Mountain road suitability improvement effect evaluation method and system for automatic driving truck

Abstract

This invention relates to a method and system for evaluating the drivability improvement effect of autonomous trucks on mountain roads. The method includes: collecting mountain road information ahead of the autonomous truck through drone aerial photography; constructing a mountain road visibility improvement information database covering geometric alignment features and the length of the forward visible path through 3D point cloud reconstruction and road vision domain calculation; acquiring vehicle operating status and forward environmental information using onboard sensors to generate and identify the available line-of-sight distance (LAS); calculating the maximum speed limit under drone-assisted and autonomous vehicle conditions based on the database and the LAS, according to the LAS safety threshold; and calculating the drivability margin improvement effect of the autonomous truck on mountain roads using an adaptive hierarchical evaluation method. This invention can quantify the drivability improvement effect of autonomous trucks on mountain roads under drone assistance, providing support for autonomous driving technology on complex road alignments and road sections with limited line-of-sight.

Description

Technical Field

[0001] This invention relates to the field of road safety assessment technology, and in particular to a method and system for evaluating the improvement of drivability of autonomous trucks on mountain roads. Background Technology

[0002] With the rapid development of autonomous driving technology, L2-L4 level autonomous driving systems have been gradually applied to trunk highway transportation scenarios. Especially in long-distance trunk logistics transportation, autonomous trucks have important application prospects for reducing driver fatigue, improving transportation efficiency and ensuring operational safety.

[0003] However, compared to plains and highways, mountain roads are characterized by continuous sharp curves, frequent changes in longitudinal slope, limited visibility, and complex obstruction relationships, significantly impacting vehicle safety. Particularly under conditions of continuous curves and terrain obstruction, the forward perception range of autonomous driving systems is easily limited, constraining the upper limit of safe speed and increasing operational risks.

[0004] Existing research mainly focuses on improving the safety of autonomous driving from the perspective of optimizing vehicle control strategies or enhancing onboard perception capabilities. However, these methods primarily rely on onboard cameras, lidar, or millimeter-wave radar for identifying the environment ahead. Their perception distance is limited by terrain occlusion and curve curvature, making it difficult to overcome the "vehicle sight distance bottleneck" in mountainous road environments.

[0005] In recent years, vehicle-mounted drone collaborative technology has gradually developed. By using drones for forward-looking aerial photography, road geometry and obstacle information at greater distances ahead of vehicles can be obtained, providing a new technical path to overcome the limitations of traditional vehicle-mounted perception line-of-sight. However, there is currently a lack of a systematic method to quantify the improvement in drivability of autonomous trucks under drone-assisted conditions in mountainous road environments, especially a safety speed comparison and evaluation mechanism based on road 3D models and safe line-of-sight constraints. Summary of the Invention

[0006] In view of this, the purpose of this invention is to provide a method and system for evaluating the improvement of drivability on mountain roads for autonomous trucks. This method, based on UAV aerial 3D reconstruction and view domain calculation combined with vehicle-mounted multi-sensor fusion perception, constructs a dynamic view domain information database for mountain roads. It then inverts the available line-of-sight under both UAV-assisted and vehicle-perception conditions, calculates safe speed limits for both conditions based on line-of-sight safety constraints, and establishes a safe speed difference index to quantitatively evaluate the improvement in drivability under UAV-assisted conditions. This provides technical support for the operational safety analysis and technical optimization of autonomous driving systems on mountain roads.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: a method for evaluating the improvement of drivability on mountain roads for autonomous trucks, comprising the following steps:

[0008] Step S1: Collect information on the mountain road ahead of the autonomous truck through drone aerial photography, and construct a mountain road visibility enhancement information database covering geometric line features and the length of the forward visibility path through 3D point cloud reconstruction and road visibility calculation.

[0009] Step S2: Use on-board sensors to acquire vehicle operating status and front environment information, and invert the vehicle's available line of sight based on the virtual obstacle generation and recognition results of multi-sensor fusion.

