Method and system for monitoring unfair driving behavior of racing track vehicle

Abstract

The invention discloses a method and a system for monitoring unfair driving behaviors of racing track vehicles. The method comprises the steps that motion data, at least including acceleration, of a racing track vehicle is collected, and a G force base line for the racing track vehicle and a racing track combination of the racing track vehicle is established according to the motion data when the racing track vehicle completes first-circle driving; calculating the real-time deviation between the current G force of the racing track vehicle and the G force base line in the subsequent circle; when the real-time deviation has an abrupt change exceeding a preset threshold value, it is determined that the racing track vehicle is an illegal vehicle and a punishment instruction is output, and the punishment instruction is used for conducting power intervention on the racing track vehicle and / or triggering at least one visual prompt and / or audible prompt through a vehicle end control interface; and when the relative position of the illegal vehicle relative to the victim vehicle reaches a deviation correction condition, cancelling the punishment instruction. The penalty-correction-revocation closed loop disclosed by the invention can meet dual requirements on stability and fairness.

Description

Technical Field

[0001] This invention relates to the automotive field, and more specifically, to a method and system for monitoring unfair driving behavior of racetrack vehicles. Background Technology

[0002] With the rapid development of the automotive industry, battery technology, and electronic control units (ECUs), various track experiences, club tournaments, and the increasingly popular "Track Day" commercial events have greatly enriched the global market. Drivers and fans can not only experience the extreme handling of high-performance vehicles on closed tracks but also test the effectiveness of chassis modifications and software calibrations. However, track driving is fast-paced, high-speed, and involves numerous cornering combinations, making collisions or off-track accidents highly likely. Traditional track supervision relies mainly on manual flag signals, fixed cameras, and track broadcasting teams. In cases of unfair driving behaviors such as "intentional collisions," "line-cutting," and "brake tests," on-duty referees often cannot obtain timely and accurate kinematic evidence, leading to misjudgments or missed calls. Furthermore, due to delayed penalties, offending drivers often cause irreversible damage to others during the race. Therefore, there is an urgent need for a technology that can quantify vehicle dynamics in real time and provide immediate feedback to ensure the fairness and safety of motorsport while also meeting the low-cost and easy-to-deploy requirements of amateur track activities.

[0003] Currently, while high-level racing events (such as F1, WEC, and Formula E) utilize onboard ECUs, high-speed cameras, and AI event replay systems, these systems rely on expensive, complex hardware such as full-vehicle wiring harnesses, differential GPS, millimeter-wave speed radar, and pose fiber optic gyroscopes. For club cups or commercial experience races, these solutions are not only costly but also require the track to establish a robust base station network and a high-precision RTK-based base station network. Furthermore, the large volume of data necessitates post-race adjudication by professionals, making it difficult to provide penalty signals within seconds on-site. Some mobile IMU-based single-vehicle "driving control scoring" apps have emerged, but their algorithms often use fixed thresholds for lateral / longitudinal G-value judgments, lacking large-sample learning for different vehicle chassis, tire grip, and track curvature. In addition, they fail to consider gradient, attitude errors, and multi-lap tire temperature drift, resulting in high false alarm rates and rendering them unsuitable for official adjudication. Moreover, these apps generally only provide scores after driving, failing to achieve real-time power intervention with the vehicle's ECU and electronic control system. Therefore, there is currently a lack of a complete solution that is cost-effective, requires no track infrastructure, adapts to different vehicles and tracks, and can enforce penalties in real time.

[0004] In summary, the key technological bottlenecks are as follows: First, how to perform personalized calibration of the handling limits of different vehicle models and different tracks without relying on expensive differential positioning or fixed camera networks; second, how to accurately distinguish between "normal limit driving" and instantaneous dynamic anomalies caused by "intentional unfair intervention" within a millisecond-level calculation cycle; and third, how to deeply couple the detection results with the vehicle's power system to achieve a "penalty-correction-revocation" closed-loop control without introducing additional unstable factors. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention aims to provide a method and system for monitoring unfair driving behavior of racetrack vehicles.

[0006] According to one aspect of the present invention, a method for monitoring unfair driving behavior of a track vehicle is provided, comprising the following steps S102-S108. S102: Acquiring motion data of the track vehicle, including at least acceleration, and establishing a G-force baseline for the track vehicle and its track combination based on the motion data during the first lap of the track vehicle. S104: In subsequent laps, calculating the real-time deviation between the current G-force of the track vehicle and the G-force baseline, wherein G-force is defined as the L2 norm of a specific force after removing gravity and normalizing with standard gravitational acceleration. S106: When the real-time deviation exhibits a sudden change exceeding a predetermined threshold, determining that the track vehicle is a violating vehicle and outputting a penalty instruction, the penalty instruction being used to intervene in the power of the track vehicle and / or trigger at least one visual and / or audible prompt through the vehicle-side control interface. S108: When the relative position of the violating vehicle relative to the victim vehicle meets the correction conditions, revoking the penalty instruction.

