A race track scene-oriented auxiliary driving method, system, vehicle and medium

Abstract

This invention relates to the field of vehicle assisted driving technology, and discloses an assisted driving method, system, vehicle, and medium for track scenarios. The method includes: constructing a track map using environmental and vehicle data collected while the vehicle is driving on the track; determining a theoretically optimal path based on the track map, preset vehicle parameters, and tire adhesion limit constraints, wherein the theoretically optimal path includes suggested speed and suggested acceleration for each path; acquiring the vehicle's current state data, and generating driving guidance information based on the current state data and the theoretically optimal path; and projecting the driving guidance information into the driver's field of vision to assist track driving. This invention constructs a high-precision track map in real time, generates an optimal performance trajectory based on the track map, the vehicle's real-time state, and physical limits, and compares it with the vehicle's current state in real time to generate guidance information, thereby assisting the driver in track driving and greatly improving the efficiency of assisted driving in track scenarios.

Description

Technical Field

[0001] This invention relates to the field of vehicle assisted driving technology, specifically to an assisted driving method, system, vehicle, and medium for track scenarios. Background Technology

[0002] With the popularization of motorsport culture, more and more civilian vehicles are equipped with high-performance features to meet the demands of track driving. However, most drivers lack track driving experience, making it difficult to fully utilize the vehicle's performance and posing safety hazards. Furthermore, existing track driving assistance systems are mostly based on preset track data, unable to adapt to the complexities of different tracks, and lack real-time feedback, making it difficult to effectively improve the driver's skill level.

[0003] Based on this, related technologies propose a solution based on remote pre-configuration, in which a remote configurator identifies the road route and maps the vehicle's speed distribution and trajectory. The trajectory defines the vehicle's path around the road route, and the speed distribution minimizes the vehicle's lap time. However, this solution has limitations: it requires prior knowledge of track data and cannot adapt to temporarily changing track conditions, such as slippery surfaces or obstacles. In addition, active control intervention may interfere with driver autonomy and increase system complexity. Furthermore, it relies on remote communication, which is unreliable in areas with weak signals, such as mountain tracks.

[0004] Furthermore, the related technology also focuses on safety warnings. Specifically, it obtains the road surface adhesion coefficient, weight, and corner radius in real time after the vehicle leaves the first corner and calculates the safe speed. When the current speed exceeds the limit, it outputs a warning signal, solving the problem of speeding and loss of control in track driving and improving safety and handling precision. However, the limitations of this solution are that it heavily relies on pre-built high-precision maps and cannot adapt to new tracks or temporarily changed track conditions. Secondly, its function is essentially to calculate a fixed "safe speed" and trigger a braking alarm, making it merely a safety warning system rather than a performance guidance tool. Even though there are existing track guidance solutions based on augmented reality head-up displays, most of them can only project a single, fixed virtual driving trajectory. They do not provide multi-dimensional performance guidance by combining information such as driver operating habits, real-time vehicle status, and vehicle operating environment. The human-computer interaction is simplistic and lacks performance-oriented feedback.

[0005] In summary, there is an urgent need for a track-specific driver assistance technology that can perceive unknown or dynamically changing track environments in real time, dynamically generate optimal performance trajectories based on the vehicle's real-time status and physical limits, and guide the driver to perform operations in an intuitive and efficient manner, in order to fill the gaps in existing technologies in the above aspects. Summary of the Invention

[0006] This invention provides an assisted driving method, system, vehicle, and medium for track scenarios, to solve the problems mentioned in the above-mentioned technical background, such as the lack of real-time perception of unknown or dynamic changes in the track environment, the inability to generate optimal driving trajectories, resulting in low efficiency of assisted driving in track scenarios and difficulty in meeting the track assisted driving requirements of vehicle limits, real-time performance, safety, and personalized performance guidance.

[0007] In a first aspect, the present invention provides an assisted driving method for track scenarios, the method comprising: A track map is constructed using environmental and vehicle data collected while the vehicle is driving on the track. Based on the track map, preset vehicle parameters, and tire adhesion limit constraints, the theoretical optimal path is determined. The theoretical optimal path includes the suggested speed and suggested acceleration for each path. Acquire the vehicle's current status data and generate driving guidance information based on the current status data and the theoretical optimal path; This projects driving guidance information into the driver's field of vision to assist track driving.

[0008] This invention constructs a track map using environmental and vehicle data collected during track driving, providing a precise geographical and environmental foundation for subsequent optimal path calculations. Based on the track map and vehicle parameters, and incorporating tire adhesion limit constraints, it generates the optimal racing line that determines the upper limit of lap times during track driving. Simultaneously, the path is bound to suggested speeds and accelerations at all points, achieving a globally optimal path for both lateral driving and longitudinal acceleration / deceleration. This maximizes vehicle performance and helps drivers achieve theoretically optimal lap times. Furthermore, based on real-time comparison of the vehicle's current state data with the theoretically optimal path, it dynamically generates driving guidance information and projects it into the driver's field of vision. This not only identifies driver deviations in line, speed, and acceleration / deceleration timing in real time, providing early correction guidance to avoid the risk of loss of control due to accumulated deviations, but also significantly reduces safety hazards caused by distraction, preventing lap time losses due to inattention. This achieves a comprehensive improvement in the performance, safety, and interactive experience of track-assisted driving.

[0009] In one optional implementation, the environmental data includes point cloud data, and the vehicle data includes pose data; using the environmental data and vehicle data collected while the vehicle is driving on the track, a track map is constructed, including: Point cloud data and pose data are fused to generate a point cloud map; Identify and extract track features from point cloud maps. Track features include at least one of the following: track surface, track boundary, center line, shoulder, and guardrail. The track features are marked in the point cloud map to obtain the track map.

[0010] This invention eliminates the need for complex preliminary preparations, such as importing third-party track maps or manually marking track features. Upon first entering an unfamiliar track, only one lap is required to complete the mapping based on the collected environmental point cloud data and the vehicle's own pose data. Simultaneously, it extracts track-specific features such as track surface, track boundaries, center line, shoulders, and guardrails, and completes feature labeling, thereby obtaining a structured track map with clear semantics. This provides a high-precision, highly robust, and lightweight structured map base for subsequent track-assisted driving.

[0011] In one optional implementation, the preset vehicle parameters include at least the road surface adhesion coefficient; the tire adhesion limit constraint is constructed based on the tire friction circle theory; based on the track map, the preset vehicle parameters, and the tire adhesion limit constraint, the theoretically optimal path is determined, including: Establish an optimization function with the objective of minimizing the single-lap travel time; Based on the road surface adhesion coefficient, the maximum adhesion limit is determined; the vector sum of the vehicle's longitudinal acceleration and lateral acceleration is calculated, and the maximum adhesion limit is used as the tire adhesion limit constraint. The track boundaries are extracted from the track map, and the optimization function is solved under the condition of satisfying the tire adhesion limit constraint to obtain the theoretical optimal path, which includes the path point sequence and the suggested speed and suggested acceleration for each path.

[0012] This invention optimizes single-lap time to precisely meet the core requirements of track driving, ensuring that the driving strategy throughout the track is globally optimized around the "shortest time," maximizing vehicle track performance from the algorithmic level. Tire grip limits are dynamically calibrated based on real-time road surface adhesion coefficients, with the vector sum of longitudinal and lateral acceleration not exceeding the maximum grip limit as a rigid constraint. Combined with track boundary constraints, this not only accurately reproduces tire dynamics, achieving optimal grip utilization and rigid locking of safety boundaries, but also ensures the compliance and absolute drivability of the planned trajectory. Finally, the theoretically optimal path obtained includes suggested speeds and accelerations for each path point. This trajectory can be directly converted into precise driving operation guidance for the driver, eliminating the need for additional judgment on acceleration / deceleration timing and cornering speed. This significantly lowers the entry barrier for track driving and provides a precise operational optimization benchmark for advanced drivers.

[0013] In one optional implementation, acquiring the vehicle's current state data includes: The collected point cloud data is matched with the track map to determine the vehicle's location on the track map. Based on the inertial measurement unit data, vehicle bus data, and positioning results collected by the vehicle, the vehicle motion state is estimated to obtain the current state data of the vehicle, which includes at least position, velocity, acceleration, and yaw rate.

[0014] This invention directly obtains the absolute positioning result of the vehicle in the track coordinate system by matching real-time collected point cloud data with a pre-constructed track map, without relying on any external positioning signals. At the same time, it can make full use of features such as track guardrails, shoulders, and track boundaries to achieve stable matching. Even in adverse scenarios such as no signal, low light, rain, and fog, it can continuously output high-precision positioning results, achieving comprehensive coverage of various track scenarios and all working conditions. In addition, the full-dimensional motion state parameters output by estimating the vehicle's motion state can provide complete and accurate data for subsequent track-assisted driving.

[0015] In one optional implementation, driving guidance information is generated based on the current state data and the theoretically optimal path, including: Based on the speed in the current state data, the target guidance path is dynamically extracted from the theoretical optimal path. The length of the target guidance path is positively correlated with the speed. The target spatiotemporal state set is used to determine the target guidance path. The target spatiotemporal state set includes the target position, target heading angle, target velocity, target longitudinal acceleration, and target lateral acceleration. Based on the current state data and the target spatiotemporal state set, a performance deviation set is calculated. The performance deviation set includes lateral position deviation, heading angle deviation, velocity deviation, and acceleration envelope deviation. Among them, the acceleration envelope deviation is determined based on the current road surface adhesion coefficient and the tire friction circle theory. The current road surface adhesion coefficient is estimated using the currently collected point cloud data. Driving guidance information is generated based on the performance deviation set. The driving guidance information includes at least steering guidance information, speed management suggestions, and operation timing prompts. Among them, the steering guidance information is generated based on lateral position deviation and / or heading angle deviation to provide suggested steering wheel angle or steering correction trend. The speed management suggestions are determined based on speed deviation and acceleration envelope deviation to provide suggested speed and suggested acceleration. The operation timing prompts are generated based on relevant information of key operation points in the target guidance path to provide prompts for future key operation points. Key operation points include at least braking point, corner apex, corner entry point, and corner exit point.