[0010] Step S3: Based on the database and the available line-of-sight distance of the vehicle, calculate the maximum speed limit for the UAV-assisted operating conditions according to the line-of-sight safety threshold. V max Maximum speed limit under vehicle operating conditions ;

[0011] Step S4: Maximum speed limit of the UAV under assisted operating conditions obtained in step S3. V max Maximum speed limit for self-driving vehicles The driving ability margin improvement index was calculated, and an adaptive grading evaluation method was used to calculate the driving ability margin improvement effect of autonomous trucks on mountain roads.

[0012] In a preferred embodiment, step S1 specifically includes:

[0013] Step S11: Using a vehicle-mounted multi-rotor drone equipped with lidar and high-definition vision sensors, aerial photography is conducted along the driving direction of the autonomous truck to collect road spatial point cloud data and image data within the range of the preset route markers.

[0014] Aerial photography parameters include: flight altitude controlled between 50-200m, flight speed not exceeding 15m / s, and flight path overlap rate ≥80%;

[0015] Step S12: Preprocess the collected point cloud data and image data, including point cloud denoising, image stitching, time synchronization and coordinate unification processing, and reconstruct the three-dimensional point cloud based on the UAV pose parameters. At the same time, perform spatiotemporal alignment and fusion with the vehicle multi-sensor data to construct a three-dimensional spatial model of the road ahead. Then, perform road vision field calculation and occlusion relationship analysis in the three-dimensional spatial model to obtain the length of the visible path ahead, Su.

[0016] Step S13: Based on the vehicle's current speed Assisted line-of-sight path length for drones Perform time-based corrections to obtain the corrected effective usable line of sight. Its expression is:

[0017] (1)

[0018] τu represents the data transmission delay between the UAV and the vehicle-mounted system.

[0019] Step S14: The road alignment feature parameters extracted from the three-dimensional spatial model and the dynamically updated effective available sight distance are stored in a structured manner to construct a dynamic visibility enhancement information database Ω for mountain roads.

[0020] The information database for improving visibility on mountain roads includes: the length of the visible path ahead under drone-assisted conditions. The range of road longitudinal position station numbers corresponding to the visible path, and the road longitudinal slope i and road surface longitudinal friction coefficient f used for sight distance safety calculation. L .

[0021] In a preferred embodiment, step S2 specifically includes:

[0022] Step S21: Use the multi-source onboard perception sensors mounted on the autonomous truck to obtain vehicle operating status information and road environment information ahead;

[0023] The operational status information includes: during the time interval from when the Level 3 autonomous driving system issues a takeover request to when the driver performs the takeover action, the vehicle uses a preset deceleration 'a'. dp Its typical value is 2.5 m·s -2 Driver perception-braking reaction time t pb_h ; Autonomous driving system perception-braking reaction time t pb_s ; Autonomous driving system perception and reaction time t p_s Driver takeover reaction time t T ;

[0024] The vehicle-mounted perception sensor includes one or more of vehicle-mounted cameras and lidar;

[0025] Step S22: Based on the vehicle's current position and the road alignment information ahead, construct a set of virtual obstacles along the vehicle's travel reference path within a preset distance range;

[0026] Obstacles include one or more of the following: vehicles ahead of the road, roadside fixed facilities, and terrain obstructions;

[0027] The set of virtual obstacles is sequentially distributed at certain intervals along the road travel direction;

[0028] Step S23: Match and analyze the virtual obstacle set with the environmental perception results after multi-sensor fusion to determine the identifiable status of each virtual obstacle and determine the position of the farthest virtual obstacle that can be stably identified under the current vehicle perception conditions.

[0029] Step S24: Calculate the distance along the road travel direction between the location of the farthest virtual obstacle and the current location of the vehicle, as the available line-of-sight S of the autonomous truck under the vehicle's perception conditions. a .

[0030] In a preferred embodiment, step S3 is implemented as follows:

[0031] Step S31: Calculate the maximum speed limit V of the UAV in the assisted operating condition under the condition of meeting line-of-sight safety. max The formula is:

[0032] (2)

[0033] Equation (2) represents Level 1 and Level 2 autonomous driving systems with UAV-assisted operating conditions;

[0034] (3)

[0035] Equation (3) is a Level 3 autonomous driving system with unmanned aerial vehicle (UAV) assistance.

[0036] (4)

[0037] Equation (4) is a Level 4 autonomous driving system with unmanned aerial vehicle (UAV) assistance.

