CN-122027636-A - Anti-cheating system and method for driving training timing

CN122027636ACN 122027636 ACN122027636 ACN 122027636ACN-122027636-A

Abstract

The invention provides a driving training timing anti-cheating system and method, wherein the system comprises a distributed data coordination module, a timing module and a data synchronization module, wherein the distributed data coordination module is used for acquiring driving training data from a plurality of timing terminals and adopting a node cluster based on Raft consistency algorithm to perform data synchronization; the node cluster is configured to enable log entries related to driving training data to be batch packages of different types according to service event types corresponding to the driving training data, trigger a log copying process to achieve batch package synchronization, a digital twin mapping module for generating a three-dimensional visual twin model based on the driving training data synchronized by the node cluster, an intelligent behavior analysis module for extracting student behavior characteristics from the three-dimensional visual twin model, identifying abnormal behavior patterns based on the machine learning model, outputting analysis results, and an early warning module for generating a risk analysis report based on the analysis results. The invention realizes the real-time monitoring of the cheating behavior.

Inventors

GAN HAIBO
LIANG JIANGHUA

Assignees

武汉木仓科技股份有限公司

Dates

Publication Date: 20260512
Application Date: 20251231

Claims (10)

1. A system for preventing cheating during driving training, the system comprising: The distributed data collaboration module is used for acquiring driving training data from a plurality of timing terminals and carrying out data synchronization by adopting a node cluster based on Raft consistency algorithm, wherein the node cluster is configured to divide log entries associated with the driving training data into batch packets of different types according to different time windows and quantity thresholds and trigger a log replication process of Raft consistency algorithm to realize synchronization of the batch packets according to the service event types corresponding to the driving training data; The digital twin mapping module is used for receiving the driving training data synchronized by the node clusters and generating a three-dimensional visual twin model based on the driving training data; The intelligent behavior analysis module is used for extracting behavior characteristics of a learner from the three-dimensional visual twin model, identifying an abnormal behavior mode based on a machine learning model and outputting an analysis result; and the early warning module is used for generating a risk analysis report based on the analysis result, wherein the risk analysis report comprises a risk grade and a treatment instruction.
2. The drive training timing anti-cheating system of claim 1, wherein the lot packets comprise a core lot packet, a critical lot packet, and a regular lot packet, the time window and quantity threshold of the core lot packet being less than the time window and quantity threshold of the critical lot packet, the time window and quantity threshold of the critical lot packet being less than the time window and quantity threshold of the regular lot packet.
3. The ride-on time anti-cheating system of claim 1, wherein the distributed data coordination module is further configured to execute a compression policy prior to triggering Raft a log replication process of a consistency algorithm; The compression strategy includes: And acquiring the current network bandwidth utilization rate and the data type of the batch of packets, determining a target compression algorithm based on the network bandwidth utilization rate and the data type, and compressing the batch of packets based on the target compression algorithm.
4. The ride-on time anti-cheating system of claim 1, wherein the cluster of nodes is further configured to: Monitoring the bandwidth utilization rate, the round trip time and the packet loss rate between the node cluster and the timing terminal, and determining a network health score based on the bandwidth utilization rate, the round trip time and the packet loss rate; Increasing the time window or number threshold when the network health score is greater than a first score; Decreasing the time window or the number threshold when the network health score is less than a second score; wherein the first score is greater than the second score.
5. The ride-on time anti-cheating system of claim 1, wherein the cluster of nodes comprises a leading node and a following node, the bulk package comprising a global sequence number and a hash value; The following node is configured to verify the correctness of the hash value and the continuity of the global sequence number after receiving the bulk packet, and if the verification fails, to request retransmission of the bulk packet from the leader node.
6. The drive training timing anti-cheating system of claim 1, wherein the drive training data comprises GPS track data, sensory data, and video stream data, and wherein the digital twinning mapping module comprises: The training field semantic modeling unit is used for digitally modeling a driving school training field, dividing the driving school training field into a plurality of training areas with semantic definition to obtain a training field semantic model, and associating a semantic tag with each training area and at least one operation specification requirement; The semantic mapping unit is used for carrying out space-time alignment on the GPS track data, the sensing data and the video stream data to obtain space-time alignment data, fusing the space-time alignment data with the training field semantic model to generate three-dimensional scene data with semantic labels, and generating the three-dimensional visualization twin model based on the three-dimensional scene data rendering.
7. The ride-on time anti-cheating system of claim 1, wherein the intelligent behavior analysis module comprises: The feature engineering unit is used for extracting student behavior features from the three-dimensional visualization twin model, and screening a key feature subset from the student behavior features based on an information gain evaluation index; the model construction and reasoning unit is used for constructing a decision tree classification model based on historical training data, inputting the key feature subset into the decision tree classification model and obtaining the analysis result.
8. The system for preventing cheating during driving training according to claim 1, wherein the intelligent behavior analysis module is further configured to perform semantic compliance judgment on the behavior of the learner based on semantic information provided by the three-dimensional visual twin model, and generate a semantic compliance judgment result, where the semantic compliance judgment result includes at least one of region residence time compliance, operation and region matching degree; the digital twin mapping module is further configured to receive the semantic compliance determination result and highlight the region determined to be semantically abnormal in a playback interface of the three-dimensional visual twin model.
9. The ride-on time anti-cheating system of claim 1, wherein the pre-warning module is further configured to: the auditing rule change of the supervision platform is monitored in real time, the rule change is analyzed through a natural language processing technology, an updating rule is obtained, and the updating rule is pushed to the timing terminal; And determining auditing failure reasons and solution suggestions when monitoring that the driving training data auditing fails based on a preset academic data auditing error code mapping table.
10. The anti-cheating method for driving training timing is characterized by comprising the following steps of: Acquiring driving training data from a plurality of timing terminals, and performing data synchronization by adopting a node cluster based on Raft consistency algorithm, wherein the node cluster is configured to divide log entries associated with the driving training data into batch packages of different types according to different time windows and quantity thresholds according to service event types corresponding to the driving training data, and trigger Raft consistency algorithm log replication flow to realize synchronization of the batch packages; receiving driving training data synchronized by the node cluster, and generating a three-dimensional visual twin model based on the driving training data; extracting student behavior characteristics from the three-dimensional visualization twin model, identifying an abnormal behavior mode based on a machine learning model, and outputting an analysis result; a risk analysis report is generated based on the analysis result, the risk analysis report including a risk level and a treatment instruction.

