CN-122001751-A - Multi-dimensional verification and anomaly backtracking method for CTC system scheduling instruction

CN122001751ACN 122001751 ACN122001751 ACN 122001751ACN-122001751-A

Abstract

The invention discloses a multi-dimensional verification and anomaly backtracking method for a CTC system scheduling instruction. The method comprises the steps of carrying out structural conversion on an instruction, endowing global unique tracking identification and complementing real-time context information, sequentially executing three layers of progressive check of format, logic and equipment states, wherein the logic check is completed by a configurable rule engine supporting hot update, the equipment state check adopts a hybrid retransmission mechanism with self-adaption capability, generating and storing a structured log in real time when each node is circulated based on the tracking identification after the instruction is issued, constructing a full-link log chain, quickly searching the full-link log according to the identification when the instruction is abnormal, and accurately positioning faults and generating reports by automatically checking the states of each node. The method realizes the full-dimension safety check of the instruction, the full-link state tracing and the minute-level abnormal positioning, and remarkably improves the dispatching safety and the operation and maintenance efficiency.

Inventors

LIU SHEN
ZHANG XIN
HUANG TIANQI
LI TAO
LONG YING
SONG XIAOYAN
XIE YU
HE YURUI
TANG BIN
ZHANG HAILIN
ZHAO DEJUN
ZHAO HONGTAO
ZHAO JUNZHE
LIU XIANG
ZHONG WEI
Guo Songkun
ZHAO ZHIHE
CHANG DONGDONG
GAO FENG
LI WEI
WANG ZHINAN
JIN HAILIN
LIU ZIFENG
Hao Guoze
SUN GUOHUA
GAO PAN

Assignees

中国铁道科学研究院集团有限公司
中国铁道科学研究院集团有限公司通信信号研究所
北京华铁信息技术有限公司
北京锐驰国铁智能运输系统工程技术有限公司

Dates

Publication Date: 20260508
Application Date: 20260210

Claims (10)

1. A multi-dimensional verification and anomaly backtracking method of a CTC system scheduling instruction is characterized by comprising the following steps: The method comprises the steps of S1, receiving a scheduling instruction, uniformly converting the scheduling instruction into a structured data format, generating a globally unique Trace ID for the instruction, and simultaneously complementing context information related to the instruction from a CTC real-time database; s2, performing three-layer progressive compliance verification based on the Trace ID, namely sequentially performing format verification, logic verification and equipment state verification on the preprocessed scheduling instruction; The format verification verifies the grammar compliance of the instruction according to a preset instruction protocol; the logic verification verifies the actual logic consistency of the instruction and the train position, the route state and the operation plan by inquiring the dynamic operation data in the CTC real-time database; the equipment state verification verifies whether the target terminal equipment has hardware conditions for executing instructions by inquiring the real-time working condition of the target terminal equipment; S3, if the instruction passes all checks, issuing the instruction, and generating and storing a structured log in real time at each node of the instruction stream based on the Trace ID, and constructing a full-link log chain comprising an issuing node, a transmitting node, a receiving node and an executing node; and S4, carrying out exception backtracking based on the full-link log chain, namely when detecting that the instruction is executed abnormally, searching and analyzing the corresponding full-link log chain by taking the Trace ID as an index, positioning abnormal nodes and judging fault reasons by checking key state parameters in each node log, and generating a backtracking report.
2. The method of claim 1, wherein the structured data format is an organized, extensible markup language format data object based on key-value pairs that includes fields for identifying instruction semantics and nested parameter sets, and wherein the globally unique Trace ID is generated from a timestamp, a random factor, and an identification Fu Pinjie related to an instruction target for achieving strong association of logs in the full-link log chain.
3. The method according to claim 1, wherein the logic verification in step S2 is performed by a configurable logic rule engine that supports defining and storing verification rules in the form of triples of "priority + trigger condition + verification algorithm", and performing rule matching using optimized Rete matching algorithm, and supports hot update of rules, and dynamically loads new rules without restarting the system.
4. The method according to claim 1, wherein the constructing the full link log chain in the step S3 specifically includes: s31, when a dispatching desk generates an instruction, recording an issuing node log containing Trace ID, instruction content, generation time and dispatcher information; S32, recording a transmission node log containing Trace ID, transmission delay, packet loss rate and channel information in the instruction communication transmission process; S33, when the terminal equipment receives the instruction, recording a receiving node log containing Trace ID, equipment ID, receiving time, equipment on-line state and decoding result; S34, dynamically recording an execution node log containing Trace ID, execution progress, an execution result and a fault code in the process of executing the instruction by the terminal equipment; And carrying out double-identification association on all the node logs through the Trace ID and Span ID for identifying the sub-processes in the nodes, and storing by adopting a double-layer architecture combining local cache and a distributed database.
5. The method of claim 4, wherein the distributed database employs an index lifecycle management strategy to divide the log data into hot data, warm data, and cold data according to time, and store the hot data, the warm data, and the cold data in storage media of different performances, respectively, and wherein the log query response time is made to be less than or equal to 1 second by optimizing the index shard and replica configuration.
6. The method of claim 1 wherein the device status check in step S2 uses a hybrid retransmission mechanism of "exponential backoff+random factor" to query the terminal device for conditions, and when the query request is overtime or unresponsive, the retransmission is performed within a set maximum retry number according to an incremental delay time and with an additional random delay, and the retransmission parameters are dynamically adjusted according to the device history communication quality.
7. The method of claim 1 wherein the exception backtracking in step S4 generates an instruction stream timing diagram by arranging the logs of each node in the full-link log chain in time sequence, and sequentially checks the log integrity of the issuing node, the link quality index of the transmitting node, the device status and decoding status of the receiving node, and the execution continuity and failure code of the executing node to precisely locate which link the exception occurred.
8. A CTC system dispatch instruction multidimensional verification and anomaly backtracking system for implementing the method of any one of claims 1 to 7, comprising: the instruction processing module is used for receiving and carrying out preprocessing of the instructions and generation of Trace IDs; The multi-dimensional verification module is used for performing format verification, logic verification and equipment state verification and comprises a configurable logic rule engine; the log chain construction and management module is used for collecting logs at each node of the instruction stream, correlating the logs through Trace ID and storing the logs into the double-layer storage architecture; The exception backtracking and analyzing module is used for retrieving the full-link log based on Trace ID when the instruction is abnormal, positioning the abnormal node and generating an analysis report; And the equipment communication management module is used for carrying out working condition inquiry and instruction transmission with the terminal equipment by adopting a timeout retransmission mechanism.
9. The system of claim 8, wherein the log chain construction and management module comprises: the log acquisition unit is deployed at each node and used for generating a structured log in real time; And the associated storage unit is used for attaching Trace ID and Span ID to the log and managing the storage, index and life cycle of the log in the local cache and the distributed database.
10. The system of claim 8, wherein the logic rules engine in the multi-dimensional verification module comprises: a rule configuration interface for defining and modifying the verification rules in a structured manner; The rule base is used for storing rules organized in a triple form of priority, triggering condition and checking algorithm; And the rule execution unit integrates the optimized Rete matching algorithm, is used for loading and executing the rules in the rule base and supporting a hot updating function.