[0016] This invention uses the vehicle's current speed as the core basis and dynamically adjusts the length of the forward path from the theoretically optimal path. The faster the vehicle speed, the longer the forward distance, which helps to achieve the optimal guidance effect in all scenarios. Furthermore, it constructs a full-dimensional performance deviation set covering lateral position, heading angle, speed, and acceleration envelope. In particular, based on the road adhesion coefficient estimated by real-time point cloud and the acceleration envelope deviation calculated by combining the tire friction circle theory, it can realize a full-dimensional quantitative evaluation of driving performance. It can not only accurately determine "how much the line deviates and how much the speed is different", but also clarify the efficiency of the current operation in utilizing grip. It provides accurate quantitative basis for guidance commands, which avoids the waste of vehicle performance by conservative driving and eliminates the risk of loss of control by exceeding the tire adhesion limit from the source of the algorithm. It significantly improves the assistance efficiency and user experience of track driving.

[0017] In one alternative implementation, driving guidance information is projected into the driver's field of vision to assist track driving, including: By using an in-vehicle display device, driving guidance information is visualized and rendered to obtain a virtual projected route; Based on the degree of deviation between the vehicle's actual driving trajectory and the theoretical optimal path, the visual attributes of the virtual projection route are dynamically adjusted. The visual attributes include at least one of color, brightness, or width.

[0018] This invention utilizes an in-vehicle display device to render driving guidance information as a virtual projection route, projecting it into the driver's field of vision. The driver can intuitively obtain all driving guidance information, completely eliminating the safety hazards caused by distraction under high-speed extreme conditions and significantly improving the safety level of track driving. At the same time, the visual attributes of the virtual projection route can be dynamically adjusted, further improving the robustness of guidance in all scenarios. The visual attributes can also be customized according to the user's visual sensitivity and driving habits, which helps to achieve a personalized guidance experience and adapt to the needs of different users.

[0019] In one alternative implementation, after projecting driving guidance information into the driver's field of vision to assist track driving, the track-oriented assisted driving method further includes: Record historical data of the vehicle during its driving on the track. The historical data includes the actual driving trajectory, vehicle status data, driver operation data, and the theoretical optimal path. Based on historical data, the process of generating the theoretically optimal path is iteratively optimized.

[0020] This invention is based on the iterative optimization of the theoretically optimal path generation process using full-dimensional historical data. This allows the path generation logic to perfectly match the actual physical characteristics of the vehicle from the source. In other words, the optimized theoretically optimal path will not result in conservative performance due to model bias, nor will it generate invalid trajectories that the vehicle cannot land on. This helps to continuously explore the vehicle's potential for extreme lap times. At the same time, by recording the driver's historical operation data, the driver's driving style can be accurately characterized, thereby iteratively optimizing the constraints and optimization objectives of path generation. This achieves a comprehensive improvement in the performance, safety, learning efficiency, and interactive experience of track-assisted driving.

[0021] Secondly, the present invention provides an assisted driving system for track scenarios, the system comprising: The module is used to build a track map using environmental and vehicle data collected while the vehicle is driving on the track. The determination module is used to determine the theoretically optimal path based on the track map, preset vehicle parameters, and tire adhesion limit constraints. The theoretically optimal path includes the suggested speed and suggested acceleration for each path. The generation module is used to acquire the vehicle's current status data and generate driving guidance information based on the current status data and the theoretical optimal path. The assistance module projects driving guidance information into the driver's field of vision to assist track driving.

[0022] This invention provides an assisted driving system for track scenarios. It utilizes a track map constructed from environmental and vehicle data collected during track driving, along with tire adhesion limit constraints, to generate a theoretically optimal path with suggested speeds and accelerations at all points. This maximizes vehicle performance and helps the driver achieve theoretically optimal lap times. Furthermore, based on real-time comparison of the vehicle's current state data with the theoretically optimal path, it dynamically generates driving guidance information and projects it into the driver's field of vision. This not only identifies driver deviations in line, speed, and acceleration / deceleration timing in real time, providing corrective guidance in advance to avoid the risk of loss of control due to accumulated deviations, but also significantly reduces safety hazards caused by distraction and prevents lap time losses due to inattention. This achieves a comprehensive improvement in the performance, safety, and interactive experience of track-assisted driving.

[0023] Thirdly, the present invention provides a vehicle, the vehicle including a controller, the controller including a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to perform the assisted driving method for track scenarios described in the first aspect or any corresponding embodiment thereof.

[0024] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the assisted driving method for a track scenario described in the first aspect or any corresponding embodiment thereof.

[0025] The assisted driving method and system for track scenarios provided by this invention are based on a real-time constructed high-precision track map and vehicle parameters, incorporating tire adhesion limit constraints to generate the theoretically optimal path that determines the upper limit of lap time in track driving. Simultaneously, the path is bound to suggested speeds and accelerations at all points, achieving a globally optimal path for lateral alignment and longitudinal acceleration / deceleration. This maximizes vehicle performance and helps the driver achieve the theoretically optimal lap time. Furthermore, based on real-time comparison of the vehicle's current state data with the theoretically optimal path, driving guidance information is dynamically generated and projected into the driver's field of vision. This not only identifies driver deviations in alignment, speed, and acceleration / deceleration timing in real time, providing corrective guidance in advance to avoid the risk of loss of control due to accumulated deviations, but also significantly reduces safety hazards caused by eye shifts and avoids lap time losses due to distraction. This achieves a comprehensive improvement in the performance, safety, and interactive experience of track-assisted driving. Attached Figure Description

[0026] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0027] Figure 1 This is a schematic diagram of the first type of assisted driving method for track scenarios according to an embodiment of the present invention; Figure 2 This is a schematic diagram of a second process for an assisted driving method for a racetrack scenario according to an embodiment of the present invention; Figure 3 This is a hardware architecture diagram of a driver assistance system; Figure 4 This is a flowchart of track map construction and positioning; Figure 5 This is a schematic diagram of path planning and visual guidance; Figure 6 This is a schematic diagram of track feature recognition and dynamic adjustment; Figure 7 This is a schematic diagram of data-driven adaptive optimization; Figure 8 This is a structural block diagram of a driver assistance system for track scenarios according to an embodiment of the present invention; Figure 9 This is a schematic diagram of the hardware structure of a vehicle according to an embodiment of the present invention. Detailed Implementation

[0028] 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, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] According to an embodiment of the present invention, an embodiment of an assisted driving method for track scenarios is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0030] This embodiment provides an assisted driving method for track scenarios. Figure 1 This is a schematic flowchart of a first type of assisted driving method for a track scenario according to an embodiment of the present invention, as shown below. Figure 1 As shown, the process includes the following steps: Step S101: Construct a track map using environmental and vehicle data collected while the vehicle is driving on the track.

[0031] In this embodiment, this step represents the vehicle using its own sensors to collect relevant data and autonomously construct a high-precision structured map of its own track. The specific content, acquisition methods, and map construction methods of the environmental and vehicle data can all be adaptively determined according to actual needs. For example, the vehicle travels a lap along the track, using its onboard LiDAR to collect environmental point cloud data (capturing track geometric features); it uses an Inertial Measurement Unit (IMU) or CAN bus (Controller Area Network) to collect vehicle pose, speed, and other data (recording the vehicle's own motion state); it employs the SLAM (Simultaneous Localization and Mapping) algorithm to align each frame of point cloud data with the vehicle pose data, compensating for point cloud motion distortion during high-speed cornering, and stitching them together to form a dense 3D point cloud map containing the entire track scene; it extracts track features, such as track boundaries and obstacles, and semantically labels these extracted features in the point cloud map, transforming it into a lightweight structured track map that can be directly called by upper-layer systems.

[0032] In one specific embodiment, a real-time high-precision mapping and localization algorithm based on extreme condition perception is proposed. This algorithm utilizes LiDAR and IMU data, employing a SLAM algorithm to generate a high-precision point cloud map of the vehicle's surrounding environment in real time, and simultaneously calculates the vehicle's centimeter-level precise position and attitude within the map, completely eliminating dependence on external signals. The SLAM process is specifically optimized for track features such as guardrails, shoulders, and cones to ensure robust localization under extreme conditions such as high-speed, high-G-value (reflecting the vehicle's extreme cornering capabilities) cornering. Specifically, the algorithm includes: 1. Point cloud acquisition and preprocessing for track feature enhancement.

[0033] In this embodiment, at the SLAM front end, the original point cloud of the LiDAR is subjected to track feature enhancement filtering, such as separating the track surface (low reflection) and guardrails / cones (high reflection) by using a reflection intensity threshold; linear features such as road shoulders and track edges are extracted based on the geometric features (curvature, normal) of the local point cloud; the extracted feature points and the original point cloud are simultaneously input into the subsequent registration and mapping process.

[0034] 2. SLAM backend optimization based on track structure constraints.

[0035] In this embodiment, the track loop is automatically identified during the mapping process, and track-specific constraint factors are introduced in the back-end optimization (such as graph-based optimization or factor graph-based SLAM framework), such as: 1) Loop closure constraint: forcing the track start and end points to be spatially closed; 2) Width consistency constraint: constraining the distance between the left and right boundaries of the track to remain constant in the straight and gentle curve areas; 3) Curvature smoothing constraint: performing spline fitting on the track centerline to ensure that the curvature is continuous and conforms to the reachable range of vehicle dynamics.

[0036] 3. Dynamic object recognition and removal.

[0037] In this embodiment, by using continuous multi-frame point cloud data and comparing the changes in the point cloud in the same area between adjacent frames, dynamic objects (such as moving vehicles and temporary obstacles) are identified and marked, and then removed when constructing the global map to ensure the static nature and long-term availability of the map.

[0038] 4. Motion distortion compensation and high-frequency positioning output.

[0039] In this embodiment, during each frame of lidar data acquisition, the angular velocity and acceleration data of the IMU during that time period are recorded synchronously. The vehicle's own motion during the radar scan is calculated by integration, and the point cloud of that frame is compensated for inverse motion to eliminate the stretching or compression of the point cloud caused by the vehicle's motion. The compensated point cloud is then used to match the global map to achieve centimeter-level high-precision positioning.

[0040] Step S102: Based on the track map, preset vehicle parameters, and tire adhesion limit constraints, determine the theoretical optimal path. The theoretical optimal path includes the suggested speed and suggested acceleration for each path.

