Timing training anti-cheating system and method based on trusted root and behavior space-time chain

Abstract

The invention provides a timing training anti-cheating system and a timing training anti-cheating method based on a trusted root and a behavior space-time chain, which can fuse time, space and behavior multi-dimensional information, ensure credibility from a data source and realize all-around and high-reliability cheating prevention through intelligent association analysis, and comprises a timing training terminal with a built-in trusted execution environment TEE module; the trusted execution environment TEE module is used for generating environment digital fingerprints and encrypting and signing the collected vehicle state sensing terminal data; the system further comprises a cloud collaborative judgment platform which is used for receiving a signed chain type data packet, and the chain type data packet comprises a plurality of data units which are linked according to a time sequence and signed by TEE; based on a preset physical rule model, performing logic consistency verification on the spatio-temporal trajectory data and the vehicle behavior data in the chained data packet; and judging whether simulation positioning cheating exists through a multi-mode causal analysis module.

Description

A Time-Based Training Anti-Cheating System and Method Based on Trust Roots and Behavioral Spatiotemporal Chains Technical Field

[0001] This invention relates to the field of driver timekeeping training technology, and more specifically, to a timekeeping training anti-cheating system and method based on a root of trust and a behavioral spatiotemporal chain. Background Technology

[0002] To meet regulatory requirements for training duration and prevent falsification of training hours, the core of driver training supervision is to safeguard training authenticity. Existing technologies primarily verify the authenticity of training from different dimensions, but each has its limitations and a systematic solution has yet to be formed. For example: 1. Technologies based on vehicle operating status verification focus on verifying the authenticity of the vehicle's operating status. For instance, Chinese patent application CN202510990372.1 discloses a scheme that uses a lifetime vehicle identification number (VIN) to prevent equipment from being moved, and fuses data from the vehicle's CAN bus (vehicle speed and RPM) with data from gyroscope sensors for analysis. This type of scheme can effectively defend against parking attendance fraud and signal injection-based cheating.

[0003] However, such technologies have significant drawbacks: their anti-cheating measures are entirely limited to vehicle behavior itself and cannot identify location-related fraud. Cheators can install the entire verification system on a real, ordinary vehicle. While the system verifies that the vehicle is actually moving, it cannot know whether that movement occurred on the designated training area or route. This could allow them to falsify valid training hours by driving on non-training roads, creating a regulatory blind spot.

[0004] 2. Techniques based on time and image comparison verification focus on the authenticity of timestamps in training records. For example, Chinese patent application CN202210758967.0 discloses a GPS-based time synchronization system. This system automatically compares the time displayed on an independent GPS time synchronization device with the system time of the timing device while simultaneously capturing images of trainees to determine if the time has been tampered with. This method aims to address the problem of easily modified timestamps in single images.

[0005] However, this technology has its shortcomings: its verification focus is singular, only addressing the authenticity of a "point in time," while failing to verify the authenticity of the vehicle's behavioral trajectory and positional changes within a "time period." It cannot identify driving trajectories forged using simulated positioning signals at the correct time point, nor can it determine whether the vehicle is in a genuine operating state, making its anti-cheating capabilities one-sided and weak.

[0006] In summary, the two main paths in existing technologies are independent of each other, leading to systemic technical deficiencies in the field of anti-cheating in time-based training: Comparison document 1 only verifies vehicle behavior, not location; comparison document 2 only verifies time, not behavior and trajectory. Neither can handle complex cheating methods that combine multiple elements of time, space, and state, such as "actual driving outside the training area" or "faking dynamic trajectories at a fixed location." Existing solutions mostly rely on directly collecting data from the device's operating system layer or the vehicle bus. This data is at risk of being intercepted, tampered with, or forged by malicious software during generation, storage, and transmission, lacking a mechanism to ensure the authenticity and integrity of data from the hardware source.

[0007] It only performs threshold judgments or simple comparisons on single types of data such as vehicle speed or timestamps, and fails to forcibly associate multimodal data such as time, space, and vehicle status, and to perform cross-dimensional logical consistency verification based on physical laws, resulting in insufficient ability to identify cheating of high-tech analog signals. Summary of the Invention

[0008] The purpose of this invention is to provide a time-based training anti-cheating system and method based on trusted roots and behavioral spatiotemporal chains. It can integrate multi-dimensional information of time, space and behavior to ensure trustworthiness from the data source and achieve comprehensive and highly reliable anti-cheating through intelligent correlation analysis.

[0009] The embodiments of the present invention are implemented as follows: A timing training anti-cheating system based on a root of trust and a behavioral spatiotemporal chain includes: a timing training terminal, fixedly installed in a training vehicle, the timing training terminal having a built-in Trusted Execution Environment (TEE) module and a BeiDou / GPS dual-mode timing module; a vehicle status sensing terminal, connected via the vehicle's OBD-II interface, used to collect the vehicle's VIN code, CAN bus signal, and gyroscope data; the TEE module is used to generate an environmental digital fingerprint for the timing training terminal and to encrypt and sign the collected positioning data, time data, and data from the vehicle status sensing terminal; the BeiDou / GPS dual-mode timing module is used to provide a unified time reference for all data; the system also includes a cloud-based collaborative adjudication platform, used to: receive signed chained data packets, the chained data packets containing multiple data units linked in chronological order and signed by the TEE; perform logical consistency verification on the spatiotemporal trajectory data and vehicle behavior data in the chained data packets based on a preset physical rule model; when the spatiotemporal trajectory data indicates a change in vehicle status, but the vehicle behavior data does not indicate a corresponding change in physical status, the multimodal causal analysis module determines that there is simulated positioning cheating.

[0010] In a preferred embodiment of the present invention, the aforementioned environmental digital fingerprint is generated by the Trusted Execution Environment (TEE) module during terminal initialization by integrating the unique device identifier of the timing training terminal, the vehicle VIN code successfully acquired for the first time, and the initial positioning coordinates, and is then bound to the cloud platform for two-way authentication.

[0011] In a preferred embodiment of the present invention, each data unit in the chained data packet includes at least: currently collected spatiotemporal data and vehicle behavior data, digital signature of the current data unit, and hash value of the previous data unit, to form an immutable chained evidence storage structure.

