwidth budget is in a tightened state, the budget constraint gating unit shortens the effective window summary of pilot density enhancement action items and prioritizes retaining pilot diversity enhancement action items or pilot focusing action items; when the computing resource budget is in a tightened state, the budget constraint gating unit prunes the summary of the returned field list in the reconfiguration candidate set into a smaller set and writes a field pruning mark in the pilot reconfiguration suggestion so that the S430 can use a matching processing link when generating the detection strategy configuration. Furthermore, the budget constraint gating unit performs conflict checks on the reconfiguration candidate set. Conflict checks include conflicts where the same reference signal entry identifier is simultaneously promoted and reduced by multiple action entries within the same effective window summary, and conflicts where the same probe identifier is excessively used by multiple action entries within the same effective window summary. When a conflict is detected, the budget constraint gating unit makes a decision based on the trigger type identifier in the source interpretation field, retaining the action entries associated with the high uncertainty level field, and writing the decision result into the conflict decision record field. After gating and pruning, the suggestion encapsulation unit encapsulates and generates pilot reconfiguration suggestions. These suggestions include at least a session identifier field, a version identifier field, a reconfiguration action entry set field, an effective window summary field, a return field list summary field, a field pruning flag field, and a conflict decision record field. The pilot reconfiguration suggestion is named as an output field at the end of this step. The pilot reconfiguration suggestion is subsequently used as an input field in S430, specifically by S430 performing detection strategy configuration generation processing on the pilot reconfiguration suggestion and outputting the detection strategy configuration. The suggested archiving unit also writes the pilot reconfiguration suggestion into the suggested archiving field of the session running state field and registers the suggestion version number, so that if a link interruption or session rollback event occurs later, the detection policy configuration can be reconciled and restored according to the suggestion version number.

[0077] In a real-world deployment scenario, the system operates collaboratively with roadside base stations and roadside sensing nodes. The channel posterior packet output by S410 contains fields with high uncertainty levels within certain alignment index range summaries, and the source explanation fields point to sparse pilot evidence and conflict suppression updates. Based on this, S420 generates pilot focusing action entries during the reconfiguration candidate generation stage and selects probe identifiers with high observation stability for reference. Simultaneously, during the budget constraint gating stage, it prunes the returned field list summary to the minimum field set according to the reported bandwidth budget and writes the field pruning mark field into the pilot reconfiguration suggestion, enabling S430 to generate a matching detection strategy configuration without changing the overall process.

[0078] S430. Perform detection strategy configuration generation processing on pilot reconfiguration suggestions to generate detection strategy configuration; In this invention, S430 occurs after the pilot reconfiguration suggestion generation and is a closed-loop implementation step that transforms the suggested actions into executable detection and acquisition strategy configurations. S430 is executed by the detection strategy configuration generation module, which comprises at least a suggestion access unit, an action parsing unit, a strategy template loading unit, a strategy orchestration unit, a strategy consistency verification unit, a strategy distribution preparation unit, and a configuration encapsulation and archiving unit. Specifically, the pilot reconfiguration suggestion obtained by S430 comes from the output of S420. The pilot reconfiguration suggestion includes at least a reconfiguration action item set field, an effective window summary field, a return field list summary field, and a field pruning mark field. During the suggestion access phase, S430 reads the session identifier field and retrieves the session configuration package summary, field mapping table summary, and probe list summary associated with the session identifier field from the session running state domain. These are used to maintain consistent field and probe definitions with previous steps during the detection strategy configuration generation process. Furthermore, S430 reads the version identifier field of the pilot reconfiguration suggestion and generates a configuration version number. The configuration version number is used for version management and auditable archiving of the detection strategy configuration, so that the system can locate the corresponding strategy configuration content by version number when evolving or rolling back.