[0041] In this embodiment, the preset vehicle parameters refer to the fixed / semi-fixed parameters that are pre-configured or calibrated and related to the vehicle's own physical characteristics. These parameters are used to accurately reflect the dynamic limits that the vehicle can reach during path planning, such as wheelbase, track width, center of gravity height, vehicle mass, moment of inertia, and road surface adhesion coefficient. The tire adhesion limit constraint determines how fast the vehicle can go, how sharply it can turn, and how hard it can brake without slipping. The theoretical optimal path represents the ideal driving trajectory that can achieve the fastest lap time on the track and is physically achievable by the vehicle.

[0042] Step S103: Obtain the current status data of the vehicle, and generate driving guidance information based on the current status data and the theoretical optimal path.

[0043] In practical applications of track driving, novice drivers often face the challenge of translating abstract track strategies and optimal racing lines into real-time operations. This presents a high barrier to entry, and traditional guidance lacks real-time deviation correction and grip limit control, resulting in either missing optimal operating opportunities and losing lap times or exceeding physical limits and causing loss of control. In this embodiment, the static theoretical optimal lap time trajectory is transformed into dynamic operating guidance that the driver can execute in real time and adapt to the extreme conditions of the track. This allows the driver to keep their eyes on the road throughout the entire process, accurately follow the optimal trajectory, maximize vehicle performance and reduce lap time within safe boundaries.

[0044] Step S104: Project driving guidance information into the driver's field of vision to assist track driving.

[0045] In this embodiment, this step aims to transform the generated abstract digital driving guidance instructions into augmented reality (AR) visual guidance that is precisely bound to the actual track surface and allows the driver to "keep their eyes on the road and intuitively perceive it," which is the final closed loop of human-computer interaction for track-assisted driving.

[0046] The track-oriented assisted driving method provided in this embodiment constructs a track map using environmental and vehicle data collected during track driving, providing a precise geographical and environmental foundation for subsequent optimal path calculation. Based on the track map and vehicle parameters, and incorporating tire adhesion limit constraints, it generates the optimal racing line that determines the upper limit of lap time during track driving. Simultaneously, the path is bound to suggested speeds and accelerations at all points, achieving a globally optimal path for lateral driving and longitudinal acceleration / deceleration. This maximizes vehicle performance and helps the driver achieve theoretically optimal lap times. Furthermore, based on real-time comparison of the vehicle's current state data and the theoretically optimal path, it dynamically generates driving guidance information and projects it into the driver's field of vision. This not only identifies driver deviations in line, speed, and acceleration / deceleration timing in real time, providing early correction guidance to avoid the risk of loss of control due to accumulated deviations, but also significantly reduces safety hazards caused by distraction, preventing lap time losses due to inattention. This achieves a comprehensive improvement in the performance, safety, and interactive experience of track-oriented assisted driving.

[0047] This embodiment provides an assisted driving method for track scenarios. Figure 2 This is a schematic diagram of a second type of assisted driving method for a track scenario according to an embodiment of the present invention, such as... Figure 2 As shown, the process includes the following steps: Step S201: Construct a track map using environmental and vehicle data collected while the vehicle is driving on the track.

[0048] In this embodiment, environmental data includes point cloud data, and vehicle data includes pose data. Point cloud data is a set of three-dimensional spatial coordinate points generated after the vehicle-mounted LiDAR scans the environment. Each point contains core information such as three-dimensional coordinates and laser reflection intensity. In a track scene, it can accurately reconstruct the geometric contours of the track and the surrounding static environment, and corely covers track-specific features such as the track surface, track boundaries, shoulders, guardrails, and buffer zones. Pose data is a set of core parameters describing the vehicle's spatial position and body posture in the track's global coordinate system, such as the vehicle's three-dimensional spatial coordinates, which reflect the vehicle's absolute position on the track map; and the vehicle's three-dimensional attitude angles, such as roll angle, pitch angle, and yaw angle, which reflect the vehicle's tilt state, pitch amplitude, and heading.

[0049] Specifically, step S201 includes: Step a1: Fuse the point cloud data with the pose data to generate a point cloud map.

[0050] In this embodiment, the SLAM algorithm is used to generate a point cloud map; see the previous text for details.

[0051] Step a2: Identify and extract track features from the point cloud map. Track features include at least one of the following: track surface, track boundary, center line, shoulder, and guardrail.

[0052] Step a3: Mark the track features in the point cloud map to obtain the track map.

[0053] In this embodiment, no complex preliminary work is required, such as importing third-party track maps or manually marking track features. Upon first entering an unfamiliar track, only one lap is needed to complete the mapping based on the collected environmental point cloud data and the vehicle's own pose data. At the same time, track scene-specific features such as track surface, track boundary, center line, road shoulder, and guardrail are extracted and labeled to obtain a structured track map with clear semantics. This provides a high-precision, highly robust, and lightweight structured map base for subsequent track-assisted driving.

[0054] Step S202: Based on the track map, preset vehicle parameters, and tire adhesion limit constraints, determine the theoretical optimal path. The theoretical optimal path includes the suggested speed and suggested acceleration for each path.

[0055] In this embodiment, the preset vehicle parameters include at least the road surface adhesion coefficient; the tire adhesion limit constraint is constructed based on the tire friction circle theory. It should be noted that the tire friction circle theory aims to provide physical constraints, that is, based on the preset road surface adhesion coefficient, the tire friction circle theory is used to lock that "the sum of the longitudinal and lateral acceleration vectors is less than or equal to the maximum adhesion limit", so as to avoid slippage and loss of control.

[0056] Specifically, step S202 includes: Step S2021: Establish an optimization function with the objective of minimizing the single-lap travel time.

[0057] Step S2022: Based on the road surface adhesion coefficient, determine the maximum adhesion limit; calculate the vector sum of the vehicle's longitudinal acceleration and lateral acceleration, and use the maximum adhesion limit as the tire adhesion limit constraint, where the vector sum does not exceed the maximum adhesion limit.

[0058] In this embodiment, the road surface adhesion coefficient represents the maximum frictional capacity between the tire and the road surface; the maximum adhesion limit represents the maximum combined acceleration that the vehicle can generate without slipping, which can be determined by multiplying the road surface adhesion coefficient and the acceleration due to gravity. Note that the calculation of the maximum adhesion limit is not theoretically unique, and there are more complex dynamic expressions that take into account factors such as vehicle load transfer, road surface slope, and dynamic changes in tire vertical load.

[0059] Step S2023: Extract the track boundary from the track map, and under the condition of satisfying the tire adhesion limit constraint, solve the optimization function to obtain the theoretical optimal path containing the path point sequence and the corresponding suggested speed and suggested acceleration of each path.

[0060] In this embodiment, the track boundary is extracted from the track map to ensure that the trajectory is always within the drivable area; in addition, numerical optimization algorithms, such as dynamic programming, model predictive control algorithms, pseudospectral methods, etc., can be used to solve the optimization function under the condition of satisfying the constraints.

[0061] In this embodiment of the invention, by minimizing the single-lap driving time as the optimization objective, it precisely aligns with the core requirements of track driving, ensuring that the driving strategy throughout the entire track is globally optimized around the "shortest time," maximizing the release of vehicle track performance from the algorithmic root. The tire adhesion limit is dynamically calibrated based on the real-time road surface adhesion coefficient, and a rigid constraint is imposed that the vector sum of longitudinal and lateral acceleration does not exceed the maximum adhesion limit. Combined with track boundary constraints, this not only accurately restores the tire dynamic characteristics, achieving optimal utilization of grip and rigid locking of safety boundaries, but also ensures the compliance and absolute drivability of the planned trajectory. Finally, the theoretically optimal path obtained includes the suggested speed and suggested acceleration for each path point. This trajectory can be directly converted into precise driving operation guidance that the driver can execute, eliminating the need for the driver to additionally judge acceleration / deceleration timing and cornering speed. This significantly lowers the entry barrier for track driving and provides a precise operational optimization benchmark for advanced drivers.

[0062] Step S203: Obtain the current state data of the vehicle, and generate driving guidance information based on the current state data and the theoretical optimal path.

[0063] In this embodiment, the current state data represents the vehicle's current location and motion state. Specifically, step S203 includes: Step S2031: Match the collected point cloud data with the track map to determine the vehicle's location on the track map.

[0064] Step S2032: Based on the inertial measurement unit data, vehicle bus data, and positioning results collected by the vehicle, the vehicle motion state is estimated to obtain the current state data of the vehicle. The current state data includes at least position, velocity, acceleration, and yaw rate.

[0065] It should be noted that existing assisted driving solutions only output single-position data, which cannot cover the full-dimensional requirements of trajectory tracking. This leads to defects such as incomplete deviation calculation and delayed / inaccurate guidance commands. The current state data output in this embodiment fully covers the vehicle's full-dimensional motion state parameters, including position, speed, acceleration, and yaw rate. Among them, the position data provides a spatial reference for calculating lateral position deviation; the speed and acceleration data provide quantitative basis for speed management and friction circle constraint verification; and the yaw rate accurately reflects the vehicle's steering dynamics and heading change trend, providing core support for the generation of steering guidance commands. This helps to improve the accuracy and real-time performance of driving guidance from the root, avoiding problems such as guidance lag and over / under-correction of steering.

[0066] In this embodiment of the invention, the absolute positioning result of the vehicle in the track coordinate system is directly obtained by matching the real-time collected point cloud data with the pre-constructed track map, without relying on any external positioning signal. At the same time, it can make full use of the features of track guardrails, shoulders, track boundaries and other features to achieve stable matching. Even in adverse scenarios such as no signal, low light, rain and fog, it can continuously output high-precision positioning results, achieving all-round coverage of various track scenarios and all working conditions. At the same time, the full-dimensional motion state parameters output by the vehicle motion state estimation can provide complete and accurate data for subsequent track-assisted driving.

[0067] Step S2033: Based on the speed in the current state data, dynamically extract the target guidance path from the theoretical optimal path. The length of the target guidance path is positively correlated with the speed.

[0068] It should be noted that existing driver assistance solutions often use a fixed look-ahead distance. This is insufficient at high speeds, leaving drivers with little reaction time to miss the optimal maneuver point; conversely, excessively long look-ahead distances at low speeds, such as when cornering, lead to redundant information interfering with driving and reducing guidance accuracy. This embodiment dynamically adjusts the look-ahead path length based on the vehicle's current speed, allowing for a longer look-ahead distance as the vehicle speed increases. This design provides ample window for driver prediction and reaction on high-speed straightaways, preventing guidance lag from the outset; it also shortens the look-ahead distance and reduces invalid information at low speeds on curves. Combined with denser sampling in high-curvature areas, this ensures precise guidance on curves, perfectly matching the dramatic changes in driving conditions—high-speed straightaways and low-speed cornering—achieving optimal guidance across all scenarios.