[0007] According to another aspect of the present invention, a monitoring system for unfair driving behavior of a track vehicle is also provided, comprising an inertial measurement unit 22, a G-force calculation unit 24, a penalty execution module 26, and a penalty cancellation module 28. The inertial measurement unit 22 is used to collect motion data including at least acceleration; the G-force calculation unit 24 is used to establish a G-force baseline for the track vehicle and its track combination based on the motion data during the first lap of the track vehicle; and to calculate the real-time deviation between the current G-force of the track vehicle and the G-force baseline in subsequent laps, wherein G-force is defined as the L2 norm of a specific force after removing gravity and normalizing with standard gravitational acceleration; the penalty execution module 26 is used to determine that the track vehicle is a violating vehicle and output a penalty command when the real-time deviation has a sudden change exceeding a predetermined threshold, the penalty command being used to intervene in the power of the track vehicle and / or trigger at least one visual and / or audible prompt through the vehicle-side control interface; and the penalty cancellation module 28 is used to cancel the penalty command when the relative position of the violating vehicle relative to the victim vehicle reaches the correction condition.

[0008] In this embodiment of the invention, the vehicle's built-in inertial measurement unit automatically samples the entire first lap to obtain the "legitimate limit" distribution of the corresponding track without requiring manual input of vehicle parameters. This baseline naturally incorporates the combined effects of vehicle weight, aerodynamics, tire grip, and track curvature, significantly reducing cross-vehicle and cross-track errors. Simultaneously, the technical chain of first-lap learning—real-time mutation detection—closed-loop intervention avoids expensive external sensors and network dependencies, providing a low-cost, portable, and real-time enforcement solution for track activities, effectively solving the core challenges of existing technologies. In this embodiment, the penalty instruction can instantly limit torque or speed, or trigger visual / audible warnings. After the offending vehicle yields or returns to the safety line, the system determines the "correction condition" based on real-time geometric relationships and removes the penalty, thus forming a closed loop of "penalty—correction—revocation." This maintains driving safety without continuously interfering with normal competition, satisfying both stability and fairness requirements. Attached Figure Description

[0009] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0010] Figure 1 This is a flowchart of a method for monitoring unfair driving behavior of track vehicles according to an embodiment of the present invention;

[0011] Figure 2 This is a structural diagram of a monitoring system for unfair driving behavior of racetrack vehicles according to an embodiment of the present invention; and

[0012] Figure 3 This is a detailed structural diagram of a monitoring system for unfair driving behavior of racetrack vehicles according to an example of the present invention. Detailed Implementation

[0013] The following embodiments are merely examples to clearly illustrate the present invention and are not intended to limit the implementation of the invention. Those skilled in the art can make other variations or modifications based on the following description, and these variations, modifications, substitutions, and alterations arising from the principles and spirit of the present invention still fall within the protection scope of the present invention.

[0014] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other.

[0015] This invention provides a method for monitoring unfair driving behavior of racetrack vehicles. Figure 1This is a flowchart of a method for monitoring unfair driving behavior of racetrack vehicles according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps S102-S108.

[0016] S102, collect motion data of the track vehicle including at least acceleration, and establish a G-force baseline for the track vehicle and its track combination based on the motion data during the first lap of the track vehicle.

[0017] S104, in subsequent laps, calculate the real-time deviation between the vehicle's current G-force and the G-force baseline, where G-force is defined as the L2 norm of a specific force after removing gravity and normalizing it to standard gravitational acceleration.

[0018] S106, when the real-time deviation has a sudden change exceeding a predetermined threshold, the track vehicle is determined to be a violating vehicle and a penalty instruction is output. The penalty instruction is used to intervene in the power of the track vehicle and / or trigger at least one visual and / or audible prompt through the vehicle-side control interface.

[0019] S108, when the relative position of the offending vehicle to the victim vehicle meets the correction conditions, the penalty order shall be revoked.

[0020] In related technologies, it is necessary to adaptively identify and deal with "unfair driving" in real time for different vehicle / track conditions without relying on human referees and external infrastructure.

[0021] In this embodiment of the invention, the vehicle's built-in inertial measurement unit is used to automatically sample the entire first lap driving process, so as to obtain the "justifiable limit" distribution of the corresponding track without the need for manual input of vehicle parameters. This baseline naturally includes the combined effects of vehicle weight, aerodynamics, tire grip and track curvature, which can significantly reduce cross-vehicle and cross-track errors.

[0022] Meanwhile, through the technical chain of first-lap learning—real-time mutation detection—closed-loop intervention, it not only avoids expensive external sensing and network dependence, but also provides a low-cost, portable, and real-time law enforcement complete solution in track activity scenarios, effectively solving the core problems of existing technologies.

[0023] In this embodiment of the invention, the penalty instruction can either instantly limit torque or speed, or trigger a visual / audible warning; after the offending vehicle completes the yielding or returns to the safety line, the system determines the "correction condition" based on real-time geometric relationships and lifts the penalty, thus forming a closed loop of "penalty-correction-revocation", which maintains driving safety and does not continuously interfere with normal competition, satisfying the dual requirements of stability and fairness.

[0024] According to an embodiment of the present invention, the G force is calculated according to the following steps:

[0025] (1) Establish the vehicle coordinate system {e x ,e y ,e z Let the original output of the accelerometer be a. raw (t);

[0026] (2) From the attitude matrix R wb (t) will be subject to gravity g w =[0,0,g0] T Projecting the force onto the vehicle system and removing it yields a specific force f. b (t)=a raw (t)-R wb (t)g w ;

[0027] (3) The resultant force G is obtained by normalizing the standard gravitational acceleration g0.