[0038] Step S32: Calculate the maximum speed limit for the vehicle under the condition of meeting line-of-sight safety requirements. The formula is:

[0039] (5)

[0040] Equation (5) represents Level 1 and Level 2 automated driving systems under autonomous vehicle operating conditions;

[0041] (6)

[0042] Equation (6) represents a Level 3 autonomous driving system under autonomous vehicle operating conditions;

[0043] (7)

[0044] Equation (7) is the Level 4 autonomous driving system for the vehicle.

[0045] In a preferred embodiment, step S4 is implemented as follows:

[0046] Step S41: Based on the difference in safe speed limits between UAV-assisted and autonomous vehicle-perception conditions, construct a drivability margin improvement index ΔV, the calculation formula of which is:

[0047] (8)

[0048] Step S42: Construct a sample dataset containing the drivability margin improvement index ΔV calculated under multiple mountain road sections or multiple operating conditions. ;

[0049] Step S43: Standardize the sample dataset and use an unsupervised clustering algorithm for adaptive clustering analysis to automatically determine the number of levels of driving suitability improvement effect;

[0050] Step S44: Determine the threshold range corresponding to each level of driving suitability improvement effect based on the clustering results, and form the hierarchical boundary between different levels;

[0051] Step S45: Establish a quantitative correspondence between each level of drivability improvement effect and the drivability margin improvement index ΔV, and conduct a graded evaluation of the drivability improvement effect of mountain roads based on the interval of the drivability margin improvement index ΔV.

[0052] In a preferred embodiment, when the pilotability margin improvement index ΔV is in the high-level range determined by the clustering algorithm, it is determined that the pilotability improvement effect is significant, indicating that the maximum safe speed under UAV assistance conditions is greatly improved and the operational safety margin is significantly enhanced.

[0053] When the pilotability margin improvement index ΔV is in the medium level range, it is judged that the pilotability improvement effect is average, indicating that the drone assistance has a certain effect on improving the upper limit of safe speed.

[0054] When the drivability margin improvement index ΔV is in the low-level range, it is judged that the drivability improvement effect is weak or has not produced an effective improvement, indicating that the drone assistance has failed to significantly improve the forward visibility conditions of the vehicle.

[0055] In a preferred embodiment, the unsupervised clustering algorithm is one or more of the following: adaptive K-means clustering algorithm, adaptive hierarchical clustering algorithm, or adaptive density clustering algorithm.

[0056] This invention also provides an evaluation system for improving the drivability of autonomous trucks on mountain roads, including a processor, a memory, and a bus. The memory stores machine-readable instructions executed by the processor. When the system is running, the processor communicates with the memory via the bus, and when the machine-readable instructions are executed by the processor, the method for evaluating the improvement of drivability of autonomous trucks on mountain roads is as described above.

[0057] Compared with the prior art, the present invention has the following beneficial effects:

[0058] (1) By using UAV aerial photography and 3D point cloud reconstruction technology, a database of information on the improved visibility of mountain roads is constructed, which breaks through the bottleneck of the traditional vehicle perception that is limited by terrain obstruction. It enables accurate calculation of the length of the visible path ahead of curves and obstructed road sections, improves the completeness and foresight of road environment information acquisition, and provides a data basis for evaluating the driving suitability improvement effect of autonomous vehicles under UAV assistance.

[0059] (2) By constructing a set of virtual obstacles and combining the results of multi-sensor fusion recognition to invert the available line of sight of the vehicle, the calculation of the vehicle line of sight is transformed from a single perception distance judgment to an identifiable stable boundary analysis, which improves the objectivity and quantifiability of the calculation of the available line of sight under the vehicle's working conditions.

[0060] (3) Based on the line-of-sight safety critical condition, the maximum safe speed limit under the UAV-assisted working condition and the self-driving working condition were calculated respectively. The comparability analysis of the upper limit of safe speed under different perception conditions was realized, and a safe speed comparison analysis mechanism based on line-of-sight safety constraints was established, providing a basis for the quantitative evaluation of the improvement effect of driving suitability of mountain roads under UAV-assisted conditions.

