Tunnel section autonomous driving safety speed determination method based on support vector regression

Abstract

The present application relates to a kind of tunnel section automatic driving safety speed determination method based on support vector regression, comprising the following steps: S1, obtains automatic driving visual range related information, tunnel section design information and weather environment information;S2, using the information obtained in step S1, using virtual test method to obtain tunnel section automatic driving effective line of sight distance and construct the database that can obtain visual range;S3, based on the database that can obtain visual range constructed in step S2, using principal component analysis to extract automatic driving available visual range characteristic parameter;S4, using the automatic driving available visual range characteristic parameter extracted in step S3, based on support vector machine regression method establishes automatic driving available visual range model;S5, based on the automatic driving available visual range model established in step S4, according to visual range safety principle, calculates tunnel section automatic driving safety speed.The method is conducive to effectively obtaining the speed value that meets the tunnel section automatic driving safety driving requirement, and then improve the safety of automatic driving.

Description

Technical Field

[0001] This invention belongs to the field of autonomous driving safety technology, specifically relating to a method for determining the safe speed of autonomous driving in tunnel sections based on support vector regression. Background Technology

[0002] With increasingly mature hardware, intelligent software theories and systems, autonomous driving technology, combined with emerging technologies such as intelligent connectivity and high-precision maps, is expected to provide effective solutions for improving vehicle driving safety and ride comfort, as well as enhancing road traffic efficiency and capacity. However, tunnels, as existing important road engineering structures, have a semi-enclosed architectural environment that will have many direct or indirect negative impacts on these beneficial effects. For example, due to the tunnel's structural obstruction, network and positioning signals are weakened, forcing autonomous vehicles to rely primarily on their autonomous intelligent systems to complete driving operations. Furthermore, the tunnel's internal structure easily obstructs visibility, and the limited lighting and visibility conditions within the tunnel result in poorer visibility for the autonomous vehicle's perception system and the driver after taking over driving compared to open roads. Clearly, these negative impacts will primarily affect the visibility conditions for autonomous driving.

[0003] However, existing research in this field is still in its early stages, mainly focusing on single or complex alignment conditions on open roads, such as circular curves, longitudinal profile alignments, and combined alignments. Few studies have addressed tunnel sections, exploring the driving state of autonomous vehicles in such environments. It is important to note that speed, as one of the most direct behavioral manifestations of a vehicle after being manipulated by a user (driver or autonomous driving system), is crucial for ensuring vehicle safety by controlling its value to meet sufficient safety threshold requirements. Furthermore, speed control is also a key task of the decision-making layer in autonomous driving systems. It is also important to note that autonomous driving systems face massive amounts of multi-source heterogeneous data during decision-making, such as tunnel road conditions, weather conditions, and system conditions; therefore, employing appropriate data processing methods is particularly crucial. Thus, effectively calculating the safe speed of autonomous vehicles in tunnel sections has become a critical issue that urgently needs to be addressed to ensure safe autonomous driving, and it has significant practical implications. Summary of the Invention

[0004] The purpose of this invention is to provide a method for determining the safe speed of autonomous driving in tunnel sections based on support vector regression. This method is beneficial for effectively obtaining speed values ​​that meet the safe driving requirements of autonomous driving in tunnel sections, thereby improving the safety of autonomous driving.

[0005] To achieve the above objectives, the technical solution adopted by this invention is: a method for determining the safe speed of automated driving in tunnel sections based on support vector regression, comprising the following steps:

[0006] Step S1: Obtain information related to the line-of-sight distance for autonomous driving, tunnel section design information, and weather and environmental information;

[0007] Step S2: Using the information obtained in step S1, a virtual testing method is used to obtain the effective line-of-sight distance for autonomous driving in the tunnel section and to construct an obtainable line-of-sight database.

[0008] Step S3: Based on the available line-of-sight database constructed in step S2, principal component analysis is used to extract the available line-of-sight feature parameters for autonomous driving.

[0009] Step S4: Using the autonomous driving obtainable line-of-sight feature parameters extracted in step S3, establish an autonomous driving obtainable line-of-sight model based on the support vector machine regression method.

[0010] Step S5: Based on the autonomous driving obtainable line-of-sight model established in step S4, calculate the safe speed for autonomous driving in the tunnel section according to the line-of-sight safety principle.