[0028] In the above formula,

[0029] a raw (t): Raw output of the accelerometer, in m·s -2 ,

[0030] t is a time variable, in seconds (s).

[0031] R wb (t): World → Vehicle attitude rotation matrix, dimensionless.

[0032] g w Gravity vector [0,0,g0]T, unit m·s -2 ,

[0033] g0: Standard gravitational acceleration, in m·s² -2 ,

[0034] f b (t): The specific force after removing gravity, in m·s. -2 ,

[0035] G(t): The composite force G, dimensionless.

[0036] In this embodiment of the invention, a dimensionless G-force is obtained through L2 normalization and compared with a G-force baseline in real time. This allows the algorithm to focus only on "relative abrupt changes" rather than absolute values, thus being compatible with dynamic factors such as tire temperature and grip decay over laps. Once a deviation exceeding a threshold is detected, the controller immediately outputs a penalty command, compressing the reaction time to violations to sub-second levels and achieving "rapid and accurate identification" of unfair driving. In this embodiment of the invention, the attitude matrix R... wb (t) Obtained from the attitude calculation of the onboard IMU (quaternion or Euler angles), the gravity vector is taken as... By projecting the data onto the vehicle system in the world coordinate system and then subtracting it, the ambiguity of "gravity direction selection" can be avoided.

[0037] According to an embodiment of the present invention, when |G(t)-G base (s)|>η k The penalty instruction is output in the above formula.

[0038] G(t): The resultant force G, dimensionless.

[0039] G base (s): G-force baseline, dimensionless.

[0040] Predetermined threshold η k Refer to the following formula for recursion:

[0041] η k+1 =(1-α)η k +αG rms (k), where,

[0042] η k Preset threshold, dimensionless.

[0043] α: Update coefficient, 0 < α < 1, dimensionless.

[0044] G rms (k): The root mean square of the filtering force G in the k-th period, dimensionless.

[0045] Furthermore, when G win (t;Δ)>G min The penalty instruction is output in time, wherein,

[0046]

[0047]

[0048] In the above formula,

[0049] T: Threshold update period, in seconds.

[0050] Δ: Length of the sliding window, in seconds.

[0051] G win (t; Δ): Root mean square of the filtering force G within the window, dimensionless.

[0052] y(τ) represents the filtered force vector, which is the output obtained by convolving the impulse response function h(τ) with the force fb(t). ||y(τ)|| represents the L2 norm of the filtered force vector y(τ), which is the adaptive threshold η. k Provides a reference so that the threshold automatically drifts with changes in operating conditions such as grip and tire temperature.

[0053] In this embodiment of the invention, considering that fixed thresholds are difficult to adapt to drifting conditions and that single-point noise is prone to false triggering, the predetermined threshold normalization achieves scale uniformity and is insensitive to occasional noise. Therefore, the penalty instruction can either instantaneously limit torque or speed, or trigger a visual / audible warning; after the violating vehicle completes the yielding or returns to the safety line, the system determines the "correction condition" based on real-time geometric relationships and removes the penalty, thus forming a closed loop of "penalty-correction-revocation," which maintains driving safety without continuously interfering with normal competition, satisfying the dual requirements of stability and fairness.

[0055] According to an embodiment of the present invention, the correction conditions include the following three items.

[0056] (1)||r(t)||≤d thr

[0057] (2)

[0058] (3) Duration ≥ τ min

[0059] In the above formula,

[0060] r(t): The relative position vector of the offending vehicle relative to the victim vehicle, in meters.

[0061] φ(t): The relative heading angle of the offending vehicle to the victim vehicle, in rad.

[0062] d thr Preset distance threshold, in meters.

[0063] φ thr Preset heading angle threshold, in rad.

[0064] τ min : The preset minimum duration, in seconds.

[0066] Furthermore, the corrective conditions also include:

[0067]

[0068] In the above formula,

[0069] ω(t): Yaw angular velocity of the offending vehicle relative to the victim vehicle, in rad·s -1 ,

[0070] ω base (t): The baseline yaw rate of the offending vehicle relative to the victim vehicle, in rad·s. -1 ,

[0071] C coh (t; Δ): A consistency indicator of the violation vehicle relative to the victim vehicle, in rad·s. -1 ,

[0072] C max Preset coherence threshold, in rad·s -1 .

[0073] In related technologies, automatically canceling penalties based solely on a fixed "ebb time" has inherent flaws: on the one hand, violating vehicles can "wear out" the timer by temporarily slowing down or deliberately increasing the distance, without actually yielding or restoring a safe geometric relationship; on the other hand, in complex curve combinations, the victim vehicle often reverses its relative position with the violating vehicle, and simply timing cannot determine whether the trajectories of the two vehicles have re-intersected, thus lacking a causal relationship with the violation.

[0074] This invention applies a dual spatial-temporal criterion to cancellation actions by simultaneously introducing three "geometric constraints"—relative distance, relative heading, and duration—and a trajectory coherence index: the system only releases power intervention when the offending vehicle is physically in a yielding posture and maintains a stable and gentle heading change within the time window. This not only prevents strategic evasion such as "cutting in immediately after the timer ends" but also avoids unnecessary performance losses caused by prolonged penalties, ultimately forming a causal closed loop, significantly reducing the "false cancellation" rate and enhancing the credibility of track enforcement.