[0069] Step S2034: Determine the target spatiotemporal state set of the target guidance path. The target spatiotemporal state set includes the target position, target heading angle, target velocity, target longitudinal acceleration, and target lateral acceleration.

[0070] It's important to note that existing track guidance solutions address the industry pain point of only matching geometric path coordinates and being disconnected from the core objective of lap time optimization. In other words, existing solutions can only tell the driver "where to go," failing to match the optimal speed, acceleration, and heading required for optimal driving. This results in drivers following the correct line but misjudging cornering speed and acceleration / deceleration timing, either exceeding grip limits and losing control or wasting performance and sacrificing lap time. This embodiment extracts a target spatiotemporal state set that fully covers all dimensions of parameters, including target position, target heading angle, target speed, and target longitudinal / lateral acceleration. This design upgrades the core objective of guidance from a "static geometric route" to a "dynamic optimal driving state," achieving full-dimensional alignment between the vehicle's current state and the state required for the theoretically optimal lap time. This ensures that every guidance action directly serves the core objective of "minimizing single-lap time," completely resolving the fundamental problem of the disconnect between geometric guidance and lap time optimization.

[0071] Step S2035: Based on the current state data and the target spatiotemporal state set, calculate the performance deviation set. The performance deviation set includes lateral position deviation, heading angle deviation, velocity deviation, and acceleration envelope deviation. Among them, the acceleration envelope deviation is determined based on the current road surface adhesion coefficient and the tire friction circle theory. The current road surface adhesion coefficient is estimated using the currently collected point cloud data.

[0072] It should be noted that this embodiment constructs a full-dimensional performance deviation set covering lateral position, heading angle, velocity, and acceleration envelope. In particular, the deviation based on the road adhesion coefficient estimated by real-time point cloud and the acceleration envelope calculated by combining the tire friction circle theory can solve the problem of the single dimension of the existing deviation calculation, which can only evaluate the degree of deviation of the line and speed, but cannot quantify the driver's utilization efficiency of tire grip, and cannot determine whether the driving operation still has performance margin or is about to exceed the safety limit; and avoids the problem of the existing use of a fixed friction coefficient as the calculation benchmark, which makes the deviation calculation completely distorted when the track is locally wet, oily, or worn. This approach, which can easily lead to guidance errors and loss of vehicle control, enables a comprehensive quantitative assessment of driving performance. It can accurately determine "how much the driving line has deviated and how much the speed has changed," and also clearly define the efficiency of the current operation in utilizing grip. This provides precise quantitative basis for guidance commands, avoiding the waste of vehicle performance by conservative driving and eliminating the risk of loss of control due to exceeding the tire adhesion limit from the algorithm's source. In addition, based on the real-time road surface adhesion coefficient, the deviation benchmark can be dynamically adjusted to adapt to the dynamic changes in track surface conditions. Even if there are sudden changes in local road surface grip, it can still output accurate deviation assessments and safe guidance, greatly improving the environmental adaptability and functional safety of the solution.

[0073] Step S2036: Generate driving guidance information based on the performance deviation set. The driving guidance information includes at least steering guidance information, speed management suggestions, and operation timing prompts. The steering guidance information is generated based on lateral position deviation and / or heading angle deviation to provide suggested steering wheel angles or steering correction trends. The speed management suggestions are determined based on speed deviation and acceleration envelope deviation to provide suggested speed and suggested acceleration. The operation timing prompts are generated based on relevant information of key operation points in the target guidance path to provide prompts for future key operation points. The key operation points include at least braking points, apex of a curve, entry point, and exit point.

[0074] It should be noted that existing assisted driving solutions suffer from drawbacks such as single guidance instructions, lack of timing prediction, and inability to be directly converted into executable operations. Based on this, the driving guidance information generated in this embodiment simultaneously covers three dimensions: steering guidance information, speed management suggestions, and operation timing prompts. These three elements form a highly complementary and complete guidance loop. Specifically, steering guidance information is generated based on spatial position and heading angle deviation, clarifying the spatial dimension of "where the vehicle should go and how much to turn the steering wheel." Speed ​​management suggestions are generated based on speed deviation and acceleration envelope deviation, clarifying the intensity dimension of "whether to increase or decrease speed and the amount of force required for the operation." Operation timing prompts are generated based on key operation points in the forward-looking path, providing early warnings for core nodes such as braking points, entry points, apex points, and exit points, clarifying the temporal dimension of "when the vehicle should perform the operation." In essence, this is a three-in-one guidance instruction system encompassing "space, intensity, and timing." It can transform abstract professional track driving experience into concrete, fully executable operation instructions. Drivers no longer need to memorize complex track strategies or predict cornering rhythms; they can simply follow the guidance to complete standardized track driving, significantly lowering the entry barrier for track driving. Furthermore, the advance warning timing prompts provide drivers with ample reaction time in high-speed conditions, ensuring that they can perform core operations such as braking, steering, and acceleration at the optimal moment. This completely solves the pain points of delayed guidance and missing the best operation point in high-speed scenarios, ensuring driving safety while maximizing lap time compression. It helps advanced drivers precisely optimize operational details and break through lap time bottlenecks, covering the needs of users at all levels, from beginners to professional players.

[0075] In this embodiment of the invention, the current vehicle speed is used as the core basis to dynamically adjust the length of the forward path from the theoretically optimal path. The faster the vehicle speed, the longer the forward distance, which helps to achieve the optimal guidance effect in all scenarios. Furthermore, a full-dimensional performance deviation set covering lateral position, heading angle, speed, and acceleration envelope is constructed. In particular, the road adhesion coefficient estimated based on real-time point cloud and the acceleration envelope deviation calculated by combining the tire friction circle theory can realize a full-dimensional quantitative evaluation of driving performance. This avoids the waste of vehicle performance by conservative driving and eliminates the risk of loss of control due to exceeding the tire adhesion limit from the source of the algorithm, which greatly improves the assistance efficiency and user experience of track driving.

[0076] Step S204: Project driving guidance information into the driver's field of vision to assist driving on the track.

[0077] Specifically, step S204 includes: Step S2041: Using the in-vehicle display device, the driving guidance information is visualized and rendered to obtain a virtual projected route.

[0078] In practical applications, track driving speeds can reach over 200 km / h. If a driver's gaze leaves the road surface for even 0.5 seconds, the vehicle travels over 27 meters. A moment's hesitation to look down at the instrument panel or central control screen can lead to a fatal accident, such as veering off the track or losing control. Traditional guidance methods rely on numerical or graphical prompts from the central control screen and instrument panel, which cannot eliminate the risk of distraction caused by a shift in gaze. This embodiment utilizes in-vehicle display devices, such as a head-up display (HUD) or instrument panel, to seamlessly integrate the virtual projected route with the actual track surface in front of the driver. The driver does not need to shift their gaze or look down at any in-vehicle equipment; they can maintain focus on the track surface throughout the entire journey and intuitively obtain all driving guidance information. This completely eliminates the safety hazards caused by distraction under high-speed extreme conditions, significantly improving the safety level of track driving.

[0079] Step S2042: Based on the degree of deviation between the vehicle's actual driving trajectory and the theoretical optimal path, dynamically adjust the visual attributes of the virtual projection route. The visual attributes include at least one of color, brightness, or width.

[0080] In practical applications, static and fixed virtual guidance routes in track scenarios can only provide a fixed target driving path and cannot provide real-time feedback on the deviation status of driving operations. Drivers cannot intuitively judge the degree of their deviation or whether the correction is adequate, which can easily lead to lap time loss due to line deviation, or even safety accidents due to serious deviations going unnoticed in time. In this embodiment, the quantified degree of trajectory deviation is transformed into intuitive visual changes, forming a closed-loop guidance system throughout the entire process. For example, when the vehicle is perfectly aligned with the optimal path, the route is displayed in green (standard state); when there is a slight deviation, it switches to yellow and the brightness is increased; when there is a serious deviation, or when the vehicle is about to cross the boundary or exceed the grip limit, it switches to bright red and the display is widened. Specifically, through the above dynamic visual feedback, drivers can perceive driving deviations in real time and quickly and accurately correct steering and speed control operations. This not only significantly improves the tracking accuracy of the optimal trajectory, maximizes vehicle performance and reduces lap time, but also corrects dangerous operations in a timely manner through strong visual warnings, rigidly avoiding the safety risks of running off the track and losing vehicle control from the human-computer interaction level.

[0081] In this embodiment of the invention, driving guidance information is rendered as a virtual projection route using an in-vehicle display device and projected into the driver's field of vision. The driver can intuitively obtain all driving guidance information, completely eliminating the safety hazards caused by distraction under high-speed extreme conditions and significantly improving the safety level of track driving. At the same time, the visual attributes of the virtual projection route can be dynamically adjusted, further improving the robustness of guidance in all scenarios. The visual attributes can also be customized according to the user's visual sensitivity and driving habits, which helps to achieve a personalized guidance experience and adapt to the usage needs of different users.

[0082] In practical applications, the optimal driving path estimated by assisted driving solutions for track scenarios inherently deviates from the actual performance of the vehicle during driving. Factors such as tire wear, brake fade, delayed power response, and changes in vehicle parameters after modifications can all cause the theoretically optimal path to be "seemingly optimal but actually unexecutable," either resulting in a conservative approach that wastes performance or exceeding the vehicle's physical limits. Furthermore, existing universal theoretical optimal paths are based on the "standard answer" of ideal driving operations, failing to consider the differences in driving habits, reaction speeds, vehicle control abilities, and risk preferences among different drivers. For example, novice drivers cannot perform extreme late braking maneuvers, while advanced drivers have specific preferences for cornering / exit rhythms. Forcibly applying universal paths either leads to excessive driver stress and increases the risk of safety issues, or the guidance instructions conflict with driver habits and have low user acceptance. Therefore, after projecting driving guidance information into the driver's field of vision to assist track driving, this embodiment of the assisted driving method for track scenarios also includes: Step b1: Record the vehicle's historical data during its journey on the track. The historical data includes the actual driving trajectory, vehicle status data, driver operation data, and the theoretical optimal path.