[0079] During the action parsing phase, the action parsing unit performs action parsing processing on each reconfiguration action entry set field, mapping each pilot reconfiguration action entry to a detection strategy atomic action. The detection strategy atomic action is defined as the smallest action unit that can be executed by a specific network element or a specific unit within a network element. It includes at least an action type identifier, a reference signal entry identifier, a probe identifier reference, an execution timing summary, and a return field list summary. Specifically, the action parsing unit maps pilot density enhancement action entries to observation period adjustment atomic actions based on the action type identifier, maps pilot diversity enhancement action entries to reference signal entry switching atomic actions, and maps pilot focusing action entries to probe activation and observation window encryption atomic actions. An execution timing summary is written into each atomic action, which includes at least an effective window summary and an alignment index range summary reference, enabling the subsequent strategy orchestration unit to schedule the atomic actions to specific runtime windows. The action parsing unit also reads the field clipping mark field. When the field clipping mark field is in the clipping state, the action parsing unit locks the returned field list summary to the clipped field set and writes the field locking mark in the atomic action. This ensures that the subsequent detection chain follows the minimum set of "processable input fields and output fields" during execution, avoiding the anomaly of mismatch between acquisition and processing standards.

[0080] During the strategy template loading phase, the strategy template loading unit loads the detection strategy template. The detection strategy template is defined as a template that organizes the atomic actions of the detection strategy into a deployable, executable, and traceable configuration structure. It includes at least a strategy identifier field, a strategy activation window field, a strategy action orchestration field, a strategy feedback field mapping field, an exception handling branch field, and a rollback condition field. Specifically, the strategy template loading unit selects the corresponding version of the detection strategy template based on the session identifier field and the time anchor identifier field in the session configuration packet summary, and writes the detection strategy template version identifier into the packet header field of the configuration encapsulation and archiving unit, enabling the detection strategy configuration to have version traceability. Further, the strategy template loading unit loads probe capability descriptions based on the probe list summary. The probe capability descriptions include at least a set of observable reference signal entries for the probe, a probe reporting channel capability summary, and a probe scheduling restriction summary. The strategy orchestration unit will subsequently use this to determine whether a certain atomic action can be executed by a certain probe, thus ensuring that the generated detection strategy configuration has implementability constraints.

[0081] During the strategy orchestration phase, the strategy orchestration unit writes the detection strategy atomic actions into the strategy action orchestration field and completes the strategy configuration generation process. Specifically, the strategy orchestration unit orchestrates atomic actions according to the effective window summary, sequentially orchestrating the observation period adjustment atomic actions, reference signal entry switching atomic actions, and probe activation and observation window encryption atomic actions within the same effective window summary, and writes the action sequence number summary into the strategy action orchestration field. During the orchestration process, the strategy orchestration unit introduces a strategy feedback field mapping field, mapping and binding the feedback field list summary with the field mapping table summary, so that when S220 extracts the pilot observation summary from the received baseband sampling packet, it can locate the corresponding observation segment according to the mapped field name. Furthermore, the strategy orchestration unit writes the anomaly handling branch field into the detection strategy configuration. The anomaly handling branch field at least includes a degradation path summary when the probe reporting channel is interrupted, when the reference signal entry switching fails, or when the observation window encryption causes a resource budget conflict. The degradation path summary can be automatically triggered by the system in the engineering implementation. The automatic triggering conditions are at least from the anomaly flag of the probe reporting channel capability summary and the tightening flag of the resource budget parameter group. After triggering, the system marks the corresponding atomic action as invalid and rolls back to the detection strategy configuration registered in the previous configuration version number, so that the closed-loop iteration can continue to run under the link degradation condition and has auditable version rollback logic.

[0082] During the policy consistency verification phase, the policy consistency verification unit performs consistency verification processing on the policy action orchestration fields and the policy feedback field mapping fields. This consistency verification processing includes at least reference signal item availability verification, probe capability matching verification, and field caliber processability verification. Reference signal item availability verification determines whether all reference signal item identifiers referenced in the policy exist in the reference signal list summary and are in an available state; probe capability matching verification determines whether the probe identifier reference matches the set of observable reference signal items; and field caliber processability verification determines whether all fields in the feedback field list summary can be mapped to the system's internal processable field name set by the field mapping table summary. If the consistency verification fails, the policy consistency verification unit writes the reason for the failure to the verification failure record field and triggers the policy orchestration unit to perform corrective orchestration. Corrective orchestration includes replacing the backup reference signal item identifier, switching the backup probe identifier reference, or further trimming the feedback field list summary. The triggering conditions and execution records of this corrective orchestration are written to the audit record field, enabling subsequent review or reproduction to trace back the actual path of policy configuration generation.