[0075] According to an embodiment of the present invention, the penalty instruction and race parameters are broadcast via a point-to-point wireless link between vehicles and reported to a ground base station. The instrument panel on the base station side is used for parameter tuning, penalty management, and audit traceability. When the base station is unavailable, the track vehicle coordinates with a decentralized multi-hop wireless relay network, wherein the decentralized multi-hop wireless relay network satisfies end-to-end opportunity constraints.

[0076] P(delay≤τ,delivery≥ρ min )≥1-ε

[0077]

[0078] K k =P k|k-1 H T HP k|k-1 H T +R) -1

[0079] In the above formula,

[0080] ε: Pre-set tolerance rate for default, dimensionless

[0081] τ: Predefined upper bound for latency, in seconds.

[0082] ρ min : The pre-set minimum delivery success rate, dimensionless

[0083] Posterior state estimation

[0084] Prior state estimation

[0085] K k Kalman gain

[0086] H: Observation matrix

[0087] P k|k-1 Prior covariance

[0088] R: Observation noise covariance

[0089] y k Observation vector

[0090] In related technologies, multi-hop or congested wireless links are prone to random latency and packet loss in track pits, tunnels, or spectator stands where they are obstructed; if penalty or revocation commands cannot be executed within the time limit τ with a success rate ρ min If the vehicle is not properly secured, the system may miss the optimal window for locking or unlocking it, potentially leading to a secondary collision. Meanwhile, measurements of the relative position, speed, and heading between vehicles are also affected by noise interference from IMU drift, GNSS multipath, and visual obstruction. If decisions are made solely based on instantaneous thresholds to determine whether a vehicle has yielded or should continue to be penalized, it's easy to misjudge a normal overtaking maneuver as a correction completed and prematurely cancel it.

[0091] To address the aforementioned uncertainties, this invention introduces a chance constraint P(delay≤τ, delivery≥ρ) at the communication layer. min The minimum reliability guarantee (≥1-ε) ensures that the law enforcement link has a quantifiable minimum reliability guarantee. Kalman fusion is employed at the perception layer, using Bayesian filtering to continuously and smoothly output high-confidence relative state estimates from IMU, GNSS, and vehicle-to-vehicle ranging data. This dual approach of lower bound on communication reliability and fusion estimation simultaneously suppresses link jitter and measurement noise amplification, significantly improving the accuracy and safety margin of penalty triggering and revocation decisions.

[0092] For Kalman and state space, let the state dimension be n, the observation dimension be m, the input dimension be r, and the output dimension be p, then

[0093] According to embodiments of the present invention, a monitoring system for unfair driving behavior of racetrack vehicles is also provided. Figure 2This is a structural diagram of a monitoring system for unfair driving behavior of racetrack vehicles according to an embodiment of the present invention, as shown below. Figure 2 As shown, the system includes an inertial measurement unit 22, a G-force calculation unit 24, a penalty execution module 26, and a penalty cancellation module 28. These will be described in detail below.

[0094] The inertial measurement unit 22 is used to collect motion data including at least acceleration; the G-force calculation unit 24 is used to establish a G-force baseline for the track vehicle and its track combination based on the motion data during the first lap of the track vehicle; and to calculate the real-time deviation between the current G-force of the track vehicle and the G-force baseline in subsequent laps, wherein G-force is defined as the L2 norm of a specific force after removing gravity and normalizing with standard gravitational acceleration; the penalty execution module 26 is used to determine that the track vehicle is a violating vehicle and output a penalty instruction when the real-time deviation has a sudden change exceeding a predetermined threshold, the penalty instruction is used to intervene in the power of the track vehicle, and / or trigger at least one visual and / or audible prompt through the vehicle-end control interface; and the penalty cancellation module 28 is used to cancel the penalty instruction when the relative position of the violating vehicle relative to the victim vehicle meets the correction conditions.

[0095] In related technologies, it is necessary to adaptively identify and deal with "unfair driving" in real time for different vehicle / track conditions without relying on human referees and external infrastructure.

[0096] In this embodiment of the invention, the vehicle's built-in inertial measurement unit is used to automatically sample the entire first lap driving process, so as to obtain the "justifiable limit" distribution of the corresponding track without the need for manual input of vehicle parameters. This baseline naturally includes the combined effects of vehicle weight, aerodynamics, tire grip and track curvature, which can significantly reduce cross-vehicle and cross-track errors.

[0097] Meanwhile, through the technical chain of first-lap learning—real-time mutation detection—closed-loop intervention, it not only avoids expensive external sensing and network dependence, but also provides a low-cost, portable, and real-time law enforcement complete solution in track activity scenarios, effectively solving the core problems of existing technologies.

[0098] In this embodiment of the invention, for a specific force after filtering, y(t) = (h*f) is defined. b (t)=∫h(τ)f b (t-τ)dτ, to ensure that y(t) and f b (t) has the same dimensions, so let the dimension of h(τ) be s. -1 Furthermore, the gain of the frequency response H(jω) is dimensionless; if matrix filtering is used, then it is stipulated that...

[0099] In this embodiment of the invention, the penalty instruction can either instantly limit torque or speed, or trigger a visual / audible warning; after the violating vehicle completes the yielding or returns to the safety line, the system determines the "correction condition" based on real-time geometric relationships and lifts the penalty, thereby forming a closed loop of "penalty-correction-revocation", which maintains driving safety and does not continuously interfere with normal competition, thus meeting the dual requirements of stability and fairness in problem ③.