[0083] In this embodiment, by recording the driver's historical operation data, such as steering wheel angle timing, accelerator / brake opening, operation response delay, and vehicle control precision, the driver's driving style, ability boundaries, and operating habits can be accurately characterized. This allows for iterative optimization of the constraints and optimization goals of path generation. For example, for novice drivers, the path tolerance range can be appropriately relaxed, the limit acceleration constraint reduced, and the early warning window for operation timing optimized to improve the safety and ease of execution of the guidance. For advanced drivers, the speed matching for entering / exiting corners and the braking point timing can be optimized to match their preferred driving rhythm, maximizing the adaptation to their operating habits within the safety boundary and helping them break through lap time bottlenecks. Ultimately, this achieves an upgrade from "general standardized guidance" to "personalized customized guidance," thus covering the needs of users at all levels, from beginners to professional players.

[0084] Step b2 involves iteratively optimizing the generation process of the theoretically optimal path based on historical data.

[0085] In this embodiment, by accumulating historical data from multiple laps, the difference between theoretical planning and actual vehicle performance can be accurately compared. This allows for iterative optimization of the underlying core elements of path generation, including correcting vehicle dynamics model parameters, optimizing the modeling accuracy of tire adhesion limit constraints, updating the calibration logic of the road adhesion coefficient, and improving the weight configuration of the minimum lap time objective function. This fundamentally ensures that the path generation logic fully aligns with the vehicle's actual physical characteristics. Specifically, the optimized theoretically optimal path avoids both conservative performance due to model bias and invalid trajectories that prevent the vehicle from landing, achieving "planning is execution, and execution achieves optimal performance," continuously unlocking the vehicle's ultimate lap time potential.

[0086] The optimal path generation process based on the theory of iterative optimization using full-dimensional historical data in this embodiment of the invention can ensure that the path generation logic fully conforms to the real physical characteristics of the vehicle from the root, and can accurately characterize the driver's driving style. It can further iteratively optimize the constraints and optimization objectives of path generation, thereby achieving a comprehensive improvement in the performance, safety, learning efficiency, and interactive experience of track-assisted driving.

[0087] It should be noted that, addressing the three core pain points of track driving—"lack of pre-installed map adaptation," "dynamic environment response," and "personalized performance enhancement"—this embodiment discloses an adaptive track driving assistance system and solution. This system requires no pre-installed track map and can use LiDAR and IMU to perceive and construct a track environment model in real time. It dynamically generates the optimal trajectory with vehicle dynamic limits as constraints and provides immersive, personalized performance guidance and real-time feedback through an Augmented Reality Head-Up Display (AR-HUD), thereby optimizing lap times and improving driver skills. Specifically, it constructs a technical architecture centered on "extreme performance exploration, dynamic environment adaptation, and personalized evolution." Through real-time LiDAR perception, dynamic limit constraints, immersive interaction, and closed-loop learning, it breaks through the safety-oriented logic of traditional assisted driving, significantly improving the safety and efficiency of assisted driving.

[0088] To further explain, the driving assistance system in this embodiment includes three core modules, and the functions and connections of each module are as follows: 1. Perception Module: It includes at least one automotive-grade 3D LiDAR (mounted at the center of the vehicle's roof, fixed with a shock absorber bracket) and one six-axis IMU (mounted near the vehicle's center of gravity); the LiDAR is used to collect point cloud data of the track environment (sampling frequency ≥10Hz, ranging accuracy ≤2cm), and the IMU is used to collect vehicle attitude data (angular velocity accuracy ≤0.01° / s, acceleration accuracy ≤0.05m / s²); both are connected to the central computing unit via Ethernet to achieve microsecond-level synchronization of timestamps.

[0089] 2. Central Computing Unit: Adopts automotive-grade ECU to execute core algorithms, including: SLAM algorithm, track feature extraction algorithm, road surface adhesion coefficient estimation algorithm, friction circle constraint trajectory planning algorithm, deviation calculation and guidance command generation algorithm, data recording and model optimization algorithm; reads vehicle dynamic parameters and real-time status (wheel speed, steering angle, etc.) through CAN bus.

[0090] 3. Human-computer interaction module: It includes at least one AR-HUD system (projection distance ≥ 5m, resolution ≥ 1920×1080) and one central control screen; the AR-HUD is used to project multimodal guidance information, and the central control screen is used for mode switching (such as mapping mode, auxiliary mode), data viewing (lap time, G-value, deviation) and system settings.

[0091] In one specific embodiment, Figure 3This is a hardware architecture diagram of an assisted driving system. As shown in the diagram, the system includes front-end perception devices such as LiDAR and IMU, providing raw data on the environment and the vehicle itself. The core computing unit (ECU) is the decision-making and computational core of the entire system, integrating four functional modules to complete the entire process from data to guidance strategy. Among these, the track SLAM and localization module fuses LiDAR point cloud and IMU pose data to autonomously construct a high-precision structured map of the track using SLAM algorithms, while simultaneously achieving real-time centimeter-level absolute positioning of the vehicle in the track coordinate system, providing spatial reference and positioning support for the entire system. The vehicle dynamics model module calibrates the vehicle's inherent physical characteristics and dynamic boundaries, constructing an adhesion limit constraint model based on the tire friction circle theory to accurately reproduce the vehicle's power, braking, and steering limits, providing rigid physical constraints for optimal trajectory planning and ensuring that the generated trajectory is feasible and does not exceed the vehicle's safety limits. The performance-optimized trajectory generator is the core strategy unit of the system, aiming to minimize... With single-lap time as the optimization objective, based on the track map and vehicle dynamics model, and under the dual constraints of track boundaries and tire adhesion limits, the system solves and generates the optimal spatiotemporal theoretical track trajectory (ideal racing line) that includes spatial path, suggested speed and acceleration at each point. A personalized learning engine is used to achieve self-learning and iterative optimization of the system. By recording vehicle driving history data and driver operation data, it iteratively optimizes the vehicle dynamics model and optimal trajectory generation logic to adapt to different drivers' driving styles, vehicle control abilities and operating habits, achieving an upgrade from general guidance to personalized customization. The head-up display (HUD) system, which outputs human-computer interaction, receives driving guidance information output by the central computing unit. Through AR technology, it accurately projects virtual guidance routes, operation sequence prompts, speed management suggestions and other content into the driver's field of vision, blending them with the actual track surface to achieve a zero-distraction immersive track driving assistance where "eyes don't leave the road and hands don't leave the wheel." It is the core interactive bridge between algorithm strategy and driver operation.

[0092] To further explain, the driving assistance solution in this embodiment includes the following core steps: Step 1: Real-time high-precision mapping and localization based on extreme condition perception: Utilizing LiDAR and IMU data, a high-precision point cloud map of the vehicle's surrounding environment is generated in real time using the SLAM algorithm. Simultaneously, the vehicle's centimeter-level precise position and attitude within this map are calculated, completely eliminating reliance on external GPS signals. Note that this SLAM process is specifically optimized for track characteristics, ensuring robust localization under extreme conditions such as high-speed, high-G cornering, and is fundamentally different from the localization methods of existing ADAS (Advanced Driver Assistance Systems) that rely on clear lane lines.

[0093] In this embodiment, this step is specifically designed to optimize the SLAM algorithm for track features such as guardrails, shoulders, cones, and curve contours. By clustering and segmenting LiDAR point clouds and fusing them with IMU attitude, centimeter-level positioning can still be achieved under extreme conditions such as high-speed cornering (i.e., lateral G-value ≥ 1.2g) and vehicle tilt. No track map needs to be pre-installed. The high-precision map can be built through one lap of "learning driving" on the first entry into the track, completely eliminating the dependence on GPS (Global Positioning System) and high-precision maps.

[0094] In one specific embodiment, Figure 4 This is a flowchart of the track map construction and localization process. As shown in the diagram, the process includes: inputting point cloud data collected by LiDAR and attitude data collected by IMU into a fusion algorithm unit. This unit includes point cloud registration and SLAM algorithms to fuse point cloud data and attitude data, completing distortion correction, stitching registration, and global optimization of multi-frame continuous point clouds to achieve 3D environment reconstruction of the entire track scene. Simultaneously, it calculates the real-time pose of the vehicle in the track's global coordinate system, completely eliminating dependence on external positioning signals. It also includes real-time road surface adhesion coefficient estimation. Based on point cloud reflectivity and texture analysis, it extracts microscopic features of the track surface, such as material, wear level, and wet / dry state, from the laser point cloud data, and calculates the road surface adhesion coefficient of each area of ​​the track in real time, upgrading the static geometric map into a semantic map with dynamic physical properties. This provides dynamic and precise input for subsequent tire adhesion limit constraints and optimal trajectory planning; the fusion algorithm unit outputs a high-precision track map with dynamic adhesion coefficients (different from traditional pure geometric point cloud maps, this map includes precise geometric features such as track boundaries, shoulders, and drivable areas, and is also bound to the real-time road surface adhesion coefficients of all points on the track. It is a structured, high-precision track map with physical semantics, providing dual constraints of spatial boundaries and physical limits for optimal trajectory planning), as well as real-time vehicle positioning coordinates (i.e., outputting the vehicle's centimeter-level absolute positioning result in the global coordinate system of the track, adapting to the positioning needs of high-speed extreme conditions and scenarios without external signals on the track, providing a precise real-time position reference for subsequent vehicle state estimation, trajectory deviation calculation, and driving guidance information generation).

[0095] Step 2: Dynamic Optimal Trajectory Generation Based on Vehicle Dynamics Limits: In the real-time constructed point cloud map, the system first identifies the track boundaries. Note that the objective function of the trajectory generation algorithm in this embodiment is not simply the "shortest path," but rather "maximizing the average speed under tire adhesion limit constraints." The algorithm calculates the maximum available grip of the vehicle at each point in real time and generates an "optimal trajectory line" (also known as the "ideal racing line") that fully utilizes this grip. Furthermore, this trajectory line is not static; the system dynamically fine-tunes the trajectory based on real-time road conditions perceived by the LiDAR, such as wet and slippery areas determined by point cloud reflectivity, to achieve true "dynamic optimization."