[0061] (4) By constructing the drivability margin improvement index ΔV, the difference in safe speed limits under UAV-assisted working conditions and autonomous vehicle perception working conditions is quantitatively analyzed, so as to realize the quantitative and graded evaluation of the drivability improvement effect of UAV-assisted conditions on mountain roads, and transform the drivability improvement effect from traditional qualitative judgment to quantifiable index evaluation, providing data basis for the optimization of autonomous driving system parameters and the formulation of operation strategies. Attached Figure Description

[0062] Figure 1 This is a flowchart of the method for evaluating the improvement of drivability of autonomous trucks on mountain roads provided in an embodiment of the present invention;

[0063] Figure 2 This is a flowchart of an embodiment of the present invention for constructing a mountain road visibility enhancement information database Ω that covers geometric linear features and the length of the forward visible path;

[0064] Figure 3 This is a flowchart illustrating the process of retrieving the vehicle's available line of sight Sa based on the virtual obstacle generation and recognition results using multi-sensor fusion, according to an embodiment of the present invention.

[0065] Figure 4 This is a flowchart illustrating the calculation of the maximum speed limit under both UAV-assisted and autonomous vehicle operating conditions, according to an embodiment of the present invention.

[0066] Figure 5This is a flowchart illustrating the improvement effect of the adaptive hierarchical evaluation method on the drivability margin of autonomous trucks on mountain roads in an embodiment of the present invention. Detailed Implementation

[0067] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0068] It should be noted that the following detailed descriptions are illustrative and intended to provide further explanation of this application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0069] It should be noted that the terminology used herein is for the purpose of describing particular implementations only and is not intended to limit the exemplary implementations according to this application; as used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise; furthermore, it should be understood that when the terms “comprising” and / or “including” are used in this specification, they indicate the presence of features, steps, operations, devices, components and / or combinations thereof.

[0070] like Figure 1-5 As shown, the method for evaluating the improvement of drivability of autonomous trucks on mountain roads proposed in this invention includes the following steps:

[0071] Step 1: Collect information about the mountainous road ahead of the autonomous truck using drone aerial photography. After 3D point cloud reconstruction and road view calculation, construct a database of enhanced visibility information for the mountainous road, encompassing geometric features and the length of the forward visible path. The flowchart for this step is shown below. Figure 2 As shown;

[0072] Step 11: Using a vehicle-mounted multi-rotor drone equipped with LiDAR and high-definition vision sensors, aerial photography is conducted along the driving direction of the autonomous truck to collect road spatial point cloud data and image data within the range of the preset route markers.

[0073] The aerial photography parameters include at least: flight altitude controlled between 50-200m, flight speed not exceeding 15m / s, and flight path overlap rate ≥80%;

[0074] Step 12: Preprocess the collected point cloud data and image data, including point cloud denoising, image stitching, time synchronization, and coordinate unification. Reconstruct the 3D point cloud based on the UAV pose parameters, and simultaneously perform spatiotemporal alignment and fusion with vehicle-mounted multi-sensor data to construct a 3D spatial model of the road ahead. Then, perform road view domain calculation and occlusion relationship analysis within the 3D spatial model to obtain the length S of the visible path ahead. u ;

[0075] The construction of the three-dimensional spatial model of the road ahead includes: using a statistical filtering algorithm to remove outliers in the point cloud, using an improved ICP algorithm to complete the registration of the flight strip point cloud, using the CSF algorithm to segment the ground point cloud and non-ground point cloud, generating the road surface DEM and terrain occlusion model, combining image semantic segmentation to extract the road centerline, and fusing to obtain a three-dimensional spatial model with semantic information.

[0076] The road view calculation and occlusion analysis include: using the center of the vehicle's front bumper as the view reference point, constructing a view cone consistent with the parameters of the onboard sensors; using ray tracing to emit view rays to each point on the road centerline; determining the intersection status of the view rays with terrain occlusion bodies; and obtaining the length S of the forward visible path based on the cumulative length of the continuously visible centerline points. u .