[0100] According to an embodiment of the present invention, the G-force calculation unit calculates the force according to the following steps:

[0101] (1) Establish the vehicle coordinate system {e x ,e y ,e z Let the original output of the accelerometer be a. raw (t);

[0102] (2) From the attitude matrix R wb (t) will be subject to gravity g w =[0,0,g0] T Projecting the force onto the vehicle system and removing it yields a specific force f. b (t)=a raw (t)-R wb (t)g w ;

[0103] (3) The resultant force G is obtained by normalizing the standard gravitational acceleration g0.

[0104]

[0105] In the above formula,

[0106] a raw (t): Raw output of the accelerometer, in m·s -2 ,

[0107] t is a time variable, in seconds (s).

[0108] R wb (t): World → Vehicle attitude rotation matrix, dimensionless.

[0109] g w Gravity vector [0,0,g0] T Unit m·s -2 ,

[0110] g0: Standard gravitational acceleration, in m·s² -2 ,

[0111] f b (t): The specific force after removing gravity, in m·s. -2 ,

[0112] G(t): The resultant force G, dimensionless;

[0113] In this embodiment of the invention, a dimensionless G-force is obtained through L2 normalization and compared with a G-force baseline in real time. This allows the algorithm to focus only on "relative abrupt changes" rather than absolute values, thus being compatible with dynamic factors such as tire temperature and grip decay over laps. Once a deviation exceeding a threshold is detected, the controller immediately outputs a penalty command, compressing the reaction time to violations to sub-second levels and achieving "rapid and accurate identification" of unfair driving. In this embodiment of the invention, the attitude matrix R... wb (t) Obtained from the attitude calculation of the onboard IMU (quaternion or Euler angles), the gravity vector is taken as... By projecting the data onto the vehicle system in the world coordinate system and then subtracting it, the ambiguity of "gravity direction selection" can be avoided.

[0114] The penalty execution module is in the case of |G(t)-G base (s)|>η k The penalty instruction will be output at that time.

[0115] In the above formula,

[0116] G(t): The resultant force G, dimensionless.

[0117] G base (s): G-force baseline, dimensionless.

[0118] The predetermined threshold η k Refer to the following formula for recursion:

[0119] η k+1 =(1-α)η k +αG rms (k)

[0120] In the above formula,

[0121] η k This predetermined threshold is dimensionless.

[0122] α: Update coefficient, 0 < α < 1, dimensionless.

[0123] G rms (k): The root mean square of the filtering force G in the k-th period, dimensionless.

[0125] The penalty execution module further applies when G win (t;Δ)>G min The penalty instruction will be output at that time.

[0126] In the above formula,

[0127]

[0128]

[0129] In the above formula,

[0130] T: Threshold update period, in seconds.

[0131] Δ: Length of the sliding window, in seconds.

[0132] G win (t;Δ): The root mean square of the filtering force G within the window, dimensionless.

[0133] In this embodiment of the invention, considering that fixed thresholds are difficult to adapt to drifting conditions and that single-point noise is prone to false triggering, the predetermined threshold normalization achieves scale uniformity and is insensitive to occasional noise. Therefore, the penalty instruction can either instantaneously limit torque or speed, or trigger a visual / audible warning; after the violating vehicle completes the yielding or returns to the safety line, the system determines the "correction condition" based on real-time geometric relationships and removes the penalty, thus forming a closed loop of "penalty-correction-revocation," which maintains driving safety without continuously interfering with normal competition, satisfying the dual requirements of stability and fairness.

[0134] 10. The system according to claim 8 or 9, characterized in that the correction conditions of the penalty revocation module include:

[0135] (1)||r(t)||≤d thr

[0136] (2)

[0137] (3) Duration ≥ τ min

[0138] In the above formula,

[0139] r(t): The relative position vector of the offending vehicle with respect to the victim vehicle, in meters.

[0140] φ(t): The relative heading angle of the offending vehicle to the victim vehicle, in rad.

[0141] d thr Preset distance threshold, in meters.

[0142] φ thr Preset heading angle threshold, in rad.

[0143] τ min : The preset minimum duration, in seconds.

[0145]

[0146] In the above formula,

[0147] ω(t): The yaw rate of the offending vehicle relative to the victim vehicle, in rad·s. -1 ,

[0148] ω base (t): The baseline yaw rate of the offending vehicle relative to the victim vehicle, in rad·s. -1 ,

[0149] C coh (t; Δ): A consistency indicator of the offending vehicle relative to the victim vehicle, in rad·s. -1 ,

[0150] C max Preset coherence threshold, in rad·s -1 .

[0151] In related technologies, automatically canceling penalties based solely on a fixed "ebb time" has inherent flaws: on the one hand, violating vehicles can "wear out" the timer by temporarily slowing down or deliberately increasing the distance, without actually yielding or restoring a safe geometric relationship; on the other hand, in complex curve combinations, the victim vehicle often reverses its relative position with the violating vehicle, and simply timing cannot determine whether the trajectories of the two vehicles have re-intersected, thus lacking a causal relationship with the violation.

[0152] This invention applies a dual spatial-temporal criterion to cancellation actions by simultaneously introducing three "geometric constraints"—relative distance, relative heading, and duration—and a trajectory coherence index: the system only releases power intervention when the offending vehicle is physically in a yielding posture and maintains a stable and gentle heading change within the time window. This not only prevents strategic evasion such as "cutting in immediately after the timer ends" but also avoids unnecessary performance losses caused by prolonged penalties, ultimately forming a causal closed loop, significantly reducing the "false cancellation" rate and enhancing the credibility of track enforcement.