[0012] In a preferred embodiment of the present invention, the above-mentioned physical rule model includes vehicle dynamics constraint rules, which are used to verify whether the deviation between the displacement change calculated by positioning data and the displacement change calculated by integrating gyroscope data within the same time period is within a reasonable threshold range.

[0013] In a preferred embodiment of the present invention, the multimodal causal analysis module is further configured to trigger a secondary verification process based on electronic fences and historical reliable locations when vehicle behavior data indicates that the vehicle is in a driving state and spatiotemporal trajectory data indicates that the vehicle position has not changed.

[0014] This invention also provides a time-based training anti-cheating method based on a trusted root and a behavioral spatiotemporal chain, applying any of the aforementioned time-based training anti-cheating systems. The method includes: terminal initialization and trusted binding, generating an environmental digital fingerprint through a Trusted Execution Environment (TEE) module and completing two-way authentication binding with a cloud platform to establish a hardware trusted root; chain-based evidence storage of teaching process data, periodically and synchronously collecting spatiotemporal data and vehicle behavior data based on a unified time benchmark during training, encapsulating them into chain-based data units within the TEE module, encrypting and signing them before uploading; cloud-based collaborative adjudication, receiving and verifying the integrity and signature validity of the chain-based data units; performing logical consistency analysis on the spatiotemporal trajectory and vehicle behavior in the data units based on a physical rule model; and performing multimodal causal analysis when a physical logical contradiction is detected between the spatiotemporal data and the behavior data to generate an anti-cheating adjudication result.

[0015] In a preferred embodiment of the present invention, the above-mentioned terminal initialization and trusted binding steps specifically include: the trusted execution environment (TEE) module securely reading and storing the vehicle VIN code and the device unique identifier; obtaining the initial high-confidence positioning coordinates; and uploading the combined hash value of the VIN code, the device unique identifier, and the initial positioning coordinates as the environmental digital fingerprint to the cloud platform to complete the registration.

[0016] In a preferred embodiment of the present invention, in the above-mentioned teaching process data chain storage step, when encapsulating each chain data unit, the hash value of the previous data unit needs to be embedded to form a data chain that is related to the previous and subsequent data.

[0017] In a preferred embodiment of the present invention, the multimodal causal analysis performed in the above-mentioned cloud-based collaborative adjudication step includes: if the analysis results show that the abnormal changes in the spatiotemporal trajectory lack corresponding vehicle behavior support, it is determined to be external simulated signal cheating; if the analysis results show that the vehicle behavior shows normal driving but the spatiotemporal trajectory is abnormally stagnant, it is determined whether it is normal training under signal obstruction environment by combining the status of the environmental digital fingerprint and the electronic fence information.

[0018] In a preferred embodiment of the present invention, the above method further includes an adjudication feedback step: feeding back the anti-fraud adjudication result and the corresponding chain of data evidence to the regulatory terminal, and providing an authenticity verification interface based on data hash value.

[0019] The beneficial effects of this invention are as follows: 1. Existing cheating methods are trending towards combination, such as simultaneously using simulated GPS signals and OBD signal injectors. Traditional solutions, due to isolated verification dimensions, are easily deceived by such combinations. The multimodal causal analysis module of this solution can discover cross-dimensional logical contradictions. For example, satellite positioning shows that the vehicle is cornering at high speed, but the gyroscope does not detect the expected lateral acceleration; or the engine speed soars but the vehicle positioning does not move at all, thus accurately determining it as a high-tech complex cheating, which cannot be achieved by a single technical path solution; 2. This solution fundamentally reconstructs the trust foundation of data generation and transmission, ensuring the authority of regulatory decisions. At the source, through the built-in Trusted Execution Environment (TEE) module, all core data is encrypted and signed in a secure isolation area at the moment of generation; in the process, a chain-like evidence storage data structure is introduced, so that each data unit contains historical hashes; in the adjudication, when the system determines cheating, it can provide a complete, self-verifying spatiotemporal evidence chain of behavior, intuitively showing when, where, and which data began to show anomalies and logical contradictions. This greatly facilitates the review and determination of responsibility by regulatory personnel, minimizing disputes.

[0020] 3. Driver training environments, such as underground garages, urban canyons, and remote road sections, are complex and varied. This solution ensures high reliability and availability through system design. For example, when entering a fully enclosed garage, GPS / BeiDou signals are completely lost. Traditional solutions that rely solely on positioning may fail or generate false alarms. This solution uses behavioral data as the primary source and known, reliable location as a secondary source for intelligent decision-making. Combined with environmental digital fingerprint verification that the vehicle has not left the vehicle, it can still effectively determine whether the vehicle is undergoing compliant static operation training, avoiding false alarms or functional paralysis caused by signal problems. The system uses a rule base such as vehicle dynamics models for verification, and the threshold can be dynamically adjusted according to vehicle type and training subject, reducing false judgments caused by fixed thresholds and improving the system's intelligence and adaptability. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 is a schematic diagram of the connection of the timed training anti-cheating system based on trusted roots and behavioral spatiotemporal chains according to an embodiment of the present invention; Figure 2 is a schematic diagram of the internal module connection of the timed training terminal according to an embodiment of the present invention; Figure 3 is a flowchart of the operation of the timed training terminal according to an embodiment of the present invention; Figure 4 is a schematic diagram of the internal unit connection of the vehicle state perception terminal according to an embodiment of the present invention; Figure 5 is a flowchart of the power-on initialization and self-test operation of the vehicle state perception terminal according to an embodiment of the present invention; Figure 6 is a flowchart of the teaching process operation of the vehicle state perception terminal according to an embodiment of the present invention; Figure 7 is a schematic diagram of the internal connection of the cloud collaborative adjudication platform according to an embodiment of the present invention; Figure 8 is a schematic diagram of the real-time data stream processing and trusted verification process of the cloud collaborative adjudication platform according to an embodiment of the present invention; Figure 9 is a schematic diagram of the asynchronous intelligent analysis and adjudication process of the cloud collaborative adjudication platform according to an embodiment of the present invention; Figure 10 is a schematic diagram of the adjudication output and feedback process of the cloud collaborative adjudication platform according to an embodiment of the present invention; Figure 11 is a flowchart of the timed training anti-cheating method based on trusted roots and behavioral spatiotemporal chains according to an embodiment of the present invention.