[0083] After the consistency check passes, the policy delivery preparation unit generates the delivery payload. The delivery payload is defined as a configuration delivery structure for actual network elements and probe execution units, and it includes at least a policy identifier field, a configuration version number field, a policy activation window field, a policy action orchestration field, and a rollback condition field. Specifically, the policy delivery preparation unit writes the delivery payload into the delivery queue in the session runtime state domain and selects a delivery channel based on the communication capability list summary in the session configuration packet summary. When the communication capability list summary indicates that a delivery channel is unavailable, the policy delivery preparation unit marks the delivery queue entry as pending delivery and automatically retryes after the channel recovers. The number of retries and the retry interval are controlled by the delivery policy parameters in the session runtime state domain, and each retry is written to the audit record field, providing a traceable description of the automation level of the delivery process. The configuration encapsulation and archiving unit ultimately encapsulates and generates a detection strategy configuration. This detection strategy configuration includes at least a session identifier field, a configuration version number field, a detection strategy template version identifier field, a strategy action orchestration field, a strategy feedback field mapping field, an exception handling branch field, a rollback condition field, and a payload reference field. The detection strategy configuration is explicitly named as an output field at the end of this step. In subsequent operations, this detection strategy configuration is used to drive the next round of received baseband sampling and pilot observation summary extraction. Specifically, the strategy action orchestration field and the strategy feedback field mapping field in the detection strategy configuration are invoked by the received baseband sampling and acquisition process in S210 and the pilot observation summary extraction process in S220. This allows the pilot reconfiguration suggestion generated in S420 to be converted into an executable configuration through S430, forming a closed-loop connection from the posterior of S410 to the subsequent observations in S210 and S220.

[0084] In a real-world deployment scenario, the system runs on roadside base stations, edge processing nodes, and several probe nodes. The pilot reconfiguration suggestion output by S420 includes pilot focusing action entries with field pruning marker fields. During the action parsing phase, S430 maps these action entries to probe activation and observation window encryption atomic actions, and during the policy orchestration phase, it orchestrates them into the policy effective window fields corresponding to several subsequent aligned index range summaries. At the same time, it locks the returned field list summary to the pruned field set and maps it to the internal field name set of the field mapping table summary. When a probe node experiences an interruption in its reporting channel within the execution window, the exception handling branch field triggers the rollback logic defined by the rollback condition field. The system automatically rolls back to the detection policy configuration corresponding to the previous configuration version number and writes the rollback record to the audit record field, enabling the system to continue the iterative operation from S210 to S410 according to the established process even in the event of communication degradation or probe anomalies.

[0085] In summary, the technical effects of this step are as follows: by converting the uncertainty records in the channel posterior packet into executable pilot reconfiguration suggestions and further generating a detection strategy configuration with version number, rollback conditions and field caliber mapping constraints, subsequent observation acquisition and pilot extraction can be continuously connected with the input caliber of prior gating and joint estimation, and an auditable automated adjustment path is provided for conflict states and sparse observation states.

Claims

1. A communication signal processing method for 5G-oriented converged sensing, characterized in that, include: S100: Obtain the network topology, reference signal list, probe list, and resource budget parameter group; perform field mapping table registration, unified time base registration, and session identifier registration processing to generate a session identifier record structure; wherein, the session identifier record structure includes a session identifier, a session start time record, a network topology snapshot identifier, a reference signal list version identifier, a probe list version identifier, a resource budget parameter group version identifier, a field mapping table version identifier, and a time base reference identifier; S200. Based on the session identifier record structure, perform received baseband sampling acquisition, pilot observation summary extraction and time alignment processing to generate aligned data packets and a quality tag table; wherein, the quality tag table includes a session identifier field, an alignment index field, a probe identifier field, a quality tag field, an anomaly source reference field and a supplementary sampling request reference field; S300. Based on the aligned data packet and quality label table, perform perception observation summary generation, candidate parameter set extraction and evidence slot record set assembly processing to generate evidence slot record set and prior constraint package; wherein, the prior constraint package includes session identifier field, time anchor point identifier field, prior rule set version identifier field, slot template version identifier field, prior confidence registration result field, constraint trigger field set and linkage review index field reference field; S400, based on the evidence slot record set and prior constraint package, performs joint channel parameter estimation, uncertainty record extraction and pilot reconfiguration suggestion generation processing to generate detection strategy configuration.