[0153] To describe in detail the specific implementation process of the above embodiments, the present invention also provides an example.

[0154] Figure 3 This is a structural diagram of a monitoring system for unfair driving behavior of racetrack vehicles according to an example of the present invention. For ease of description, each functional unit in the diagram is represented in block diagram form, but in actual products, it can be implemented by ASIC, DSP, MCU, SoC, or a combination thereof, or by multiple circuit boards distributed in different locations on the vehicle. To avoid duplication with other embodiments in the specification, only the distinctive details corresponding to the accompanying drawings are described below; the remaining steps and functions can be found in the foregoing claims and embodiments.

[0155] The track vehicle 10 can be a gasoline-powered go-kart, an electric go-kart, a Formula One car, or a touring car. The vehicle chassis power supply system provides 12V / 24V DC power to the embedded monitoring device (electric models can also directly draw 48V from the high-voltage battery and step it down via a DC-DC converter module, and a CAN or LIN bus is reserved so that this system can read key ECU parameters such as engine speed, throttle opening, and motor torque).

[0156] The integrated embedded system 20 is encapsulated in an IP65 aluminum alloy housing measuring approximately 90mm × 60mm × 25mm. This housing is mounted to the side wall of the seat via a four-point elastic shock-absorbing bracket, which reduces vibration and facilitates maintenance. The integrated embedded system 20 includes:

[0157] Main control SoC: adopts dual-core ARM Cortex-A55 + Cortex-M33 architecture Structure; High-speed kernel running Linux RT kernel executes G-force computing and communication protocol stack, The low-speed core runs the bare-metal program to manage power and watchdog timer.

[0158] High-reliability storage: 8GB eMMC for recording raw IMU data and penalty dates. The system includes an OTA firmware package; the onboard 16MB QSPI Nor Flash memory stores the BootLoader. With redundant firmware.

[0159] Power management: Wide voltage Buck-Boost (9–60V) + supercapacitor. Provides voltage drop buffer for instantaneous torque limiting action.

[0160] Safety isolation: Three-way optocouplers isolate power intervention from CAN FD transceiver The signal ensures that an MCU malfunction does not affect the vehicle's original ECU decisions.

[0161] Embedded software 30 consists of four layers:

[0162] Hardware Abstraction Layer (HAL): Encapsulates IMU, GNSS, UWB, Wheel-Speed, Steering-Angle and other drivers; periodically package raw 6-DoF IMU data (2 (kHz) and body bus data (100Hz).

[0163] Core algorithm layer: The G-force calculation unit runs in the first lap phase, at 20ms intervals. Periodic calculation f b (t) and update the G-force baseline distribution histogram (200 bins). (following the loop) The unfair driving behavior detector operates in stages, completing deduplication and FFT within a 2ms cycle. Band-limited, L2 normalization and threshold recursion η k When ΔG>η k And window energy G win >G min The penalty execution module is triggered at any time.

[0164] Communication stack layer: adopts Wi-Fi 6Mesh + BLE 5.3 dual-mode; BLE broadcasts 20 bytes of penalty metadata for low-power listening by neighboring vehicles, and Wi-Fi Mesh is responsible for uploading full logs to the base station or upper-level drone relay and receiving OTA.

[0165] Fault tolerance and OTA layer: dual-partition A / B gray-scale upgrade, with CRC32 and RSA-2048 signature verification enabled during the upgrade process; if the startup counter > 3, the previous stable version will be rolled back.

[0166] The vehicle-mounted sensor 40 includes the following units:

[0167] IMU combined inertial navigation: triaxial gyroscope + triaxial accelerator up to 2000dps / 16g; Internal data transmission synchronous temperature compensation, noise density <70μg / Hz.

[0168] Dual-frequency GNSS & RTK: used for global trajectory alignment, post-race replay, and safety fencing; automatically switches to inertial navigation for dead calculation in obstructed areas of the track.

[0169] UWB ranging nodes: measure vehicle-to-vehicle or vehicle-to-pillar distances with an accuracy of 3–5 cm, improving the Kalman observation dimension for relative position estimation.

[0170] Steering angle / wheel speed sensor: used for sideslip angle calculation and collision intensity reconstruction.

[0171] Behavior execution module 50 includes the following units:

[0172] Power intervention: Sends torque limiting command to ECU (limits throttle in fuel vehicles, reduces inverter duty cycle in electric vehicles); insertion delay <15ms.

[0173] HMI prompts: LED light strip flashing in three colors, steering wheel vibration motor, and multi-tone buzzer combination indicate the level of violation; providing real-time feedback to the driver.

[0174] Failsafe: If communication or algorithm errors are frequently triggered, the watchdog enters Grace-Mode—only retaining HMI prompts and not issuing torque limits; to prevent accidental power reduction leading to rear-end collisions.

[0175] The track operation terminal device 60 and user interface 70 include the following units:

[0176] The 60 operational devices are ruggedized and equipped with Wi-Fi 6E<EA. A tablet-type device running the Race-Control Dashboard based on Qt / QML:

[0177] The main interface displays the penalty status (green / yellow / ) of all participating vehicles in real time. (Red block), G-force baseline deviation heatmap.