[0077] Step 13: Based on the vehicle's current speed V u The length of the line-of-sight path S assisted by the drone u By performing time-based corrections, the corrected effective usable line-of-sight S is obtained. u ', its expression is:

[0078] (1)

[0079] The τ u For data transmission latency between the drone and the vehicle system;

[0080] Step 14: The road alignment feature parameters extracted from the three-dimensional spatial model and the dynamically updated effective available sight distance are stored in a structured manner to construct a dynamic visibility enhancement information database Ω for mountain roads;

[0081] The information database for improving visibility on mountain roads includes: the length of the forward visible path S under drone-assisted conditions. u The range of road longitudinal position station numbers corresponding to the visible path, and the road longitudinal slope i and road surface longitudinal friction coefficient f used for sight distance safety calculation. L .

[0082] Step 2: Utilize onboard sensors to acquire vehicle operating status and forward environment information, and invert the vehicle's available line of sight based on the virtual obstacle generation and recognition results from multi-sensor fusion; the flowchart for this step is as follows. Figure 3 As shown;

[0083] Step 21: Use the multi-source onboard perception sensors on the autonomous truck to obtain information on the vehicle's operating status and the road environment ahead;

[0084] The operational status information includes at least the following: for a Level 3 autonomous driving system, during the time interval from issuing a takeover request to the driver performing the takeover action, the vehicle uses a preset deceleration 'a'.dp Its typical value is 2.5 m·s -2 Driver perception-braking reaction time t pb_h ; Autonomous driving system perception-braking reaction time t pb_s ; Autonomous driving system perception and reaction time t p_s Driver takeover reaction time t T ;

[0085] The vehicle-mounted perception sensor includes at least one or more of the following: vehicle-mounted camera and lidar.

[0086] Step 22: Based on the vehicle's current position and the road alignment information ahead, construct a set of virtual obstacles along the vehicle's reference path within a preset distance range;

[0087] The obstacles include at least vehicles in front of the road, roadside fixed facilities, or terrain obstructions;

[0088] The set of virtual obstacles is sequentially distributed at certain intervals along the road travel direction;

[0089] Step 23: Match and analyze the virtual obstacle set with the environmental perception results after multi-sensor fusion to determine the identifiable status of each virtual obstacle and determine the position of the farthest virtual obstacle that can be stably identified under the current vehicle perception conditions.

[0090] Among them, the multi-sensor fusion perception result is structured data that includes the target's spatial location, size, detection confidence, and number of tracking frames;

[0091] Among them, the matching analysis adopts the spatial distance matching criterion, and the matching is considered successful when the planar distance between the virtual obstacle and the perceived target is less than or equal to 1.5m;

[0092] Among them, the identifiable state is divided into three categories: stable identifiable, unstable identifiable, and unidentifiable. The criteria for stable identifiable state are a detection confidence level ≥ 0.7 and a continuous tracking frame count ≥ 3 frames.

[0093] Among them, the farthest stable identifiable virtual obstacle position is the road centerline position corresponding to the farthest stable identifiable virtual obstacle along the driving direction;

[0094] The location of the road centerline corresponding to the farthest virtual obstacle refers to the location of the centerline point obtained by vertically projecting the bottom center coordinates of the virtual obstacle onto the road centerline; if the projection point is outside the preset aerial photography range or not within the vehicle's driving path, the virtual obstacle will not participate in the line-of-sight calculation.

[0095] Step 24: Calculate the distance along the road travel direction between the location of the farthest virtual obstacle and the current location of the vehicle, which is taken as the available line-of-sight S of the autonomous truck under the vehicle's perception conditions. a .

[0096] Using the same vehicle front bumper center as in step S13 as the calculation reference point, the bottom center coordinates of the farthest stable and recognizable virtual obstacle are taken.

[0097] The coordinate transformation matrix generated in step S12 is used to unify the coordinates of the virtual obstacles to the global coordinate system of the road 3D spatial model;

[0098] The coordinates of the virtual obstacle are vertically projected onto the road centerline extracted in step S12, and the coordinates of the projection point are calculated.

[0099] By summing the three-dimensional distances from the centerline point corresponding to the current position of the vehicle to the adjacent centerline points of the projection point, the available line of sight S of the vehicle can be obtained. a ;

[0100] For virtual obstacles in an unstable, identifiable state, the calculation result is multiplied by a safety correction factor of 0.8.