2. The method according to claim 1, characterized in that, The process of registering a field mapping table includes: The session configuration generation process includes: the topology access submodule performing structure verification on the network topology; the reference signal registration submodule performing reference signal resource consistency verification on the reference signal list; the probe registration submodule performing probe capability caliber unification processing on the probe list; the budget registration submodule writing resource budget parameter groups into the budget record field; and the field mapping table registration submodule forming a field mapping table. The field mapping table includes field name specifications, field type specifications, unit specifications, timestamp caliber specifications, and default value caliber specifications.

3. The method according to claim 1, characterized in that, The process of unified time base registration and session identifier registration includes: The unified time reference registration process includes: the time source access submodule selecting the clock source access path based on the time synchronization capability description of the probe list; the time synchronization verification submodule performing multi-probe timestamp consistency verification and reference signal observation window consistency verification; the drift monitoring submodule entering continuous monitoring state after passing the time synchronization verification; and the time reference record generation submodule generating a time reference record, which includes a time anchor point identifier, clock source type description, time synchronization status summary, time synchronization verification result summary, drift monitoring status summary, and timestamp caliber specification reference identifier. The session identifier registration process includes: a session identifier generation submodule calling the time anchor identifier in the time base record and the network domain identifier information in the network topology to generate a session identifier; a session identifier conflict verification submodule performing comparison verification with the historical session identifier archive domain and conflict window verification of concurrent sessions within the same network domain; a session identifier record structure generation submodule generating a session identifier record structure; and a session state binding submodule binding the session identifier record structure to the running state domain of the session configuration package.

4. The method according to claim 1, characterized in that, The process of performing baseband sampling and acquisition includes: The receiving baseband sampling acquisition and processing includes: a sampling triggering submodule forming sampling triggering rules based on the period and triggering conditions of the reference signal list; a sampling window orchestration submodule generating sampling window orchestration results from the network topology, reference signal list, and probe list; an RF front-end access submodule and a baseband sampling link access submodule performing link access processing; a sampling quality pre-inspection submodule performing sampling continuity checks, timestamp monotonicity checks, and field parsing availability checks; and a receiving baseband sampling packet encapsulation submodule encapsulating the receiving baseband sampling packet. The receiving baseband sampling packet includes a session identifier field, a time anchor identifier field, a network topology snapshot identifier field, a reference signal list version identifier field, a probe identifier field, a sampling window orchestration result field, a sampling granularity description field, a sampling data payload field, and a pre-inspection anomaly reference field.

5. The method according to claim 1, characterized in that, The process of pilot observation summary extraction and time alignment includes: The pilot observation summary extraction and time alignment processing includes: an observation window positioning submodule locating sampling segments in the sampling window arrangement result field of the received baseband sampling packet based on the time-frequency resource location description of the reference signal list; a summary extraction submodule extracting pilot observation summaries from the located sampling segments; a time alignment submodule mapping the pilot observation summaries to a unified time reference to generate an alignment index; an alignment consistency verification submodule performing cross-probe alignment slot coverage consistency verification and alignment index continuity verification; and an alignment data packet encapsulation submodule encapsulating and generating alignment data packets. The alignment data packets include a session identifier field, a time anchor identifier field, an alignment index field, a pilot observation summary set field, an alignment deviation record field, and an alignment anomaly reference field.

6. The method according to claim 1, characterized in that, The process of generating a sensory observation summary includes: The sensing observation summary generation process includes: an input access unit performing packet header parsing and summary header parsing on the aligned data packet and pilot observation summary; an alignment index parsing unit mapping the alignment index field to a sensing time window index; a sensing resource orchestration unit generating a sensing resource orchestration result; a sensing feature extraction unit extracting a sensing feature summary according to the sensing resource orchestration result; a cross-source consistency judgment unit performing consistency judgment on the sensing observation summary and pilot observation summary; and a summary encapsulation unit encapsulating the sensing observation summary to obtain the sensing observation summary. The sensing observation summary includes a sensing time window index field, a sensing probe identifier field, a sensing feature summary field, a sensing confidence clue field, and an anomaly marker field.