[0096] In this embodiment, this step estimates the road surface adhesion coefficient (i.e., μ value) in real time through reflectivity analysis and texture feature extraction of lidar point clouds, and constructs an objective function of "maximizing average speed" using the tire friction circle theory as the core constraint. It dynamically allocates longitudinal (acceleration / braking) and lateral (steering) acceleration to generate an optimal trajectory that adapts to real-time road conditions. When a low-adhesion area (such as a wet road surface or oil stains) is detected, the trajectory is automatically adjusted to move outward to obtain stronger grip or to adjust the power output based on the current trajectory, thus solving the defects of static trajectory in existing technologies.

[0097] Step 3: Immersive Performance Guidance and Real-Time Feedback Based on AR-HUD: The calculated "optimal trajectory line" is immersively projected using AR-HUD. Note that the guidance information in this embodiment goes beyond a single "virtual track." The system calculates and displays multi-dimensional performance indicators in real time, including: ① Dynamic marking of the optimal braking / acceleration point (updated in real time with vehicle speed and road conditions); ② Real-time G-value table (lateral / longitudinal G-values), intuitively displaying the vehicle's current dynamic limits and helping the driver perceive these limits; ③ Precise apex warnings and cornering speed suggestions, such as apex warnings designed to guide the driver to cut corners at the optimal position; ④ Dynamic feedback of the guide line (color: green - minimal deviation, yellow - moderate deviation, red - excessive deviation; width: the larger the deviation, the thicker the line). This multimodal feedback constitutes a "virtual track engineer," rather than a simple "navigator," a feature completely absent in general ADAS systems.

[0098] In one specific embodiment, Figure 5 This is a schematic diagram of path planning and visual guidance. It should be noted that this diagram shows the principle of the panoramic application of the track driving assistance system in a complete circular track scenario. It can fully present the entire process logic of the system from global optimal path planning to real-time visual guidance, and intuitively demonstrate how the system helps the driver follow the professional-grade optimal driving line, balancing lap time performance and driving safety. It is also a concrete panoramic presentation of the core functions of the track driving assistance solution mentioned above.

[0099] It needs to be explained that, Figure 5 The track boundary in the diagram represents the edge of the physically drivable area of ​​the track, defining a rigid spatial constraint for the vehicle's entire journey. All routes planned by the system are strictly confined within the track boundary, fundamentally mitigating the safety risk of vehicles running off the track. The ideal racing curve (green dashed line, i.e., the theoretical optimal path mentioned earlier) is a professional-grade optimal driving line obtained by the system through global optimization based on a high-precision track map, vehicle dynamics parameters, and tire adhesion limit constraints, with the objective of "minimizing single-lap time." It fully covers the straight sections and all curves of the circular track and is the ideal driving route for achieving the fastest lap time. The entry point (the optimal position for the vehicle to begin turning and enter the curve, determining the initial line rhythm), the apex (the point within the curve where the vehicle is closest to the inside of the track, the core position with the smallest turning radius, directly determining the maximum efficiency of cornering), and the exit point (the optimal position for the vehicle to complete the turn, straighten, and exit the curve, determining the acceleration efficiency on the straightaway after exiting the curve) are the three core key nodes for cornering driving, and also the core warning points for the system's operation timing prompts. These points provide advance warnings to the driver, ensuring that they perform braking, steering, and acceleration operations at the optimal time. The virtual projection route (yellow solid line) is the visual embodiment of the system's guidance instructions, projected onto the driver's field of vision by AR-HUD and other head-up display devices, achieving centimeter-level spatial binding with the actual track surface. Note that the image dynamically captures the forward section of the road from the vehicle's current position to the upcoming curve, providing a concrete guide that the driver can directly follow. This completely solves the distraction problem of traditional guidance requiring the driver to look down at the instrument panel, achieving immersive driving guidance where "eyes don't leave the road."

[0100] In one specific embodiment, Figure 6 This diagram illustrates track feature recognition and dynamic adjustment. It should be noted that this diagram represents an optimal trajectory dynamic adjustment and multi-dimensional safe driving guidance system based on real-time track feature recognition in a track scenario. It intuitively demonstrates how the system dynamically adapts to changes in track conditions by recognizing core features such as track surface characteristics and key operational nodes, simultaneously optimizing the driving trajectory and guidance strategy, ultimately achieving the dual goals of "maximum performance release and rigid control of safety boundaries."

[0101] It needs to be explained that, Figure 6The dynamic trajectory optimization of the circular track in the diagram is based on the dynamic iteration of the optimal trajectory according to the identification of road surface adhesion characteristics. It can intuitively present the core logic of the system's real-time identification and dynamic optimization of the optimal trajectory for track road surface adhesion characteristics. Note that the system in the diagram uses laser point cloud to identify the wet and slippery road surface area in the curve (i.e., the low adhesion coefficient track characteristic) in real time, and captures the change in working condition where the road surface grip is significantly reduced in this area. For the racing car curve based on the principle of ideal dry road surface planning, the system combines the real-time identification of low adhesion coefficient to recalibrate the tire adhesion limit constraint, and completes the trajectory optimization within the safe range of the track boundary. That is, the original limit line close to the center of the curve is shifted outward, increasing the cornering radius and reducing the lateral acceleration required for cornering, matching the upper limit of tire grip under low adhesion road surface, which not only avoids the vehicle from exceeding the grip limit and causing slippage and loss of control, but also maximizes the cornering efficiency within the safe boundary. At the same time, the virtual projection route projected to the driver's field of vision is updated, and the optimized trajectory is transformed into a visual guide that the driver can directly execute in real time, realizing a real-time closed loop of "working condition identification, trajectory optimization, and guidance update".

[0102] To explain further, Figure 6 The multi-dimensional driving guidance visualization in the diagram is based on the identification of key track features, providing comprehensive driving guidance and safety boundary control. It intuitively presents the core logic of the system's extraction of key track operation features and the implementation of multi-dimensional driving guidance. Note that the system extracts core track operation features, identifies and marks key braking points before curves, and provides visual warnings through blue blocks. At the same time, the optimal driving trajectory ahead is intuitively presented through a yellow virtual projection route, providing drivers with basic steering and line guidance. In addition, the G-value table in the diagram is a concrete manifestation of the tire friction circle theory. Based on real-time vehicle status data, the system intuitively presents the real-time G-value (the combined acceleration of the vehicle's longitudinal and lateral acceleration) in the table, allowing drivers to perceive at a glance the current tire grip utilization level and the remaining distance from the adhesion limit, accurately controlling the aggressiveness of driving operations, and rigidly locking the safety boundary from the human-computer interaction level to avoid exceeding the tire grip limit.

[0103] Step 4: Personalized Driving Model Evolution Based on Driver-in-the-Loop Data: The system records complete driving data for each lap, including vehicle trajectory, speed, system guidance information, and the driver's actual actions. During non-driving periods, the system can utilize this historical data to offline train and optimize the "ideal racing line" calculation model from Step 2 using machine learning algorithms (such as imitation learning or reinforcement learning). The optimized model parameters can be updated to the vehicle system via OTA (Over-The-Air) technology, enabling the system to continuously evolve and ultimately generate highly personalized optimal paths for specific vehicles, tire configurations, and even specific drivers. Note that this learning process is "driver-in-the-loop." The system not only learns the theoretical optimal solution but also analyzes the specific driver's driving habits (such as cornering style and braking timing) to generate a "personalized optimal trajectory" that conforms to both physical limits and the driver's characteristics, achieving an evolution from a "general coach" to a "personalized coach."

[0104] In this embodiment, this step records the vehicle status, driver operations (steering wheel angle, throttle / brake opening) and guidance deviation data for each lap in real time. The trajectory planning model is optimized offline using machine learning algorithms (improved imitation learning, reinforcement learning). The optimized model is updated to the vehicle system in real time via OTA, gradually adapting to specific vehicles (dynamic parameters), tire configurations, and driver driving styles (such as aggressive cornering and braking habits), achieving an evolution from "general optimal" to "specific optimal".

[0105] In one specific embodiment, Figure 7 This is a diagram illustrating data-driven adaptive optimization. It's important to note that this diagram showcases a driver assistance system (ADAS) designed for a racing scenario. It demonstrates a self-learning, iterative, closed-loop architecture based on actual user driving data to achieve personalized path planning, fully presenting the system's iterative logic from "general standardized optimal guidance" to "personalized customized guidance tailored to each user." Further explanation is needed. Figure 7 The machine learning optimization module compares the system's baseline trajectory with the driver's actual operation across all dimensions, accurately characterizing the driver's driving style (aggressive / conservative), vehicle control ability, operational response rhythm, cornering / exit habits, risk preferences, and other unique features to clarify the driver's capability boundaries. Simultaneously, it uncovers the driver's optimal maneuvers in specific track sections (e.g., the driver's line and speed control strategy on a corner, resulting in a faster lap time than the system's initial plan), inversely incorporating the driver's track driving experience to correct the inherent limitations of the generalized standardized model. Finally, after optimization and training by the machine learning module, a personalized path planning model specific to the driver is generated and pushed back to the vehicle system via OTA (Over-The-Air) updates, completing the iterative upgrade of system functions.

[0106] In one specific embodiment, an adaptive track driving assistance system and assistance scheme are also provided, which aims to provide the driver with visual guidance of the optimal driving path through real-time perception and calculation, so as to assist the driver in achieving faster lap times on the track. Specific implementation methods include: Phase 1, System Initialization and Track Learning: Preparations for the vehicle before driving in a real track environment, including the following steps: Step 1: Hardware System Setup and Integration: Install an automotive-grade 3D LiDAR on the center of the roof of the selected vehicle (such as a performance sedan or race car) using a high-strength shock-absorbing bracket. Simultaneously, install a six-axis inertial measurement unit (IMU) near the vehicle's chassis or center of gravity, and connect the LiDAR and IMU to the vehicle's central computing unit (ECU) via CAN bus or Ethernet.

[0107] Step 2, Sensor Timestamp Synchronization: After the system powers on, sensor synchronization initialization is performed first. The central computing unit uses a precise clock synchronization protocol to unify the timestamps of the LiDAR, IMU, and vehicle CAN bus to a microsecond-level error range, ensuring that each frame of data corresponds to the vehicle state at the same moment during subsequent data fusion.

[0108] Step 3: Static Coordinate System Joint Calibration: On a flat and open field, using a standard-sized rectangular calibration board, collect multiple sets of LiDAR point cloud data and IMU attitude data under different attitudes. Through calculation, determine the precise spatial transformation relationship (rotation matrix R and translation vector T) between the LiDAR coordinate system and the IMU coordinate system, thus completing the extrinsic parameter calibration and providing a foundation for subsequent data fusion.