[0101] Step 3: Based on the database and the available line-of-sight distance of the vehicle, and according to the line-of-sight safety threshold, calculate the maximum speed limit under both UAV-assisted and vehicle-assisted operating conditions; the flowchart for this step is as follows. Figure 4 As shown;

[0102] Step 31: Calculate the maximum speed limit V of the UAV in the assisted operating condition under the condition of meeting line-of-sight safety. max The formula is:

[0103] (2)

[0104] Equation (2) represents Level 1 and Level 2 autopilot systems with UAV-assisted operating conditions; S u 'This is the length of the enhanced visible path after reconstruction of the 3D field of view based on UAV forward-looking aerial photography and time-corrected.'

[0105] (3)

[0106] Equation (3) is a Level 3 autonomous driving system with unmanned aerial vehicle (UAV) assistance.

[0107] (4)

[0108] Equation (4) is a Level 4 autonomous driving system with unmanned aerial vehicle (UAV) assistance.

[0109] Step 32: Calculate the maximum speed limit for the vehicle under the condition of meeting line-of-sight safety requirements. The formula is:

[0110] (5)

[0111] Equation (5) represents Level 1 and Level 2 automated driving systems under autonomous vehicle operating conditions; S a The usable line of sight is obtained by relying solely on the multi-source vehicle-mounted perception sensors on the vehicle itself under occlusion conditions in complex mountainous terrain.

[0112] (6)

[0113] Equation (6) represents a Level 3 autonomous driving system under autonomous vehicle operating conditions;

[0114] (7)

[0115] Equation (7) is the Level 4 autonomous driving system for the vehicle.

[0116] Step 4: Calculate the improvement effect of autonomous truck drivability margin on mountain roads using an adaptive hierarchical evaluation method; the flowchart for this step is shown below. Figure 5 As shown;

[0117] Step 41: Based on the difference in safe speed limits between UAV-assisted and autonomous vehicle-perception-based operating conditions, construct a drivability margin improvement index ΔV, the calculation formula of which is:

[0118] (8)

[0119] Step 42: Construct a sample dataset containing the drivability margin improvement index ΔV calculated under multiple mountain road sections or multiple operating conditions. ;

[0120] Step 43: Standardize the sample dataset and use an unsupervised clustering algorithm for adaptive clustering analysis to automatically determine the number of levels of driving suitability improvement.

[0121] Step 44: Determine the threshold range corresponding to each level of drivability improvement effect based on the clustering results, and form the hierarchical boundary between different levels;

[0122] Step 45: Establish a quantitative correspondence between each level of drivability improvement and the drivability margin improvement index ΔV, and evaluate the drivability improvement effect of mountain roads according to the interval of ΔV:

[0123] When ΔV is in the high-level interval determined by the clustering algorithm, it is judged that the pilotability improvement effect is significant, indicating that the maximum safe speed under UAV assistance conditions is greatly improved and the operational safety margin is significantly enhanced.

[0124] When ΔV is in the medium level range, it is judged that the improvement effect of the pilotability is average, indicating that the drone assistance has a certain effect on improving the upper limit of safe speed.

[0125] When ΔV is in the low-level range, it is judged that the improvement effect on drivability is weak or no effective improvement is produced, indicating that the drone assistance has failed to significantly improve the forward visibility conditions of the vehicle.

[0126] The unsupervised clustering algorithm is one or more of the following: adaptive K-means clustering algorithm, adaptive hierarchical clustering algorithm, or adaptive density clustering algorithm.

[0127] In summary, this invention provides a method for evaluating the drivability improvement effect of autonomous trucks on mountain roads. This method acquires 3D data of mountain roads through UAV aerial photography, constructs a dynamic road visibility information database, and combines onboard multi-sensor fusion perception to invert the available line-of-sight distance of the vehicle. Under line-of-sight safety constraints, it calculates the maximum safe speed limit under both UAV-assisted and vehicle-perceived conditions. By constructing a drivability margin improvement index ΔV, it quantitatively evaluates and classifies the drivability improvement effect of autonomous trucks under UAV assistance in mountain road environments. This invention can identify key road sections and influencing factors that limit the safe operation of autonomous trucks, providing a basis for the formulation and effectiveness verification of existing drivability improvement measures for mountain roads. Compared with existing evaluation methods that rely on design parameters or single simulation analysis, this invention performs calculations and analysis based on actual road 3D environmental data, resulting in evaluation results that better reflect actual operating conditions, a wider range of applications, stronger engineering operability, and greater potential for widespread application.