7. The method according to claim 1, characterized in that, The process of candidate parameter set extraction and evidence slot record set assembly includes: The candidate parameter set extraction and evidence slot record set assembly process includes: a candidate parameter extraction unit extracting candidate parameter clues from pilot observation summaries and perception observation summaries to generate a candidate parameter set; a candidate parameter deduplication unit performing deduplication and merging processing on the candidate parameter set; a slot template loading unit loading slot templates; a slot assembly unit mapping pilot observation summaries and perception observation summaries to evidence slots; a slot conflict merging unit performing conflict merging processing; a slot quality binding unit binding a quality mark table to the evidence slots; and an evidence slot record set encapsulation unit encapsulating and generating an evidence slot record set. The evidence slot record set includes a session identifier field, a time anchor identifier field, a slot template version identifier field, a slot number field, a slot type field, a set of slot input fields, a quality mark reference field, and a conflict summary field.

8. The method according to claim 1, characterized in that, The process of performing joint channel parameter estimation includes: The joint channel parameter estimation process includes: an input access unit reading the prior constraint packet, the evidence slot record set, and the candidate parameter set and performing consistency comparison; a caliber alignment unit performing version alignment judgment on the slot template version identifier field and the prior rule set version identifier field; an evidence screening and gating unit performing gating screening processing on the evidence slot record set based on the prior confidence registration result field; a parameter entry arrangement unit performing estimation target arrangement processing on the candidate parameter set; a joint estimation core unit performing joint channel parameter estimation processing on the estimation batch unit, the evidence gating mark table, and the parameter-evidence binding record; an uncertainty generation unit generating uncertainty records; and a posterior encapsulation unit encapsulating the channel posterior packet, which includes a session identifier field, a time anchor identifier field, a parameter posterior record field, an uncertainty record field, an evidence adoption summary field, a conflict source explanation field, a delay aggregation reference field, and a version identifier field.

9. The method according to claim 1, characterized in that, The process of uncertainty record extraction and pilot reconfiguration suggestion generation includes: The uncertainty record extraction and pilot reconfiguration suggestion generation process includes: a posterior access unit reading channel posterior packets and establishing an uncertainty index; an uncertainty analysis unit performing trigger interpretation processing on the uncertainty index to generate a reconfiguration trigger list; a reference signal entry mapping unit mapping parameter identifiers in the reconfiguration trigger list to a set of reference signal entry identifiers; a reconfiguration candidate generation unit generating a reconfiguration candidate set based on the reconfiguration mapping record; a budget constraint gating unit performing gating and pruning processing on the reconfiguration candidate set based on the resource budget parameter group; and a suggestion encapsulation unit encapsulating and generating a pilot reconfiguration suggestion. The pilot reconfiguration suggestion includes a session identifier field, a version identifier field, a reconfiguration action entry set field, an effective window summary field, a return field list summary field, a field pruning mark field, and a conflict resolution record field.

10. The method according to claim 1, characterized in that, The process of generating the detection strategy configuration includes: The detection strategy configuration generation process includes: a suggestion access unit reading pilot reconfiguration suggestions and generating a configuration version number; an action parsing unit performing action parsing processing on each reconfiguration action item set field and mapping it to detection strategy atomic actions; a strategy template loading unit loading the detection strategy template; a strategy orchestration unit writing the detection strategy atomic actions into the strategy action orchestration field; a strategy consistency verification unit performing consistency verification processing on the strategy action orchestration field and the strategy feedback field mapping field; a strategy delivery preparation unit generating the delivery payload; and a configuration encapsulation and archiving unit encapsulating and generating the detection strategy configuration. The detection strategy configuration includes a session identifier field, a configuration version number field, a detection strategy template version identifier field, a strategy action orchestration field, a strategy feedback field mapping field, an exception handling branch field, a rollback condition field, and a delivery payload reference field.

Images