[0178] The "Parameter Tuning" page allows for remote distribution of α, T, Δ, and d parameters. thr φ thr It includes more than 30 calibration parameters and supports one-click switching between different competition groups.

[0179] The "Replay" page is indexed by vehicle and time, using L2 norm trajectories and IMU data. Data and penalty logs enable post-match review and accident evidence collection.

[0180] In summary, according to the above embodiments of the present invention, a method and system for monitoring unfair driving behavior of a track vehicle are provided. The method includes: S102, collecting motion data of the track vehicle, including at least acceleration, and establishing a G-force baseline for the track vehicle and its track combination based on the motion data during the first lap of the track vehicle. S104, in subsequent laps, calculating the real-time deviation between the current G-force of the track vehicle and the G-force baseline, wherein G-force is defined as the L2 norm of a specific force after removing gravity and normalizing with standard gravitational acceleration. S106, when the real-time deviation exhibits a sudden change exceeding a predetermined threshold, determining that the track vehicle is a violating vehicle and outputting a penalty instruction, which is used to intervene in the power of the track vehicle and / or trigger at least one visual and / or audible prompt through the vehicle-side control interface. S108, when the relative position of the violating vehicle relative to the victim vehicle meets the correction conditions, revoking the penalty instruction. This invention determines the "correction conditions" and removes penalties based on real-time geometric relationships, thereby forming a closed loop of "penalty-correction-revocation". This maintains driving safety without continuously interfering with normal competition, thus meeting the dual requirements of stability and fairness.

[0181] This invention provides a unified declaration regarding symbols and units.

[0182] Coordinates and Rotation:

[0183] R wb (t)——World→Vehicle system attitude rotation matrix (dimensionless);

[0184]

[0185] Signal:

[0186] a raw (t)(m·s *2 ),

[0187] f b (t)=a raw (t)-R wb (t)g w (m·s -2 ),

[0188] y(t)=(h*f b (t)(m·s) -2 ),

[0189] h(τ)(s -1 ),

[0190] H(jω) (dimensionless).

[0191] Scalar:

[0192] g0 standard gravity (m·s) -2 ),

[0193] G(t) = ||f b || / g0 (dimensionless)

[0194] G rms G win G base ,η k (Dimensionless).

[0195] Time / Window:

[0196] T(s), Δ(s), τ, τ min (s).

[0197] geometry:

[0198] r(t)(m), d thr (m), C coh C max (rad·s -1 ).

[0199] Probability and Communication: ρ min ,ε (dimensionless).

[0200] Dimensions:

[0201] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. 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 method for monitoring unfair driving behavior of racetrack vehicles, characterized in that, include: Collect motion data of the track vehicle, including at least acceleration, and establish a G-force baseline for the track vehicle and its track combination based on the motion data during the first lap of the track vehicle. In subsequent laps, the real-time deviation between the current G-force of the track vehicle and the G-force baseline is calculated, where G-force is defined as the L2 norm of a specific force after removing gravity and normalizing with standard gravitational acceleration. When the real-time deviation has a sudden change exceeding a predetermined threshold, the track vehicle is determined to be a violating vehicle and a penalty instruction is output. The penalty instruction is used to intervene in the power of the track vehicle and / or trigger at least one visual and / or audible prompt through the vehicle-side control interface. as well as When the relative position of the offending vehicle to the victim vehicle meets the correction conditions, the penalty order is revoked.

2. The method according to claim 1, characterized in that, The force G is calculated according to the following steps: (1) Establish the vehicle coordinate system {e x ,e y ,e z Let the original output of the accelerometer be a. raw (t); (2) From the attitude matrix R wb (t) will be subject to gravity g w =[0,0,g0] T Projecting the force onto the vehicle system and removing it yields a specific force f. b (t)=a raw (t)-R wb (t)g w ; (3) The resultant force G is obtained by normalizing the standard gravitational acceleration g0. In the above formula, a raw (t): Raw output of accelerometer, in m·s -2 , t is a time variable, in seconds (s). R wb (t): World → Vehicle attitude rotation matrix, dimensionless. g w Gravity vector [0,0,g0]T, unit m·s -2 , g0: Standard gravitational acceleration, in m·s² -2 , f b (t): The specific force after removing gravity, in m·s. -2 , G(t): The composite force G, dimensionless.

3. The method according to claim 2, characterized in that, When |G(t)-G base (s)|>η k The penalty instruction will be output in time. In the above formula, G(t): The resultant force G, dimensionless. G base (s): G-force baseline, dimensionless. The predetermined threshold η k Refer to the following formula for recursion: or k+1 =(1-a)n k +αG rms (k) In the above formula, η k The predetermined threshold is dimensionless. α: Update coefficient, 0 < α < 1, dimensionless. G rms (k): The root mean square of the filtering force G in the k-th period, dimensionless.

4. The method according to claim 3, characterized in that, When G win (t;Δ)>G min The penalty instruction will be output in time. In the above formula, In the above formula, T: Threshold update period, in seconds. Δ: Length of the sliding window, in seconds. G win (t; Δ): Root mean square of the filtering force G within the window, dimensionless. y(t) represents the filtered specific force vector, which is the output obtained by convolving the specific force fb(t) with the impulse response function h(t).