Description

Anti-cheating system and method for driving training timing Technical Field The invention relates to the technical field of driving training management, in particular to a system and a method for preventing cheating during driving training. Background The traditional driving training timing management system faces serious challenges in terms of guaranteeing the reality at the time of learning and preventing cheating, and mainly has the following technical defects: 1. Data consistency and reliability are inadequate in that existing systems typically employ simple client-server or asynchronous message synchronization between timing terminals, driving school platforms, and superior supervisory platforms. The method is difficult to deal with network fluctuation and node failure, is easy to lose, repeat or contradict the learning record caused by data synchronization failure, conflict or manual tampering, seriously damages legal rights and interests of students and driving schools, and provides a multiplicative opportunity for learning counterfeiting. 2. The anti-cheating measures are passive and single, and the current anti-cheating measures are mostly independent of face recognition, GPS positioning or hardware anti-disassembly, and are easy to bypass by novel cheating technologies (such as positioning simulation and video playback). Meanwhile, the means lack of cooperation, so that formed data fragmentation is difficult for auditors to intuitively and efficiently screen out complex cheating modes such as 'horse race cheating', 'track instantaneous movement' from massive unstructured data. 3. The auditing process is highly dependent on manual work, and the efficiency is low, namely the auditing work of the supervision platform is seriously dependent on manual checking of text and chart logs one by one, and the overall, visual and traceable visual reproduction capability of the training process is lacking. Aiming at the increasingly growing training data, the manual auditing is time-consuming and labor-consuming, the unification of the standards is difficult to ensure, and the elaborate camouflage cheating behavior is easy to miss, so that the method becomes a bottleneck of industry supervision. Therefore, there is an urgent need to provide a system and a method for preventing cheating during driving training, which can guarantee credibility from a data source, provide intuitive process tracing, and have integrated driving training supervision technology with intelligent analysis and real-time early warning capabilities. Disclosure of Invention In view of the foregoing, it is necessary to provide a system and a method for preventing cheating during driving training, which are used for solving the technical problems that in the prior art, the timing data is unreliable, the training process cannot be directly traced back, and intelligent analysis and early warning cannot be performed, so that the cheating during driving training cannot be effectively prevented. In order to solve the above technical problems, the present invention provides a system for preventing cheating during driving training, comprising: The distributed data collaboration module is used for acquiring driving training data from a plurality of timing terminals and carrying out data synchronization by adopting a node cluster based on Raft consistency algorithm, wherein the node cluster is configured to divide log entries associated with the driving training data into batch packets of different types according to different time windows and quantity thresholds and trigger a log replication process of Raft consistency algorithm to realize synchronization of the batch packets according to the service event types corresponding to the driving training data; The digital twin mapping module is used for receiving the driving training data synchronized by the node clusters and generating a three-dimensional visual twin model based on the driving training data; The intelligent behavior analysis module is used for extracting behavior characteristics of a learner from the three-dimensional visual twin model, identifying an abnormal behavior mode based on a machine learning model and outputting an analysis result; and the early warning module is used for generating a risk analysis report based on the analysis result, wherein the risk analysis report comprises a risk grade and a treatment instruction. In one possible implementation, the batch packages include a core batch package, a critical batch package, and a regular batch package, the time window and quantity threshold of the core batch package being less than the time window and quantity threshold of the critical batch package, the time window and quantity threshold of the critical batch package being less than the time window and quantity threshold of the regular batch package. In one possible implementation, the distributed data collaboration module is further configured to execute a compression policy b