[0128] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0129] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0130] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0131] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0132] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

[0133] This patent is not limited to the above-described preferred embodiments. Anyone can derive other various methods for evaluating the driving suitability of autonomous trucks on mountain roads based on the inspiration of this patent. All equivalent changes and modifications made within the scope of this patent application shall be covered by this patent.

Claims

1. A method for evaluating the improvement of drivability on mountain roads for autonomous trucks, characterized in that, Includes the following steps: Step S1: Collect information on the mountain road ahead of the autonomous truck through drone aerial photography, and construct a mountain road visibility enhancement information database covering geometric line features and the length of the forward visibility path through 3D point cloud reconstruction and road visibility calculation. Step S2: Use on-board sensors to acquire vehicle operating status and front environment information, and invert the vehicle's available line of sight based on the virtual obstacle generation and recognition results of multi-sensor fusion. Step S3: Based on the database and the available line-of-sight distance of the vehicle, calculate the maximum speed limit for the UAV-assisted operating conditions according to the line-of-sight safety threshold. V max Maximum speed limit under vehicle operating conditions ; Step S4: Maximum speed limit of the UAV under assisted operating conditions obtained in step S3. V max Maximum speed limit for self-driving vehicles The driving ability margin improvement index was calculated, and an adaptive grading evaluation method was used to calculate the driving ability margin improvement effect of autonomous trucks on mountain roads.

2. The method for evaluating the improvement of drivability on mountain roads for autonomous trucks according to claim 1, characterized in that, Step S1 specifically includes: Step S11: Using a vehicle-mounted multi-rotor drone equipped with lidar and high-definition vision sensors, aerial photography is conducted along the driving direction of the autonomous truck to collect road spatial point cloud data and image data within the range of the preset route markers. Aerial photography parameters include: flight altitude controlled between 50-200m, flight speed not exceeding 15m / s, and flight path overlap rate ≥80%; Step S12: Preprocess the collected point cloud data and image data, including point cloud denoising, image stitching, time synchronization and coordinate unification processing, and reconstruct the three-dimensional point cloud based on the UAV pose parameters. At the same time, perform spatiotemporal alignment and fusion with the vehicle multi-sensor data to construct a three-dimensional spatial model of the road ahead. Then, perform road vision field calculation and occlusion relationship analysis in the three-dimensional spatial model to obtain the length of the visible path ahead, Su. Step S13: Based on the vehicle's current speed Assisted line-of-sight path length for drones Perform time-based corrections to obtain the corrected effective usable line of sight. Its expression is: （1） τu represents the data transmission delay between the UAV and the vehicle-mounted system. Step S14: The road alignment feature parameters extracted from the three-dimensional spatial model and the dynamically updated effective available sight distance are stored in a structured manner to construct a dynamic visibility enhancement information database Ω for mountain roads. The information database for improving visibility on mountain roads includes: the length of the visible path ahead under drone-assisted conditions. The range of road longitudinal position station numbers corresponding to the visible path, and the road longitudinal slope i and road surface longitudinal friction coefficient f used for sight distance safety calculation. L .

3. The method for evaluating the improvement of drivability on mountain roads for autonomous trucks according to claim 2, characterized in that, Step S2 specifically includes: Step S21: Use the multi-source onboard perception sensors mounted on the autonomous truck to obtain vehicle operating status information and road environment information ahead; The operational status information includes: during the time interval from when the Level 3 autonomous driving system issues a takeover request to when the driver performs the takeover action, the vehicle uses a preset deceleration 'a'. dp Its typical value is 2.5 m·s -2 Driver perception-braking reaction time t pb_h ; Autonomous driving system perception-braking reaction time t pb_s ; Autonomous driving system perception and reaction time t p_s Driver takeover reaction time t T ; The vehicle-mounted perception sensor includes one or more of vehicle-mounted cameras and lidar; Step S22: Based on the vehicle's current position and the road alignment information ahead, construct a set of virtual obstacles along the vehicle's travel reference path within a preset distance range; Obstacles include one or more of the following: vehicles ahead of the road, roadside fixed facilities, and terrain obstructions; The set of virtual obstacles is sequentially distributed at certain intervals along the road travel direction; Step S23: Match and analyze the virtual obstacle set with the environmental perception results after multi-sensor fusion to determine the identifiable status of each virtual obstacle and determine the position of the farthest virtual obstacle that can be stably identified under the current vehicle perception conditions. Step S24: Calculate the distance along the road travel direction between the location of the farthest virtual obstacle and the current location of the vehicle, as the available line-of-sight S of the autonomous truck under the vehicle's perception conditions. a .