[0023] Icons: Timing Training Terminal A; Trusted Execution Environment (TEE) Module A01; Beidou / GPS Dual-Mode Timing Module A02; First Data Interface and Communication Module A03; Vehicle Status Perception Terminal B; Main Control and Security Preprocessing Unit B01; Vehicle Signal Security Acquisition Unit B02; High-Precision Inertial Measurement Unit B03; Anti-Disassembly and Communication Unit B04; Cloud-based Collaborative Adjudication Platform C; Secure Access and Data Gateway Module C01; Trusted Verification and Evidence Chain Restoration Engine C02; Intelligent Adjudication Engine C03; Adjudication Management and Feedback Portal Module C04. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0025] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0026] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0027] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided to make this application more comprehensive and complete, and to fully convey the concept of the exemplary embodiments to those skilled in the art.

[0028] Furthermore, the described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. Numerous specific details are provided in the following description to give a thorough understanding of embodiments of this application. However, those skilled in the art will recognize that the technical solutions of this application can be practiced without one or more of the specific details, or other methods, components, apparatuses, steps, etc., can be employed. In other instances, well-known methods, apparatuses, implementations, or operations are not shown or described in detail to avoid obscuring various aspects of this application.

[0029] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0030] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0031] Please refer to Figures 1 and 2 for the first embodiment. This embodiment provides a timing training anti-cheating system based on trusted roots and behavioral spatiotemporal chains, including a timing training terminal A, a vehicle status perception terminal B, and a cloud-based collaborative adjudication platform C that communicates with the two terminals.

[0032] The timing training terminal A is fixedly installed in the training vehicle. The timing training terminal A has a built-in Trusted Execution Environment (TEE) module A01, a Beidou / GPS dual-mode timing module A02, a first data interface, and a communication module A03.

[0033] The Trusted Execution Environment (TEE) module A01 is an integrated chip used to generate the environmental digital fingerprint of the timing training terminal A, and to encrypt and sign the collected location data, time data, and data from the vehicle status perception terminal B. It also integrates a secure storage area located inside the TEE, used to securely store the device's unique identifier, which is burned into the chip, the environmental digital fingerprint, the encryption key, and a temporary vehicle VIN code. This area cannot be directly accessed or read from the outside. It also integrates an encryption engine that supports Chinese national cryptographic algorithms SM2 / SM3 / SM4 or international standards such as AES-256 and SHA-256, providing hardware-level acceleration for data signing and encryption.

[0034] The environmental digital fingerprint is generated by the Trusted Execution Environment (TEE) module A01 during terminal initialization by integrating the unique device identifier of the timing training terminal A, the vehicle VIN code successfully acquired for the first time, and the initial positioning coordinates, and is then bound to the cloud platform for two-way authentication.

[0035] The BeiDou / GPS dual-mode timing module A02 provides a unified time reference for all data, employing a high-precision timing and positioning module supporting both BeiDou-3 and GPS L1 / L5 frequency bands. Its core functions include: prioritizing the reception of BeiDou satellite timing signals, seamlessly switching to GPS when BeiDou signals are weak, and outputting a 1PPS signal with an accuracy better than 100 nanoseconds and the corresponding UTC time message; completing positioning data acquisition and simultaneously outputting positioning information such as latitude, longitude, speed, and heading angle; and integrating a network positioning assistance module, combining 4G / 5G cellular network and Wi-Fi positioning capabilities to provide auxiliary position reference when satellite signals are completely interrupted.

[0036] The first data interface and communication module A03 includes at least a vehicle data interface, an auxiliary sensor interface, and an encrypted communication module. The vehicle data interface is a CAN bus interface, connected to the vehicle's OBD-II port via a dedicated isolator, used for secure, read-only monitoring of CAN signals such as the vehicle's VIN code, speed, RPM, and ignition status. This interface is designed to physically prevent write-back to avoid interference with the original vehicle circuitry. The auxiliary sensor interface can be expanded to connect to a built-in six-axis inertial measurement unit (IMU), including a three-axis gyroscope and a three-axis accelerometer, to provide redundant motion data in case of malfunctions in the vehicle status sensing terminal B. The encrypted communication module supports TLS 1.3 or a national cryptographic standard-based encrypted transmission protocol, establishing a secure channel with the cloud platform via a 4G / 5G network.

[0037] Please refer to Figure 3. The operation flow of the timing training terminal A is as follows: Terminal initialization and establishment of trusted root: A11: The terminal is securely installed inside the dashboard of the training vehicle, connected to the vehicle's constant power and ACC power supply, and connected to the OBD-II port; A12: TEE secure startup: After the terminal is powered on, the TEE secure startup verification is performed first to ensure that its running code has not been tampered with; A13: Environmental information collection and fingerprint generation: The TEE passively listens to and securely reads the vehicle's VIN code through the CAN interface; at the same time, the TEE obtains the initial high-reliability positioning coordinates provided by the Beidou / GPS module. The initial high-reliability positioning coordinates must meet the conditions of satellite number ≥ 6 and positioning accuracy factor HDOP < 2.0; the TEE concatenates the device unique ID, vehicle VIN code, and initial coordinates, and uses the SM3 hash algorithm to generate a unique environmental digital fingerprint.

[0038] A14: Cloud-based two-way authentication binding. The terminal sends the environmental digital fingerprint to the cloud platform through an encrypted channel. The cloud platform binds the fingerprint with the training vehicle information registered with the driving school. The training vehicle information includes the license plate, VIN, and permitted training area. It also generates a digital certificate or token for the terminal and sends it to the TEE for secure storage. Thus, the terminal, vehicle, and initial location are uniquely bound for life, forming an unclonable hardware root of trust.

[0039] The teaching process uses a chain-based data storage system. During training, A21 (spatial-temporal synchronous data acquisition) is executed periodically. In each acquisition cycle, such as 1 second, the high-precision timestamp output by the Beidou / GPS dual-mode timing module A02, such as 2025-10-27 10:00:00.123456 UTC, triggers a global data acquisition. Under the same nanosecond-level time base, the system synchronously acquires the following data: 1. Latitude and longitude and speed from the positioning module; 2. Vehicle speed and RPM from the CAN bus; 3. Encrypted gyroscope data packets from the vehicle status perception terminal B.