[0011] Furthermore, in step S1, the autonomous driving line-of-sight related information includes at least: autonomous vehicle perception sensor technical parameters and installation information, autonomous vehicle perception function information, autonomous vehicle driver takeover reaction time, autonomous vehicle preset braking deceleration, autonomous driving system or driver takeover braking deceleration, and autonomous driving system perception reaction time.

[0012] The tunnel section design information includes at least: tunnel alignment design information, tunnel cross section design information, tunnel cross passage and parallel passage design information, as well as the location and shape information of traffic signs, lighting, and traffic monitoring facilities that may obstruct the view;

[0013] The weather environment information includes at least the weather type, wherein when the weather type is fog, it includes at least the fog intensity level and visibility.

[0014] Furthermore, the technical parameters of the autonomous vehicle's perception sensors include at least: perception sensor type, detection distance, horizontal field of view, vertical field of view, horizontal angular resolution, and vertical angular resolution;

[0015] The installation information perceived by the autonomous vehicle includes at least: installation height and number of installations;

[0016] The perception information of the autonomous vehicle includes at least: static obstacle perception algorithm.

[0017] Furthermore, step S2 is implemented as follows:

[0018] Step S21: Based on the acquired autonomous driving line-of-sight related information, tunnel section design information and weather environment information, a virtual tunnel and environment model and autonomous driving test scenario are built in the software environment using a virtual testing method to map the real tunnel section autonomous driving scenario.

[0019] Step S22: Conduct virtual testing to obtain the effective line-of-sight distance for autonomous driving in the tunnel section;

[0020] Step S23: Using autonomous driving line-of-sight related information, tunnel section design information, and weather environment information as the data link input parameter set, and using the effective line-of-sight distance of autonomous driving in the tunnel section as the data link output parameter, the data link input-output parameters are matched to construct an obtainable line-of-sight database.

[0021] Furthermore, step S3 is implemented as follows:

[0022] Step S31: For the data chain input parameter set in the available database, examine the correlation between parameters, and obtain the Pearson, Spearman, or Kendall correlation results and significance according to the parameter value type;

[0023] Step S32: Perform KMO test and Bartlett's test of sphericity on the parameters to ensure the structural validity of the parameter data;

[0024] Step S33: Perform principal component analysis on the parameters. Using the component eigenvalue greater than 1 as the extraction criterion, extract the principal components that reflect the line-of-sight information of autonomous driving, the tunnel section design information and the weather environment information as the line-of-sight feature parameters that autonomous driving can acquire, and obtain the component score coefficient matrix.

[0025] Furthermore, step S4 is implemented as follows:

[0026] Step S41: Based on the component score coefficient matrix obtained in step S33, update the data chain input parameters in the available line-of-sight database, add the corresponding available line-of-sight feature parameter data, and use it as an input variable;

[0027] Step S42: Set the effective line-of-sight distance of the tunnel section that matches the available line-of-sight feature parameter data to obtain the data link output parameters in the line-of-sight database as output variables;

[0028] Step S43: Use radial basis functions as the kernel function of the support vector machine, and use cross-validation to optimize the parameters of the support vector machine to establish the support vector machine model;

[0029] Step S44: Normalize the samples composed of input and output variables, randomly divide the samples according to a 7:3 ratio, and use them as the model training set and test set respectively. Then, substitute them into the support vector machine model for training and obtain the optimal model as the autonomous driving line-of-sight model.

[0030] Furthermore, the implementation method of step S5 is as follows:

[0031] Step S51: Based on the line-of-sight safety principle, enable the autonomous driving system to obtain the line-of-sight distance S. a With the demand for autonomous driving line of sight S r equal;

[0032] Step S52: When the autonomous driving automation level of the vehicle is Level 3, the safe speed V for autonomous driving in the tunnel section is... safe The calculation formula is:

[0033]

[0034] In the formula, A d For braking deceleration after the automatic driving system or driver takes over, A dp Pre-set braking deceleration for autonomous vehicles, t S For the perception and reaction time of the autonomous driving system, t T For the driver takeover reaction time of autonomous vehicles, i G The longitudinal slope of the road in the tunnel section; where the road is uphill in the tunnel section, i G >0; When the tunnel section is downhill, i G <0;

[0035] When the autonomous vehicle's driving automation level is 4 or 5, the safe speed V for autonomous driving in the tunnel section is... safe The calculation formula is:

[0036]

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] (1) This invention uses multi-source heterogeneous data such as autonomous driving line-of-sight related information, tunnel section design information and weather environment information as model input, and adopts a support vector machine model suitable for this type of data to calculate the safe speed of autonomous driving in tunnel section, which makes the calculation results of this invention more realistic and effective, filling the gap in this field.