5. The method according to claim 1, characterized in that, The correction conditions include: ||r(t)||≤d thr Duration ≥ τ min In the above formula, r(t): The relative position vector of the offending vehicle relative to the victim vehicle, in meters. The relative heading angle of the offending vehicle with respect to the victim vehicle, in rad. d thr Preset distance threshold, in meters. Preset heading angle threshold, in rad. τ min : The preset minimum duration, in seconds.

6. The method according to claim 5, characterized in that, The correction conditions also include: In the above formula, ω(t): The yaw rate of the offending vehicle relative to the victim vehicle, in rad·s. -1 , ω base (t): The baseline yaw rate of the offending vehicle relative to the victim vehicle, in rad·s. -1 , C coh (t; Δ): The consistency index of the violating vehicle relative to the victim vehicle, in rad·s -1 , C max Preset coherence threshold, in rad·s -1 .

7. The method according to any one of claims 1-6, characterized in that, The penalty instructions and race parameters are broadcast via point-to-point wireless links between vehicles and reported to the ground base station. The instrument panel on the base station side is used for parameter setting, penalty management and audit traceability. When the base station is unavailable, the track vehicles cooperate with a decentralized multi-hop wireless relay network, wherein the decentralized multi-hop wireless relay network satisfies an end-to-end opportunity constraint: P(delay≤τ,delivery≥ρ min )≥1-e K k =P k|k-1 H T (HP k|k-1 H T +R) -1 In the above formula, ε: Pre-set permissible default rate, dimensionless τ: Predefined upper bound for latency, in seconds. ρ min : The pre-set minimum delivery success rate, dimensionless Posterior state estimation Prior state estimation K k Kalman gain H: Observation matrix P k|k-1 Prior covariance R: Observation noise covariance y k : Observation vector.

8. A monitoring system for unfair driving behavior of racetrack vehicles, characterized in that, include: An inertial measurement unit (IMU) is used to acquire motion data, including at least acceleration. The G-force calculation unit is used to establish a G-force baseline for the track vehicle and its track combination based on the motion data during the first lap of the track vehicle. And used in subsequent laps to calculate the real-time deviation between the current G-force of the track vehicle and the G-force baseline, wherein G-force is defined as the L2 norm of a specific force after removing gravity and normalizing with standard gravitational acceleration. The penalty execution module is used to determine that the track vehicle is a violating vehicle and output a penalty instruction when the real-time deviation has a sudden change exceeding a predetermined threshold. The penalty instruction is used to intervene in the power of the track vehicle and / or trigger at least one visual and / or audible prompt through the vehicle-side control interface. as well as The penalty cancellation module is used to cancel the penalty instruction when the relative position of the violating vehicle with respect to the victim vehicle meets the correction conditions.

9. The system according to claim 8, characterized in that, The G-force calculation unit calculates according to the following steps: (1) Establish the vehicle coordinate system {e x ,e y ,e z Let the original output of the accelerometer be a. raw (t); (2) From the attitude matrix R wb (t) will be subject to gravity g w =[0,0,g0] T Projecting the force onto the vehicle system and removing it yields a specific force f. b (t)=a raw (t)-R wb (t)g w ; (3) The resultant force G is obtained by normalizing the standard gravitational acceleration g0. In the above formula, a raw (t): Raw output of accelerometer, in m·s -2 , t is a time variable, in seconds (s). R wb (t): World → Vehicle attitude rotation matrix, dimensionless. g w Gravity vector [0,0,g0]T, unit m·s -2 , g0: Standard gravitational acceleration, in m·s² -2 , f b (t): The specific force after removing gravity, in m·s. -2 , G(t): The resultant force G, dimensionless; The penalty execution module when |G(t)-G base (S)|>η k The penalty instruction will be output in time. In the above formula, G(t): The resultant force G, dimensionless. G base (s): G-force baseline, dimensionless. The predetermined threshold η k Refer to the following formula for recursion: or k+1 =(1-a)n k +αG rms (k) In the above formula, η k The predetermined threshold is dimensionless. α: Update coefficient, 0 < α < 1, dimensionless. G rms (k): The root mean square of the filtering force G in the kth period, dimensionless; The penalty execution module further when G win (t;Δ)>G min The penalty instruction will be output in time. In the above formula, In the above formula, T: Threshold update period, in seconds. Δ: Length of the sliding window, in seconds. G win (t; Δ): Root mean square of the filtering force G within the window, dimensionless. y(t) represents the filtered specific force vector, which is the output obtained by convolving the specific force fb(t) with the impulse response function h(t).

10. The system according to claim 8 or 9, characterized in that, The correction conditions for the penalty revocation module include: ||r(t)||≤d thr Duration ≥ τ min In the above formula, r(t): The relative position vector of the offending vehicle relative to the victim vehicle, in meters. The relative heading angle of the offending vehicle with respect to the victim vehicle, in rad. d thr Preset distance threshold, in meters. Preset heading angle threshold, in rad. τ min : The preset minimum duration, in seconds. In the above formula, ω(t): The yaw rate of the offending vehicle relative to the victim vehicle, in rad·s. -1 , ω base (t): The baseline yaw rate of the offending vehicle relative to the victim vehicle, in rad·s. -1 , C coh (t; Δ): The consistency index of the violating vehicle relative to the victim vehicle, in rad·s -1 , C max Preset coherence threshold, in rad·s -1 .

Images