[0040] A22: Encapsulation and signing within the TEE. All raw data is sent to the TEE. Within the TEE, the program packages the current data with the calculated hash value of the previous data unit to form a new data block. Using the private key stored in the TEE, this data block is digitally signed, such as using the SM2 algorithm.

[0041] A23: Construct a chain of data units, which must include at least a timestamp, a block index, the hash value of the previous data unit, packet data, and a signature of all of the above. The packet data should include at least location data, such as latitude and longitude coordinates and speed; vehicle data, such as bus transmission data and speed; and sensor data, such as encrypted gyroscope data hashes. The hash value calculated for the current data unit will be temporarily stored in the TEE secure storage area for the construction of the next data unit, thus forming a unidirectional, irreversible hash chain.

[0042] The signed chain of data units is uploaded to the cloud in real time via an encrypted communication module. If the network is interrupted, the data will be encrypted and cached in the terminal's external storage, and will be re-transmitted in order after the network is restored, maintaining the chain structure.

[0043] The vehicle status perception terminal B is connected through the vehicle OBD-II interface and is used to collect vehicle VIN code, CAN bus signal and gyroscope data. Please refer to Figure 4. It includes: main control and safety preprocessing unit B01, vehicle signal safety acquisition unit B02, high-precision inertial measurement unit B03 and anti-disassembly and communication unit B04.

[0044] The main control and security preprocessing unit B01 includes a microcontroller (MCU), a secure storage module, and a real-time clock module. The MCU is an automotive-grade MCU with a floating-point unit and a hardware encryption engine, responsible for coordinating various modules, parsing protocols, fusing data, and performing local preprocessing. The secure storage unit has built-in EEPROM or Flash memory for storing the terminal's unique ID, vehicle VIN code cache, and the session key generated by pairing with the timing training terminal A. The real-time clock module has an independent power supply to maintain time and status records when the vehicle is off.

[0045] The vehicle signal safety acquisition unit B02 includes an OBD-II interface and isolation circuit, as well as a multi-protocol parsing chip.

[0046] The terminal connects to the vehicle via a standard OBD-II connector. The core design of the OBD-II interface and isolation circuit lies in the signal acquisition isolation circuit, which employs a high-speed digital isolator or optocoupler to ensure that the terminal can only listen to signals on the vehicle's CAN bus and other buses, and absolutely cannot write any instructions or signals to the vehicle's bus. This fundamentally eliminates the risk of the vehicle's electronic system security being affected by terminal malfunctions or malicious attacks. The isolation circuit also provides signal injection filtering; it includes a hardware filter that can filter out, to a certain extent, unconventional high-frequency injected signals, such as abnormal waveforms generated by certain racing machines.

[0047] The multi-protocol parsing chip is an integrated chip that can automatically identify and parse multiple vehicle protocols such as ISO 15765-4 (CAN) and ISO 14230 (KWP2000), and accurately locate and read key signals such as VIN code, vehicle speed, engine speed, ignition status, throttle opening, and gear position.

[0048] The high-precision inertial measurement unit B03 is a sensor fusion processor added to a traditional six-axis motion sensor. The sensor fusion processor has a built-in Kalman filter fusion algorithm, which can directly output stable attitude, angle and linear acceleration data after noise reduction and attitude compensation.

[0049] The anti-tampering and communication unit B04 connects to the high-precision inertial measurement unit B03. The high-precision inertial measurement unit B03 itself monitors for abnormal movement or vibration of the terminal, corresponding to unauthorized disassembly. It can also use a sealed label or simple circuit continuity detection to determine if the casing has been illegally opened. Its interface connects to the timing training terminal A via a dedicated cable with a physical locking mechanism or strong pairing, using Bluetooth / BLE 5.0 with distance awareness. This connection is used to: receive a unified high-precision timestamp calibrated by BeiDou / GPS distributed by the timing training terminal A, serving as the time reference for all data from this terminal; and send encrypted behavioral data packets to the timing training terminal A in real time for chain encapsulation.

[0050] Please refer to Figure 5. The specific operation flow of the vehicle status perception terminal B is as follows: Terminal power-on initialization and self-test B11: Insert the terminal into the vehicle's OBD-II interface, and the terminal is powered by the vehicle's power supply to start; B12: After the main control and safety preprocessing unit B01MCU starts, it first performs hardware self-tests on the IMU, memory, and communication interface, and checks the anti-disassembly status. If disassembly marks are detected, a high-risk event is recorded locally; B13: The multi-protocol parsing chip automatically scans the bus, identifies the vehicle communication protocol, and sends a standard diagnostic request to read the vehicle's VIN code. The read VIN code is cached in the secure storage area; B14: Handshake pairing with the timing training terminal A, and bidirectional authentication with the timing training terminal A's TEE module through a dedicated secure communication link; exchange session keys, and receive the first authoritative timestamp calibrated by satellite timing, completing local RTC synchronization.

[0051] Please refer to Figure 6. Synchronous acquisition and preprocessing of behavioral data during the teaching process: B21: Each data acquisition cycle is strictly synchronized with the timing training terminal A, such as 1Hz, triggered by a synchronization time pulse signal or a timestamped acquisition command from the timing training terminal A; B22: Parallel acquisition of multi-source data, including vehicle CAN signal acquisition, which, through isolation circuitry, listens to and captures raw messages of signals such as vehicle speed and engine speed on the CAN bus within the same microsecond-level time window; IMU raw data acquisition, synchronously reading the raw digital outputs of the gyroscope and accelerometer, for example, through the SPI interface; B23: Local data fusion and preprocessing. The MCU parses the CAN message into specific physical quantities based on the preset or adaptively learned DBC file; the MCU or IMU's built-in processor fuses the raw data from the six axes to calculate and output more stable vehicle pitch angle, roll angle, and composite acceleration; key information is packaged into a concise but complete data structure, which includes at least: the autonomous terminal's master timestamp, vehicle speed, engine speed, ignition status, longitudinal acceleration, lateral acceleration, and yaw rate; B24: local signature and forwarding, using the session key negotiated with the timing training terminal A, performs HMAC calculation or lightweight encryption on the generated behavioral data digest to generate a data integrity check code; the raw data digest and check code are sent to the timing training terminal A in real time through a secure communication link.