4. The method for evaluating the improvement of drivability on mountain roads for autonomous trucks according to claim 3, characterized in that, The implementation method for step S3 is as follows: Step S31: Calculate the maximum speed limit V of the UAV in the assisted operating condition under the condition of meeting line-of-sight safety. max The formula is: （2） Equation (2) represents Level 1 and Level 2 autonomous driving systems with UAV-assisted operating conditions; （3） Equation (3) is a Level 3 autonomous driving system with unmanned aerial vehicle (UAV) assistance. （4） Equation (4) is a Level 4 autonomous driving system with unmanned aerial vehicle (UAV) assistance. Step S32: Calculate the maximum speed limit for the vehicle under the condition of meeting line-of-sight safety requirements. The formula is: （5） Equation (5) represents Level 1 and Level 2 automated driving systems under autonomous vehicle operating conditions; （6） Equation (6) represents a Level 3 autonomous driving system under autonomous vehicle operating conditions; （7） Equation (7) is the Level 4 autonomous driving system for the vehicle.

5. The method for evaluating the improvement of drivability on mountain roads for autonomous trucks according to claim 4, characterized in that, The implementation method for step S4 is as follows: Step S41: Based on the difference in safe speed limits between UAV-assisted and autonomous vehicle-perception conditions, construct a drivability margin improvement index ΔV, the calculation formula of which is: （8） Step S42: Construct a sample dataset containing the drivability margin improvement index ΔV calculated under multiple mountain road sections or multiple operating conditions. ; Step S43: Standardize the sample dataset and use an unsupervised clustering algorithm for adaptive clustering analysis to automatically determine the number of levels of driving suitability improvement effect; Step S44: Determine the threshold range corresponding to each level of driving suitability improvement effect based on the clustering results, and form the hierarchical boundary between different levels; Step S45: Establish a quantitative correspondence between each level of drivability improvement effect and the drivability margin improvement index ΔV, and conduct a graded evaluation of the drivability improvement effect of mountain roads based on the interval of the drivability margin improvement index ΔV.

6. The method for evaluating the improvement of drivability on mountain roads for autonomous trucks according to claim 5, characterized in that, When the pilotability margin improvement index ΔV is in the high-level range determined by the clustering algorithm, it is judged that the pilotability improvement effect is significant, indicating that the maximum safe speed under drone assistance conditions is greatly improved and the operational safety margin is significantly enhanced. When the pilotability margin improvement index ΔV is in the medium level range, it is judged that the pilotability improvement effect is average, indicating that the drone assistance has a certain effect on improving the upper limit of safe speed. When the drivability margin improvement index ΔV is in the low-level range, it is judged that the drivability improvement effect is weak or has not produced an effective improvement, indicating that the drone assistance has failed to significantly improve the forward visibility conditions of the vehicle.

7. The method for evaluating the improvement of drivability on mountain roads for autonomous trucks according to claim 5, characterized in that, The unsupervised clustering algorithm is one or more of the following: adaptive K-means clustering algorithm, adaptive hierarchical clustering algorithm, or adaptive density clustering algorithm.

8. A system for evaluating the improvement of drivability on mountain roads for autonomous trucks, comprising a processor, a memory, and a bus, wherein the memory stores machine-readable instructions executed by the processor; characterized in that, When the system is running, the processor and the memory communicate via a bus, and the machine-readable instructions are executed by the processor as described in any one of claims 1 to 7.

Images