[0109] Step 4: Vehicle Dynamics Parameter Input: Key dynamic parameters of the vehicle, such as wheelbase, center of gravity height, tire friction coefficient range, and maximum steering angle, are pre-input or read from the central computing unit via the CAN bus. These parameters will be used in the subsequent path planning model to ensure that the generated path conforms to the vehicle's physical limits.

[0110] Step 5: Enter Track Mapping Mode: When entering a new track for the first time, the driver selects "Track Mapping Mode" through the human-machine interface (such as the central control screen). In this mode, the system guides the driver to drive a full lap along the center line of the track at a low and constant speed (such as 40-60 km / h) to collect track environment data.

[0111] Step 6: High-precision point cloud map construction: In the "track mapping mode," the central computing unit fuses LiDAR scan data with IMU attitude and position data in real time, and uses the SLAM algorithm to construct a dense 3D point cloud map of the track. This map accurately includes all static environmental information such as the track surface, shoulders, and guardrails.

[0112] Step 7: Key Track Features Extraction and Vectorization: The central computing unit automatically processes the high-precision point cloud map generated in Step 6. Through clustering, edge detection, and fitting algorithms, it accurately extracts the core features of the track, including: the left and right boundary lines of the track, the center line of the track, the shoulder area, and the starting line position. These features are stored in a vectorized data structure to form a structured track model.

[0113] Step 8: Offline Calculation of the Ideal Racing Line: Based on the structured track model generated in Step 7 and the vehicle dynamics parameters input in Step 4, the central computing unit uses a path optimization algorithm (such as variational optimal control or dynamic programming) to calculate the theoretical "ideal racing line." The objective function of this algorithm is typically to minimize the total time to traverse the entire track while satisfying tire grip constraints. The calculation result is a smooth baseline path that includes position, velocity, and acceleration suggestions.

[0114] It should be noted that in this embodiment, the objective function is defined as "maximizing the average speed under the constraint of tire dynamic adhesion limit," rather than simply geometric adaptation or fixed constraint optimization, directly corresponding to the core requirement of "shortest lap time" in track driving. In a specific embodiment, the process of the ideal racing line is as follows: 1) Input data: The system loads track geometry information (such as track centerline, corner radius, and slope) extracted from high-precision point cloud map and pre-input vehicle dynamic parameters, such as wheelbase, center of gravity height, and tire friction coefficient μ.

[0115] 2) Constructing the optimization objective function: The objective function is "to minimize the total time to traverse the entire track while satisfying the tire adhesion limit," specifically including: A. The objective function can be expressed as: Minimize T = ∫(1 / v)ds; where v is the instantaneous velocity of the vehicle, and s is the arc length along the center line of the track. The physical meaning of this objective is: under the premise of satisfying all constraints, maintain the highest speed at every point on the track as much as possible, thereby achieving the shortest total time.

[0116] B. Constraints: Physical boundaries under tire adhesion limits. To achieve the above objectives, the vehicle's trajectory must strictly adhere to its physical limits. Using the tire friction circle theory as the core constraint for path planning, at any given moment, the total grip force provided by the tire is finite, and its longitudinal acceleration a... x (Acceleration or braking) and lateral acceleration a y The vector sum of the (steering) forces cannot exceed the maximum adhesion limit between the tire and the road surface. The mathematical expression for this constraint is:

[0117] Among them, a xThe longitudinal acceleration of the vehicle (m / s²), a y Here, μ is the lateral acceleration of the vehicle (m / s²), μ is the real-time tire friction coefficient, and g is the gravitational acceleration (approximately 9.8 m / s²).

[0118] 3) Dynamic solution process: a x With a y The trade-offs and allocations are considered. Along the track arc length s, the optimal control input sequence is solved, namely the steering wheel angle δ(s) and throttle / brake opening θ(s), thus obtaining the optimal state sequence (path x(s), y(s), and speed v(s)). The solution logic is as follows: (1) State space definition, including: State variables: [x, y, ψ, v], representing the vehicle's lateral position, longitudinal position, heading angle, and velocity, respectively.

[0119] Control variables: [δ, θ], representing steering wheel angle and accelerator / brake opening, respectively.

[0120] (2) Vehicle dynamics model: The system has a built-in simplified vehicle dynamics model to describe how the state variables change with the control variables, including: v x =v×cos(ψ); v y =v×sin(ψ); ψ'=(v / L)×tan(δ), where L is the wheelbase; a x =(F long / m)-a resist , where F long For longitudinal force, m is mass, and a is... resist This is due to air resistance and other drags that cause deceleration.

[0121] (3) Optimal control solution: The system adopts the optimal control algorithm to solve the above dynamic model under the premise of satisfying the friction circle constraint and the track boundary constraint. That is, the boundary can be substituted into the above algorithm to calculate iteratively. The specific process is as follows: Track discretization: Discretize the track arc length s into N points, such as s0, s1, ..., s N-1 Define a state variable at each discrete point: velocity v. i and longitudinal acceleration Simultaneously, regarding the vehicle's lateral movement, as the vehicle travels along the ideal track line, the lateral acceleration is determined by the curvature κ and the velocity v: a y =v 2 ×κ.

[0122] Friction circle constraint: at each discrete point, and satisfy ; Total time T = Σ(2 / (v) i +v i+1 ))×Δs i (Trapezoidal integral approximation).

[0123] (4) The final solution to the optimal control problem is a complete spatiotemporal trajectory. It not only includes a smooth geometric path (x(s), y(s)) that can unleash the full potential of the vehicle, but more importantly, it assigns optimal speed suggestions v(s), acceleration suggestions ax(s) and ay(s) to each point on the path.

[0124] Phase Two, Real-time Driving Assistance and Guidance, represents a fundamental leap from "retrospective display" to "proactive guidance," constructing an immersive interactive system with "real-time performance optimization" as its core objective. This includes the following steps: Step 9, Real-time High-Precision Positioning: When the vehicle is traveling at high speed on the track, the system enters "Real-time Assist Mode". The central computing unit matches the real-time point cloud scanned by the current LiDAR with the pre-built high-precision point cloud map to obtain the vehicle's centimeter-level high-precision position and attitude (heading angle, pitch angle, etc.) in the map coordinate system.

[0125] Step 10, Vehicle Motion State Fusion Estimation: The central computing unit fuses the high-precision positioning results from Step 9, the angular velocity and acceleration data from the IMU, and the wheel speed information from the vehicle's CAN bus, and collects and outputs a smooth, high-frequency, and high-reliability vehicle motion state estimate, including: real-time position, speed, acceleration, yaw rate, etc.

[0126] Step 11: Dynamic Look-Ahead Path Generation and Spatiotemporal State Matching: Based on the current vehicle state estimated in Step 10, the system dynamically extracts a look-ahead path segment from the "ideal racing line." This process employs a speed adaptive mechanism, i.e., the look-ahead distance L... lookahead =L base +k×v current L base Given the base distance (e.g., 30 meters), k is the velocity coefficient, and v... current The current vehicle speed is used. Higher speeds allow for longer forward distances, providing drivers with ample time to anticipate changes. Simultaneously, the system performs curvature analysis on the forward path segment, automatically increasing the density of sampling points in high-curvature areas such as curves to ensure precise guidance. The target matched by the system is not a static geometric point, but a "target spatiotemporal state packet" containing multi-dimensional information. This packet is obtained from the associated data of the forward path points and includes: target position (x...). ref , yref ), Target heading angle ψ ref Target velocity v ref Target longitudinal acceleration and target lateral acceleration .

[0127] Step 12, Real-time Calculation and Evaluation of Driving Performance Deviation: The system compares the vehicle's current real-time state with the "target spatiotemporal state packet" obtained in Step 11 in multiple dimensions to calculate a series of performance deviations: Lateral position deviation: The vertical distance from the vehicle's center of gravity to the target path reference line.

[0128] Heading angle deviation: The angle difference between the vehicle's current heading angle and the tangent direction of the target path reference point.

[0129] Speed ​​deviation: Δv=v current -v ref This deviation is directly related to lap time; a positive deviation indicates excessive speed, while a negative deviation in the acceleration zone indicates that performance has not been fully utilized.

[0130] Acceleration envelope bias: The system is based on the real-time estimated tire-road friction coefficient μ est Using the friction circle theory, calculate the measured longitudinal acceleration of the current vehicle. With lateral acceleration The synthesized vector is evaluated relative to the current target state. , The position and distance from the boundary within the friction circle. This deviation is used to determine the efficiency and rationality of grip utilization. Simultaneously, based on this deviation, the system assesses in real-time how close the current driving state is to the ideal performance boundary, providing quantitative input for the guidance decision in step 13.

[0131] Step 13: Generation of Personalized Guidance Instructions Based on Performance Optimization: The system no longer uses a single PID controller, but instead generates a comprehensive guidance strategy through a performance-optimized guide. The core of this guide is to satisfy the vehicle's instantaneous dynamic constraints (based on real-time μ). est Given the premise that the vehicle is smoothly and efficiently guided from its current state to a target state sequence, we solve an optimal problem. The generated guidance instructions are a comprehensive instruction set, including: Steering guidance: Suggested steering wheel angle or steering correction trend.

[0132] Speed ​​management recommendations: Based on the speed deviation Δv and acceleration envelope state, semantic recommendations such as "maintain", "gentle braking" or "aggressive acceleration" are given.

[0133] Operation timing prompts: Based on forward-looking information, provide preparatory prompts for upcoming key operation points (such as braking points, apex of the curve, and exit point of the curve).

[0134] Step 14: Visual Guidance Information Rendering: The central computing unit will visualize the generated guidance information through the HUD or dashboard. Specifically, a semi-transparent, glowing virtual path line (such as an AR augmented reality effect) will be projected onto the real road in front of the driver. When the vehicle deviates from the path, the virtual line will shift relative to the real road, visually indicating the correct direction.

[0135] Step 15, Dynamic Color and Width Feedback: The color and width of the virtual guide lines can dynamically change according to the magnitude of driving deviation. For example, when the deviation is within a very small range, the line is green; as the deviation increases, it turns yellow; and when the deviation is too large, it turns red. At the same time, the line width can also be thickened as the deviation increases, providing a stronger visual warning.