[0052] The cloud-based collaborative adjudication platform C is used to: receive signed chained data packets, which contain multiple data units linked chronologically and signed by a TEE. Each data unit in the chained data packet includes at least: currently collected spatiotemporal data and vehicle behavior data, the digital signature of the current data unit, and the hash value of the previous data unit, forming an immutable chained evidence structure; perform logical consistency verification on the spatiotemporal trajectory data and vehicle behavior data in the chained data packets based on a preset physical rule model. The physical rule model includes vehicle dynamics constraint rules, which are used to verify whether the deviation between the displacement change calculated from the positioning data and the displacement change calculated from the gyroscope data integration within the same time period is within a reasonable threshold range; when the spatiotemporal trajectory data indicates a change in vehicle state, but the vehicle behavior data does not indicate a corresponding change in physical state, the multimodal causal analysis module determines that there is simulated positioning fraud. The multimodal causal analysis module is also configured to: trigger a secondary verification process based on electronic fences and historical trusted locations when the vehicle behavior data indicates that the vehicle is in a driving state, while the spatiotemporal trajectory data indicates that the vehicle position has not changed.

[0053] The cloud-based collaborative adjudication platform C is the core analysis and decision-making hub of the system. It adopts a microservice architecture and has high availability, high concurrency, and elastic scalability capabilities. Please refer to Figure 7. It includes a secure access and data gateway module C01, a trusted verification and evidence chain restoration engine C02, an intelligent adjudication engine C03, and an adjudication management and feedback portal module C04.

[0054] The secure access and data gateway module C01 includes a two-way authentication gateway and a data receiving and buffering queue. The two-way authentication gateway ensures that all terminal connections are authenticated using two-way TLS or national cryptographic SSL certificates, guaranteeing that the access user is only the legitimate time-training terminal A. The data receiving and buffering queue receives encrypted chained data units uploaded by terminals, performs preliminary format verification, and then places them into a highly reliable message queue, such as Kafka / RocketMQ, for buffering, achieving traffic shaping and asynchronous processing.

[0055] The trusted verification and evidence chain reconstruction engine C02 is used to complete signature verification, hash chain continuity verification, and spatiotemporal alignment and evidence chain reconstruction. Specifically, the signature verification service retrieves data units from the queue, uses the corresponding terminal's pre-set public key to obtain and verify its TEE digital signature during initial binding, and only data with valid signatures proceeds to subsequent processes. The hash chain continuity verification service groups the data that has passed signature verification by device ID, sequentially calculates the hash value of each data unit, and compares it with the "previous unit hash value" recorded in the next data unit. Any mismatch will cause that segment of the evidence chain to be marked as "broken," triggering a serious fraud alert. Spatiotemporal alignment and evidence chain reconstruction reassembles verified continuous data units from the same terminal into a complete and tamper-proof "behavioral-spatiotemporal evidence chain" in the time dimension for use by the subsequent analysis engine.

[0056] The intelligent adjudication engine C03 serves as the core, and includes a multi-dimensional rule model library and a multi-modal causal analysis module.

[0057] The multi-dimensional rule model library includes a vehicle dynamics rule model, an electronic fence and training scenario rule library, and a temporal behavior rule model.

[0058] Vehicle dynamics rule models are one of the core models. For example, a displacement-acceleration-gyroscope correlation model is established based on car / truck vehicle types. One specific verification rule is: if ΔS gps / Δt>20 km / h and ΔS imu If the distance is less than 5 meters, it is marked as "physical contradiction - suspected simulated location". ΔS gps Δt is the GPS displacement difference, ΔS is the time difference, and ΔS is the GPS displacement difference. gps / Δt is the average velocity, ΔS imu The displacement is calculated by double integration using a gyroscope and an accelerometer.

[0059] The electronic fence and training scenario rule base is used to store the electronic fence polygons of legal training sites, road sections that are allowed for training, maximum speed limits, typical training durations, etc.

[0060] The temporal behavior rule model is used to define the behavior patterns of normal training, such as "ignition ON" should precede "vehicle speed > 0"; "lateral acceleration (gyroscope)" should peak synchronously with "heading angle change (GPS)" when turning, etc.

[0061] The multimodal causal analysis module includes a feature extractor, a contradiction detection and hypothesis generator, and a scenario simulation and verification unit.

[0062] Feature extractors are used to extract key feature sequences from the chain of evidence, such as velocity curves, acceleration spectra, position sequences, and steering angle sequences.

[0063] The contradiction detection and hypothesis generator: When the rule engine detects anomalies that violate basic physical laws, such as velocity without displacement, this module will analyze the details of the contradiction in depth and generate possible cheating hypotheses, such as "simulating GPS signals" or "interfering with CAN signals by a racing machine".

[0064] For complex contradictions, the scenario simulation and verification module calls a lightweight vehicle motion model and attempts to simulate using two modes: "normal training" and "hypothetical cheating". It compares which mode can better explain all the observed positioning, CAN, and IMU data, thereby elevating the analysis from rule violation to the level of optimal explanation.

[0065] The adjudication management and feedback portal includes an adjudication result generator and regulatory visualization and API interfaces.

[0066] The adjudication result generator integrates the outputs of the rule engine and the causal analysis module to generate structured adjudication results, including: device ID, time period, such as the anomaly type of simulated location cheating, confidence level, key evidence pointers, pointing to specific abnormal data segments in the evidence chain.

[0067] The regulatory visualization and API interface provide traffic management departments with a web portal that visualizes the real-time training status, historical trajectories, and anomaly alarm heatmaps of all vehicles. It also provides an API that allows the regulatory system to query adjudication results on demand and independently verify the returned data hash value in a blockchain repository or locally.

[0068] Please refer to Figure 8-10. The execution flow of the cloud-based collaborative adjudication platform C includes: C1: Real-time data stream processing and trusted verification; C11: The data gateway receives chained data units from the terminal, transmitted via the TLS channel; C12: Signature verification and chain verification, signature verification, verifying the signature and public key of the chained data unit; hash verification, i.e., chain verification, comparing the hash values ​​of the previous chained data unit and the current chained data unit, and simultaneously calculating and temporarily storing the hash value of the current chained data unit for verifying the hash value of the next chained data unit; C13: Storing the verified data units and their associated relationships into the spatiotemporal database and hash index library to form a complete evidence chain that can be quickly queried.