Description

Multi-dimensional verification and anomaly backtracking method for CTC system scheduling instruction Technical Field The invention relates to the technical field of safety control of a rail transit signal system, and particularly provides a multi-dimensional verification and backtracking method which takes compliance verification and anomaly tracking into consideration aiming at a scheduling instruction processing link of a scheduling centralized control system (Centralized Traffic Control, CTC for short). Background The CTC system is used as a core carrier for track traffic train operation scheduling, and bears key scheduling instruction generation and issuing functions such as train route opening, train number change, speed limitation, shunting operation scheduling and the like, and the accuracy of instruction processing directly relates to train operation safety and transportation efficiency. With the improvement of the rail traffic operation density and the intelligent development, the complexity and the safety requirement of dispatching instructions are increasingly severe, the identities of dispatchers are changed from the traditional director to director and operators, and the operation intensity and the safety risk are synchronously increased. In actual operation, the wrong issuing or abnormal execution of the instruction may cause safety accidents such as train conflict, route occupation, transportation delay and the like, even constitute general class D and above accidents, but the prior art is difficult to meet the actual requirements of full-dimension verification, full-link tracing and rapid abnormal positioning, and technical innovation is needed. In the prior art, the multi-dimensional verification and anomaly backtracking of the CTC system scheduling instruction mainly adopts the following methods: static checking method for instruction format The scheme is based on a preset protocol template or a regular rule, and performs single-dimension matching verification on coding fields, parameter numbers, data types and value formats of the modulating instructions, only verifies compliance of instruction grammar layers, and does not relate to any dynamic operation data such as real-time train positions, route occupation states, equipment working conditions and the like. However, the method can only check the grammar compliance of the instruction, can not identify the instruction with legal format but logic error (for example, issuing an open instruction to an occupied route), has serious potential safety hazard, the check rule depends on manual presetting and maintenance, the matching logic needs to be rewritten when the instruction type is newly added or the protocol is regulated, the adaptation period is long, the response is slow, a flexible processing mechanism is lacked for edge scenes such as fuzzy parameters, inconsistent versions and the like, the error interception rate is high, only the original text of the instruction is stored, the rule matching detail and associated data in the check process are not recorded, and the error source is difficult to position after the abnormality occurs. Binary feedback method for instruction execution result After the terminal equipment executes the instruction, the scheme returns a simple identification of success/failure to the CTC system through a communication link, the system only records the instruction ID and the final execution result, and no process data feedback such as transmission process, equipment state, execution progress or detailed fault cause exists. However, the feedback information of the method is extremely deficient, and the method can only judge whether an execution result is yes or not, so that specific failure reasons such as transmission packet loss, equipment failure, logic conflict, poor link quality and the like cannot be distinguished, so that a failure source is difficult to accurately define, the exception checking efficiency is low, operation and maintenance personnel need to manually check a multi-source log, the average checking time exceeds 2 hours, the pre-judging and intercepting capability is lacking, the equipment offline and resource occupation saturated scenes and other conditions which do not have execution conditions cannot be identified before the instruction is issued, potential safety hazards exist in the instruction blind issuing, and key link parameters such as transmission delay, packet loss rate and the like cannot be recorded, so that the influence of the link quality on the instruction execution cannot be quantified. Log segmentation method based on content semantics The scheme identifies semantic boundaries (such as event start/end identification) in the log through Natural Language Processing (NLP) technology, and realizes logic complete segmentation. Typical implementations include a BERT model-based log feature extraction segmentation scheme, keyword matching semantic segmentation algorithms, etc.,