[0039] (2) Based on existing autonomous driving virtual testing technology, this invention avoids the drawback that theoretical calculations can easily lead to calculation results that are too ideal than the actual situation. At the same time, it can save costs and ensure test safety more than on-site testing.

[0040] (3) The safe speed for autonomous driving in tunnel sections calculated by the present invention can provide a theoretical basis for the formulation of speed limit schemes for tunnel sections, and make up for the shortcomings of existing technical solutions that only specify the design and operation conditions for a single level of autonomous driving vehicle and are difficult to propose convenient management measures from the perspective of road traffic management. Attached Figure Description

[0041] Figure 1 This is a flowchart illustrating the implementation of the method for determining the safe speed of autonomous driving in tunnel sections based on support vector regression, provided in an embodiment of the present invention.

[0042] Figure 2 This is a flowchart illustrating the construction of an acquireable line-of-sight database in an embodiment of the present invention;

[0043] Figure 3 This is a flowchart of extracting obtainable line-of-sight feature parameters for autonomous driving in an embodiment of the present invention;

[0044] Figure 4 This is a flowchart of an embodiment of the present invention for establishing an autonomous driving obtainable line-of-sight model based on the support vector machine regression method. Detailed Implementation

[0045] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0046] It should be noted that the following detailed descriptions are exemplary 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.

[0047] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments 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.

[0048] like Figure 1 As shown, this embodiment provides a method for determining the safe speed of autonomous driving in tunnel sections based on support vector regression, including the following steps:

[0049] Step S1: Obtain information related to the line-of-sight distance for autonomous driving, tunnel section design information, and weather environment information.

[0050] The autonomous driving line-of-sight related information includes at least: autonomous vehicle perception sensor technical parameters and installation information, autonomous vehicle perception function information, autonomous vehicle driver takeover reaction time, autonomous vehicle preset braking deceleration, autonomous vehicle braking deceleration after autonomous driving system or driver takeover, and autonomous driving system perception reaction time.

[0051] The tunnel section design information includes at least: tunnel alignment design information, tunnel cross-section design information, tunnel cross passage and parallel passage design information, as well as the location and shape information of traffic signs, lighting, and traffic monitoring facilities that may obstruct the view.

[0052] The weather environment information includes at least the weather type, wherein when the weather type is fog, it includes at least the fog intensity level and visibility.

[0053] In this embodiment, the technical parameters of the autonomous vehicle's perception sensors include at least: sensor type, detection distance, horizontal field of view, vertical field of view, horizontal angular resolution, and vertical angular resolution. The autonomous vehicle's perception installation information includes at least: installation height and number of sensors installed. The autonomous vehicle's perception function information includes at least: a static obstacle perception algorithm.

[0054] Specifically, information related to the line-of-sight for autonomous driving can be obtained through on-site data collection or by gathering existing research results. Tunnel design information can be obtained through on-site data collection or by providing relevant data from road design departments. Weather and environmental information can be obtained through on-site data collection or by statistical analysis of publicly available weather data.

[0055] Step S2: Using the autonomous driving line-of-sight related information, tunnel section design information and weather environment information obtained in Step S1, a virtual testing method is used to obtain the effective line-of-sight distance for autonomous driving in the tunnel section and construct an obtainable line-of-sight database.

[0056] like Figure 2 As shown, the implementation method of step S2 is as follows:

[0057] Step S21: Based on the acquired autonomous driving line-of-sight information, tunnel section design information, and weather environment information, a virtual tunnel and environment model, as well as an autonomous driving test scenario, are built in the software environment using a virtual testing method to map the real tunnel section autonomous driving scenario.

[0058] Among them, the software environment-based virtual testing method can be achieved independently or jointly with scenario-based autonomous driving virtual testing software such as CarSim and PreScan. The modeling effectiveness of the above software has been widely verified in the field.

[0059] Step S22: Conduct virtual testing to obtain the effective line-of-sight distance for autonomous driving in the tunnel section.