[0069] C2: Asynchronous Intelligent Analysis and Adjudication; C21: For newly added continuous data segments, the rule engine is triggered; Example: Dynamics Validation: Select time period T1 to T2. Calculate the spherical distance ΔS between two points. gps Δt = T2 - T1, therefore V gps = ΔS gps / Δt.

[0070] Simultaneously, Simpson integrals were performed on the gyroscope and accelerometer data during this period to calculate the vehicle's displacement ΔS in the northeast-central coordinate system. imu .

[0071] Application rule: If |V gps - (ΔS imu / Δt)|>Vs, where Vs is the speed threshold, such as 15 km / h, which triggers alarm A1.

[0072] C22: Causal analysis module intervention: When alarm A1 is generated, the causal analysis module is activated.

[0073] Feature Analysis: Module Discovery V gps = 60 km / h, but ΔS imu ≈ 0, and the IMU data spectrum shows that the vehicle is in a near-static micro-vibration mode.

[0074] Hypothesis generation: Primary hypothesis H1: GPS signal is simulated; secondary hypothesis H2: IMU malfunction.

[0075] Cross-validation: Check the CAN bus vehicle speed signal V during the same time period. can If V can If V ≈ 0, then H1 is strongly supported; if V can If the value is approximately 60, the contradiction shifts to "the machine injects a CAN signal but the IMU does not move," triggering another set of analysis procedures.

[0076] Scenario assessment: Considering that the "environmental digital fingerprint" status is normal, the device has not been disassembled, and the vehicle is located in a non-signal-shielded area, the final determination is "high-confidence simulated GPS positioning cheating".

[0077] C23: Triggering conditions for adjudication in special scenarios such as signal obstruction: The rule engine detects V gps ≈ 0 but V can >0 and ΔS imu >0.

[0078] Causal analysis process: Query the vehicle's historical frequently used training routes and geofences linked to the cloud.

[0079] Determine if the last reliable GPS point is located in a "known signal obstruction area," such as near the entrance of an underground parking garage. Using Kalman filtering or particle filtering algorithms, based on the last reliable location, IMU integral data, and CAN vehicle speed, perform dead reckoning (DR) to generate a projected trajectory.

[0080] The endpoint of the calculated trajectory is compared with the first location point after the GPS signal is restored. If the distance between the two is within a reasonable error and the entire calculation process is within the electronic fence, it is judged as "normal training under signal obstruction", and the training time is valid.

[0081] C3: Decision Output and Feedback C31: Generate Decision Report. The decision report generated by the platform shall include at least the terminal number, vehicle identification number, time range, final decision, anomaly type, confidence level, evidence chain hash value, key contradictions, and recommended actions.

[0082] C32: Multi-channel feedback pushes high-risk rulings, such as those related to cheating, to the driving school management terminal and monitoring screen in real time; the ruling report and its associated original data chain hash value are synchronously stored on the blockchain to ensure that the ruling result itself is tamper-proof and judicially traceable; and the ruling query and evidence verification interface is opened to the regulatory system.

[0083] The second embodiment of the present invention also provides a time-based training anti-cheating method based on a root of trust and a behavioral spatiotemporal chain, applying the time-based training anti-cheating system of the first embodiment (see Figure 11). The method includes: S1: Terminal initialization and trusted binding, generating an environmental digital fingerprint through the Trusted Execution Environment (TEE) module and completing two-way authentication binding with the cloud platform to establish a hardware root of trust; S2: Chain-based evidence storage of teaching process data, periodically and synchronously collecting spatiotemporal data and vehicle behavior data based on a unified time reference during training, and storing the data in the Trusted Execution Environment (TEE). The module encapsulates data into chained data units, encrypts and signs them, and then uploads them. When encapsulating each chained data unit, the hash value of the previous data unit must be embedded to form a data chain that is related to the previous and subsequent data units. S3: Cloud-based collaborative adjudication receives and verifies the integrity and signature validity of the chained data units. Based on a physical rule model, it performs logical consistency analysis on the spatiotemporal trajectory and vehicle behavior in the data units. When a physical logical contradiction is detected between the spatiotemporal data and the behavioral data, it performs multimodal causal analysis to generate an anti-fraud adjudication result. The anti-fraud adjudication result and the corresponding chained data evidence chain are fed back to the regulatory terminal, and an authenticity verification interface based on the data hash value is provided.

[0084] In the cloud-based collaborative adjudication process, the multimodal causal analysis includes: if the analysis results show that the abnormal changes in the spatiotemporal trajectory lack corresponding vehicle behavior support, it is determined to be cheating by the external simulated signal; if the analysis results show that the vehicle behavior shows normal driving but the spatiotemporal trajectory is abnormally stagnant, it is determined whether it is normal training under signal obstruction environment by combining the status of the environmental digital fingerprint and the electronic fence information.

[0085] The specific steps of terminal initialization and trusted binding include: the Trusted Execution Environment (TEE) module A01 securely reads and stores the vehicle VIN code and the device's unique identifier; obtains the initial high-confidence positioning coordinates; and uploads the combined hash value of the VIN code, device unique identifier, and initial positioning coordinates as the environmental digital fingerprint to the cloud platform to complete the registration.

[0086] The third embodiment, based on the timed training anti-cheating system based on trusted roots and behavioral spatiotemporal chains in the first embodiment and the timed training anti-cheating method based on trusted roots and behavioral spatiotemporal chains in the second embodiment, elaborates in detail how the multimodal causal analysis module in the cloud-based collaborative adjudication platform C can accurately identify and judge the high-tech complex cheating of "high-speed cornering without centrifugal force" through specific data processing procedures and physical models.

[0087] When the rule engine detects a significant difference between the GPS-calculated displacement and the IMU-integrated displacement within a certain time period based on the vehicle dynamics model, which constitutes a Level 1 alarm, it will submit the corresponding chain of evidence data slices to the causal analysis module.

[0088] A typical data slicing structure is summarized in Table 1 below: Table 1:

[0089] As can be seen from the slice, within 30 seconds, the GPS data depicted a clear turning trajectory with a speed of approximately 55-58 km / h and a heading angle change of approximately 90 degrees. However, the CAN vehicle speed provided by the vehicle's CAN bus remained at 0, and the values ​​of the gyroscope Z-axis angular velocity reflecting the yaw / turning rate and the lateral acceleration reflecting the centrifugal force were extremely small, close to the level of stationary noise.