[0136] Step 16, Key Curve Information Prompt: When the vehicle approaches a key braking point or apex on the track, the system can dynamically display the suggested braking force or optimal cornering speed in the form of icons or numbers at a specific location on the HUD, providing the driver with quantitative reference information.

[0137] Step 17: Real-time Lap Time and Performance Data Recording: The system utilizes high-precision positioning data to accurately calculate the moment the vehicle crosses the starting line, thereby displaying the current lap time, the historical best lap time, and the real-time difference between the best lap time and the current lap time. Simultaneously, it records the vehicle's speed, G-force, and other performance data for each lap.

[0138] Phase Three, Data Replay and Adaptive Optimization: This phase utilizes personalized "driver-in-the-loop" modeling and evolution as a core architecture of the track driving assistance system. It resolves the inherent contradiction between general assistance systems and individual driver differences. Through a data-driven approach, the system can understand, adapt to, and ultimately achieve efficient collaboration with specific drivers, thus realizing a leap from "one-size-fits-all" to "personalized instruction." This includes the following steps: Step 18: Complete Recording of Driving Data: The system packages and stores complete data for each lap, including but not limited to: high-precision trajectory, vehicle status, driver operations (steering wheel angle, accelerator and brake opening), system-generated guidance path, and real-time deviation data. Each data package includes the lap number and a timestamp.

[0139] Step 19, Offline Data Analysis and Path Optimization: After the track day, the data recorded in Step 18 can be replayed and analyzed using dedicated software or a cloud platform. By comparing the driver's actual racing line with the system's planned "ideal racing line," the system analyzes which sections the driver was faster or slower. Using this data, machine learning algorithms (such as imitation learning or reinforcement learning) are employed to fine-tune the path planning model from Step 8, making the generated path more closely resemble the vehicle's actual optimal solution on the track.

[0140] Step 20, Model Update and Personalized Push: The optimized path planning model parameters from Step 19 are updated back to the vehicle system via OTA technology. Through multiple learning and iterations, the system can generate highly personalized optimal racing lines for specific vehicles, tire configurations, and even specific drivers, achieving continuous evolution of the assistance effect.

[0141] This embodiment also provides a driver assistance system for track scenarios, such as... Figure 8 As shown, it includes: Module 801 is used to build a track map using environmental and vehicle data collected while the vehicle is driving on the track.

[0142] The determination module 802 is used to determine the theoretically optimal path based on the track map, preset vehicle parameters, and tire adhesion limit constraints. The theoretically optimal path includes the suggested speed and suggested acceleration for each path.

[0143] The generation module 803 is used to acquire the current state data of the vehicle and generate driving guidance information based on the current state data and the theoretical optimal path.

[0144] The auxiliary module 804 is used to project driving guidance information into the driver's field of vision to assist track driving.

[0145] The track-oriented assisted driving system provided in this embodiment of the invention can execute the track-oriented assisted driving method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method execution. Further functional descriptions of the above modules are the same as in the corresponding embodiments described above, and will not be repeated here.

[0146] The track-oriented assisted driving system in this embodiment of the invention utilizes a track map constructed from environmental and vehicle data collected during track driving, along with tire adhesion limit constraints, to generate a theoretically optimal path with suggested speeds and accelerations at all points. This maximizes vehicle performance and helps the driver achieve theoretically optimal lap times. Furthermore, based on real-time comparison of the vehicle's current state data with the theoretically optimal path, it dynamically generates driving guidance information and projects it into the driver's field of vision. This not only identifies issues such as driver deviations in line, speed, and acceleration / deceleration timing in real time, providing corrective guidance in advance to avoid the risk of loss of control due to accumulated deviations, but also significantly reduces safety hazards caused by distraction and avoids lap time losses due to inattention. This achieves a comprehensive improvement in the performance, safety, and interactive experience of track-oriented assisted driving.

[0147] Figure 9 This is a schematic diagram of the structure of a vehicle provided in an embodiment of the present invention.

[0148] The following is a detailed reference. Figure 9 The diagram illustrates a structural schematic suitable for implementing a vehicle according to an embodiment of the present invention. The vehicle may include a processor (e.g., a central processing unit, graphics processor, etc.) 901, which can perform various appropriate actions and processes according to a program stored in a read-only memory ROM 902 or a program loaded from memory 908 into a random access memory RAM 903. The RAM 903 also stores various programs and data required for vehicle operation. The processor 901, ROM 902, and RAM 903 are interconnected via a bus 904. An input / output (I / O) interface 905 is also connected to the bus 904.

[0149] Typically, the following devices can be connected to I / O interface 905: input devices 906 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 907 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 908 including, for example, magnetic tapes, hard disks, etc.; and communication devices 909. Communication device 909 allows the vehicle to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 9 Vehicles with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown, and more or fewer devices may be implemented or have instead.

[0150] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 909, or installed from a memory 908, or installed from a ROM 902. When the computer program is executed by the processor 901, it performs the functions defined in the assisted driving method for track scenarios according to embodiments of the present invention. Figure 9 The vehicle shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of the present invention.

[0151] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the assisted driving method for track scenarios shown in the above embodiments is implemented.

[0152] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A driving assistance method for track scenarios, characterized in that, The method includes: A track map is constructed using environmental and vehicle data collected while the vehicle is driving on the track. Based on the track map, preset vehicle parameters, and tire adhesion limit constraints, a theoretically optimal path is determined, which includes a suggested speed and a suggested acceleration for each path. The current state data of the vehicle is obtained, and driving guidance information is generated based on the current state data and the theoretical optimal path. The driving guidance information is projected into the driver's field of vision to assist driving on the track.

2. The assisted driving method for track scenarios according to claim 1, characterized in that, The environmental data includes point cloud data, and the vehicle data includes pose data; the process of constructing a track map using the environmental data and vehicle data collected while the vehicle is driving on the track includes: Point cloud data and pose data are fused to generate a point cloud map; Track features are identified and extracted from the point cloud map, including at least one of the following: track surface, track boundary, center line, shoulder, and guardrail. The track features are marked in the point cloud map to obtain the track map.

3. The assisted driving method for track scenarios according to claim 1, characterized in that, The preset vehicle parameters include at least the road surface adhesion coefficient; the tire adhesion limit constraint is constructed based on the tire friction circle theory; the determination of the theoretically optimal path based on the track map, preset vehicle parameters, and tire adhesion limit constraint includes: Establish an optimization function with the objective of minimizing the single-lap travel time; Based on the road surface adhesion coefficient, the maximum adhesion limit is determined; the vector sum of the longitudinal acceleration and lateral acceleration of the vehicle is calculated, and the maximum adhesion limit is taken as the tire adhesion limit constraint. The track boundaries are extracted from the track map, and the optimization function is solved under the condition of satisfying the tire adhesion limit constraint to obtain the theoretical optimal path containing the path point sequence and the suggested speed and suggested acceleration for each path.

4. The assisted driving method for track scenarios according to claim 1, characterized in that, The step of obtaining the current status data of the vehicle includes: The collected point cloud data is matched with the track map to determine the vehicle's location on the track map. Based on the inertial measurement unit data and vehicle bus data collected by the vehicle, as well as the positioning results, the vehicle motion state is estimated to obtain the current state data of the vehicle. The current state data includes at least position, velocity, acceleration, and yaw rate.

5. The assisted driving method for track scenarios according to claim 4, characterized in that, The process of generating driving guidance information based on the current state data and the theoretically optimal path includes: Based on the speed in the current state data, a target guidance path is dynamically extracted from the theoretically optimal path, and the length of the target guidance path is positively correlated with the speed. The target spatiotemporal state set of the target guidance path is determined, and the target spatiotemporal state set includes the target position, target heading angle, target velocity, target longitudinal acceleration, and target lateral acceleration; Based on the current state data and the target spatiotemporal state set, a performance deviation set is calculated. The performance deviation set includes lateral position deviation, heading angle deviation, velocity deviation, and acceleration envelope deviation. The acceleration envelope deviation is determined based on the current road surface adhesion coefficient and the tire friction circle theory. The current road surface adhesion coefficient is estimated using the currently collected point cloud data. Driving guidance information is generated based on the performance deviation set. The driving guidance information includes at least steering guidance information, speed management suggestions, and operation timing prompts. The steering guidance information is generated based on lateral position deviation and / or heading angle deviation to provide suggested steering wheel angles or steering correction trends. The speed management suggestions are determined based on speed deviation and acceleration envelope deviation to provide suggested speed and suggested acceleration. The operation timing prompts are generated based on relevant information of key operation points in the target guidance path to provide prompts for future key operation points. The key operation points include at least braking points, apex bends, entry points, and exit points.

6. The assisted driving method for track scenarios according to claim 1, characterized in that, The step of projecting the driving guidance information into the driver's field of vision to assist track driving includes: The driving guidance information is visualized and rendered using an in-vehicle display device to obtain a virtual projected route; Based on the degree of deviation between the actual driving trajectory of the vehicle and the theoretical optimal path, the visual attributes of the virtual projection route are dynamically adjusted, and the visual attributes include at least one of color, brightness, or width.

7. The assisted driving method for track scenarios according to any one of claims 1 to 6, characterized in that, After projecting the driving guidance information into the driver's field of vision to assist track driving, the method further includes: Record historical data of the vehicle during its driving on the track. The historical data includes the actual driving trajectory, vehicle status data, driver operation data, and the theoretical optimal path. Based on the historical data, the generation process of the theoretically optimal path is iteratively optimized.

8. A driver assistance system for racing scenarios, characterized in that, The system includes: The module is used to build a track map using environmental and vehicle data collected while the vehicle is driving on the track. The determination module is used to determine the theoretically optimal path based on the track map, preset vehicle parameters, and tire adhesion limit constraints. The theoretically optimal path includes the suggested speed and suggested acceleration for each path. The generation module is used to acquire the current state data of the vehicle and generate driving guidance information based on the current state data and the theoretical optimal path; The auxiliary module is used to project the driving guidance information into the driver's field of vision to assist driving on the track.

9. A vehicle, characterized in that, The vehicle includes a controller, which includes a memory and a processor. The memory and the processor are communicatively connected to each other. The memory stores computer instructions, and the processor executes the computer instructions to perform the assisted driving method for track scenarios as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the assisted driving method for a track scenario as described in any one of claims 1 to 7.

Images