[0090] Then, multi-dimensional features are extracted and aligned simultaneously.

[0091] The module first performs time synchronization interpolation on the evidence chain to ensure that all modal data are aligned at the millisecond level, and then extracts key physical feature sequences. For example, in this embodiment, it is necessary to extract spatiotemporal trajectory features and vehicle behavior features.

[0092] The spatiotemporal trajectory characteristics include: calculating the instantaneous turning radius R and centripetal acceleration A from continuous GPS points. The calculation formula is: A = V² / R, where V is the GPS velocity and R is the radius obtained by the geometric method of three consecutive points. The rate of change of heading angle is calculated as the theoretical yaw rate.

[0093] Vehicle behavior characteristics include: CAN signal characteristics: vehicle speed and rotational speed that remain at 0; and physical characteristics proposed in the IMU: directly measured lateral acceleration, yaw rate, and the change in heading angle obtained by integrating the yaw rate. The lateral displacement Δy is estimated by performing a double integral on the lateral acceleration and combining it with other axial accelerations.

[0094] The extracted features are then compared across modalities to quantify the contradictions.

[0095] In this embodiment, contradiction A is manifested as the lack of centripetal acceleration.

[0096] At time t=10:03:15, according to GPS trajectory calculations, the vehicle is turning at V =57 km / h (≈15.8 m / s) with a radius R =45 meters.

[0097] According to Newton's laws of motion, the theoretical centripetal acceleration that the vehicle should experience at this moment is: A = V² / R = (15.8)² / 45 ≈ 5.55 m / s² ≈ 0.57g. However, the actual measured lateral acceleration at this moment is approximately 0.02 m / s² (≈0.002g).

[0098] The quantification value of the contradiction is: |5.55 – 0.02| ≈ 5.53 m / s², which is far beyond the error of the IMU sensor, which is usually <0.05g and within the normal body roll compensation range.

[0099] Contradiction B is manifested in the lack of yaw motion. The GPS heading angle changes by approximately 90 degrees within 30 seconds, and the average yaw rate should be 3 degrees / second. The yaw rate directly measured by the IMU fluctuates within ±0.2 degrees / second, and the heading angle change obtained by integration is less than 2 degrees, which is significantly different from 90 degrees.

[0100] The contradiction point C reflects the vehicle's power system status. The CAN bus speed and RPM are 0, indicating that, according to feedback from the vehicle's electronic system, the vehicle is not moving and the engine is not outputting power.

[0101] Based on the aforementioned contradictions, an initial set of hypotheses is generated: High probability H1: The GPS signal is forged by a high-precision simulator, generating a virtual dynamic trajectory that conforms to the road shape; Low probability H2: The IMU sensor is completely malfunctioning, but the probability of multiple axis data simultaneously failing completely and outputting near-zero values ​​is extremely low. Medium probability H3: The vehicle is lifted and runs on a roller "running machine," with the wheels spinning freely, but the CAN vehicle speed being 0 negates this hypothesis, unless the CAN signal is also deeply injected and tampered with.

[0102] Then, multi-source evidence fusion and hypothesis testing are performed.

[0103] The module calls other data sources to verify the above assumptions: verify H1, simulate GPS: check signal metadata: query the signal-to-noise ratio in the original GPS message to see the number of visible satellites and the continuity of satellite IDs. Simulated signals often have a stable number of satellites but abnormal IDs, or abnormally uniform SNR.

[0104] Check network location consistency: During the same period, are the cellular network base station location or Wi-Fi location coordinates collected by the terminal consistent with the GPS trajectory? If the network location shows that the vehicle is at a fixed point, such as a garage, while the GPS is moving, then H1 is strongly supported.

[0105] Invoking the environmental digital fingerprint status confirms that the terminal hardware itself has not been disassembled, ruling out the possibility that the device was moved for the purpose of the crime, and focusing the focus of the problem on the signal level.

[0106] Excluding H2 / H3: IMU self-test and context verification: Check the Z-axis acceleration data of the IMU reflecting vertical vibration within the same time period. If the vehicle is on a "running machine," there should be vibrations at a specific frequency; if it is truly stationary, then it is an environmental noise pattern. In this example, the IMU data pattern matches the characteristics of a truly stationary vehicle.

[0107] In-depth CAN signal analysis: Examine other CAN signals besides vehicle speed, such as wheel speed pulse signals and transmission gear positions. If all are zero or empty values, and consistent with the IMU being stationary, the possibility of actual vehicle movement is further ruled out.

[0108] Then, based on the scenario reconstruction and final decision, and integrating all the analysis, the module reconstructs the "storyline": Most likely scenario: The vehicle is actually stationary (CAN speed = 0, IMU shows it is stationary), and the ignition switch is on (IGN = ON). The cheater used a professional GPS simulator to inject a preset "virtual training trajectory" containing high-speed cornering into the timing terminal. Since the vehicle did not actually move, the IMU could not detect any corresponding lateral acceleration or yaw motion.

[0109] The final decision logic chain is as follows: If GPS shows high-speed movement and significant changes in trajectory curvature; and (IMU-measured centripetal acceleration ≈ 0 and yaw rate ≈ 0); and (theoretical centripetal acceleration >> measured value + reasonable error tolerance); and (CAN vehicle speed ≈ 0 and engine output is zero); and (network positioning contradicts GPS trajectory or GPS signal metadata is abnormal); and (environmental digital fingerprint is normal, device is not removed from vehicle), then it is determined that: trajectory forgery cheating was performed using an external high-precision GPS signal simulator. Confidence level = 0.99.

[0110] Finally, an interpretable adjudication report is output. The module outputs a structured and interpretable adjudication report, which can be directly used for supervision. The output adjudication report is as follows: Adjudication ID: ADJ_20231027_100300_001, Conclusion: Invalid training hours; Cheating type: GPS signal simulation trajectory forgery, Confidence level: 0.99; Main contradiction: Detection of an irreconcilable conflict between high-dynamic GPS trajectory and the static state of physical sensors.

[0111] Key evidence summary period: 10:03:00-10:03:30, GPS track shows that a 90-degree turn with a radius of about 45 meters and a speed of 55-58 kilometers per hour was completed.

[0112] Physical contradiction: Theoretically, the centripetal acceleration should be ≥0.57g, but the measured lateral acceleration of the IMU is <0.002g, with the deviation exceeding the threshold by 500%.

[0113] Vehicle status: The CAN bus continuously reports that the vehicle speed is 0 and the engine speed is 0.

[0114] Supporting evidence: The network positioning coordinates clustered at a fixed point P during the same period, which is inconsistent with the GPS dynamic trajectory.

[0115] Device status: Environmental digital fingerprint verification passed, ruling out overall device relocation.

[0116] Evidence chain hash anchor: 0x89ab...cdef, recommended action: immediately freeze the learning hours for this period, trigger a high-level alarm, and recommend conducting an on-site technical inspection.

[0117] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0118] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices or units, or may be electrical, mechanical or other forms of connection.

[0119] The units described as separate components may or may not be physically separate. As will be appreciated by those skilled in the art, the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0120] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0121] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or grid device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0122] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A time-based training anti-cheating system based on trusted roots and behavioral spatiotemporal chains, characterized in that, include: A timing training terminal is fixedly installed in a training vehicle. The timing training terminal has a built-in Trusted Execution Environment (TEE) module and a Beidou / GPS dual-mode timing module. The vehicle status perception terminal, connected via the vehicle's OBD-II interface, is used to collect vehicle VIN code, CAN bus signals, and gyroscope data. The Trusted Execution Environment (TEE) module is used to generate an environmental digital fingerprint for the timing training terminal and to encrypt and sign the collected positioning data, time data, and data from the vehicle status perception terminal. The BeiDou / GPS dual-mode timing module provides a unified time reference for all data. The system also includes a cloud-based collaborative adjudication platform, used to: receive signed chained data packets, which contain multiple data units linked in chronological order and signed by the TEE; perform logical consistency verification between the spatiotemporal trajectory data and vehicle behavior data in the chained data packets based on a preset physical rule model; and determine the existence of simulated positioning fraud through a multimodal causal analysis module when the spatiotemporal trajectory data indicates a change in vehicle status, but the vehicle behavior data does not indicate a corresponding change in physical status.

2. The time-based training anti-cheating system based on trusted roots and behavioral spatiotemporal chains according to claim 1, characterized in that, The environmental digital fingerprint is generated by the Trusted Execution Environment (TEE) module during terminal initialization by combining the unique device identifier of the timing training terminal, the vehicle VIN code successfully acquired for the first time, and the initial positioning coordinates, and is then bound to the cloud platform for two-way authentication.

3. The time-based training anti-cheating system based on trusted roots and behavioral spatiotemporal chains according to claim 1, characterized in that, Each data unit in the chained data packet includes at least: the currently collected spatiotemporal data and vehicle behavior data, the digital signature of the current data unit, and the hash value of the previous data unit, to form an immutable chained evidence storage structure.

4. The time-based training anti-cheating system based on trusted roots and behavioral spatiotemporal chains according to claim 1, characterized in that, The physical rule model includes vehicle dynamics constraint rules, which are used to verify whether the deviation between the displacement change calculated from positioning data and the displacement change calculated from the integration of gyroscope data within the same time period is within a reasonable threshold range.

5. The time-based training anti-cheating system based on trusted roots and behavioral spatiotemporal chains according to claim 1, characterized in that, The multimodal causal analysis module is also configured to trigger a secondary verification process based on electronic fences and historical reliable locations when the vehicle behavior data indicates that the vehicle is in a driving state and the spatiotemporal trajectory data indicates that the vehicle position has not changed.

6. A time-based training anti-cheating method based on trusted roots and behavioral spatiotemporal chains, characterized in that, The anti-cheating system for timed training according to any one of claims 1-5, the method comprising: terminal initialization and trusted binding, generating an environmental digital fingerprint through the Trusted Execution Environment (TEE) module and completing two-way authentication binding with the cloud platform to establish a hardware trusted root; chain-based evidence storage of teaching process data, periodically and synchronously collecting spatiotemporal data and vehicle behavior data based on a unified time benchmark during training, encapsulating them into chain-based data units within the TEE module and uploading them after encryption and signing; cloud-based collaborative adjudication, receiving and verifying the integrity and signature validity of the chain-based data units; performing logical consistency analysis on the spatiotemporal trajectory and vehicle behavior in the data units based on a physical rule model; and performing multimodal causal analysis to generate an anti-cheating adjudication result when a physical logical contradiction is detected between the spatiotemporal data and the behavior data.

7. The anti-cheating method for timing training based on trusted roots and behavioral spatiotemporal chains according to claim 6, characterized in that, The terminal initialization and trusted binding steps specifically include: the Trusted Execution Environment (TEE) module securely reading and storing the vehicle VIN code and the device unique identifier; obtaining initial high-confidence positioning coordinates; and uploading the combined hash value of the VIN code, device unique identifier, and initial positioning coordinates as the environment digital fingerprint to the cloud platform to complete registration.

8. The anti-cheating method for timing training based on trusted roots and behavioral spatiotemporal chains according to claim 6, characterized in that, In the chain-based evidence storage step of the teaching process, when encapsulating each chain data unit, the hash value of the previous data unit needs to be embedded to form a data chain that is related to the previous and subsequent data units.

9. The anti-cheating method for timing training based on trusted roots and behavioral spatiotemporal chains according to claim 6, characterized in that, In the cloud-based collaborative adjudication step, the execution of multimodal causal analysis includes: if the analysis results show that the abnormal changes in the spatiotemporal trajectory lack corresponding vehicle behavior support, it is determined to be external simulated signal cheating; if the analysis results show that the vehicle behavior shows normal driving but the spatiotemporal trajectory is abnormally stagnant, it is determined whether it is normal training under signal obstruction environment by combining the status of the environmental digital fingerprint and the electronic fence information.

10. The anti-cheating method for timing training based on trusted roots and behavioral spatiotemporal chains according to claim 6, characterized in that, The method also includes an adjudication feedback step: feeding back the anti-fraud adjudication result and the corresponding chain of data evidence to the regulatory terminal, and providing an authenticity verification interface based on data hash value.

Images