[0060] Among them, the effective line-of-sight distance for autonomous driving in tunnel sections refers to the driving path distance between the autonomous vehicle and a stationary obstacle, which is the farthest stationary obstacle in the tunnel section ahead that the autonomous vehicle can detect.

[0061] Step S23: Using autonomous driving line-of-sight related information, tunnel section design information, and weather environment information as the data link input parameter set, and using the effective line-of-sight distance of autonomous driving in the tunnel section as the data link output parameter, the data link input-output parameters are matched to construct an obtainable line-of-sight database.

[0062] The establishment of a line-of-sight database can rely on data management software such as SPSS, Origin, Excel, and MATLAB.

[0063] Step S3: Based on the available line-of-sight database constructed in step S2, principal component analysis is used to extract the available line-of-sight feature parameters for autonomous driving.

[0064] like Figure 3 As shown, the implementation method of step S3 is as follows:

[0065] Step S31: For the data chain input parameter set in the available database, examine the correlation between parameters, and obtain the Pearson, Spearman, or Kendall correlation results and significance based on the parameter value type.

[0066] Step S32: Perform the KMO (Kaiser-Meyer-Olkin) test and Bartlett's test of sphericity on the parameters to ensure the structural validity of the parameter data.

[0067] Step S33: Perform principal component analysis on the parameters. Using the component eigenvalue greater than 1 as the extraction criterion, extract the principal components that reflect the line-of-sight information of autonomous driving, the tunnel section design information and the weather environment information as the line-of-sight feature parameters that autonomous driving can acquire, and obtain the component score coefficient matrix.

[0068] The steps required to perform principal component analysis can be achieved using data management software such as SPSS and Origin.

[0069] Step S4: Using the autonomous driving obtainable line-of-sight feature parameters extracted in step S3, establish an autonomous driving obtainable line-of-sight model based on the support vector machine regression method.

[0070] like Figure 4 As shown, the implementation method of step S4 is as follows:

[0071] Step S41: Based on the component score coefficient matrix obtained in step S33, update the data chain input parameters in the available view distance database, add the corresponding available view distance feature parameter data, and use it as an input variable.

[0072] Step S42: Set the effective line-of-sight distance of the tunnel section that matches the available line-of-sight feature parameter data to obtain the data link output parameters in the line-of-sight database as output variables.

[0073] Step S43: Use radial basis functions as the kernel function of the support vector machine, and use cross-validation to optimize the parameters of the support vector machine to establish the support vector machine model.

[0074] Step S44: Normalize the samples composed of input and output variables, randomly divide the samples according to a 7:3 ratio, and use them as the model training set and test set respectively. Then, substitute them into the support vector machine model for training and obtain the optimal model as the autonomous driving line-of-sight model.

[0075] The steps required for support vector machine regression modeling can be implemented using MATLAB and the Libsvm toolbox.

[0076] Step S5: Based on the autonomous driving obtainable line-of-sight model established in step S4, calculate the safe speed for autonomous driving in the tunnel section according to the line-of-sight safety principle.

[0077] In this embodiment, step S5 is implemented as follows:

[0078] Step S51: Based on the line-of-sight safety principle, enable the autonomous driving system to obtain the line-of-sight distance S. a With the demand for autonomous driving line of sight S r equal.

[0079] When considering the safety principle of sight distance, the required parking sight distance value is usually compared with the available sight distance value.

[0080] Step S52: When the autonomous driving automation level of the vehicle is Level 3, the safe speed V for autonomous driving in the tunnel section is... safe The calculation formula is:

[0081]

[0082] In the formula, A d For braking deceleration after the automatic driving system or driver takes over, A dp Pre-set braking deceleration for autonomous vehicles, t S For the perception and reaction time of the autonomous driving system, t T For the driver takeover reaction time of autonomous vehicles, i GThe longitudinal slope of the road in the tunnel section; where the road is uphill in the tunnel section, i G >0; When the tunnel section is downhill, i G <0.

[0083] When the autonomous vehicle's driving automation level is 4 or 5, the safe speed V for autonomous driving in the tunnel section is... safe The calculation formula is:

[0084]

[0085] The safe speed can be calculated based on the key information above and the solution provided in this embodiment.

[0086] In summary, this invention provides a method for determining the safe speed of autonomous driving in tunnel sections based on support vector regression. By utilizing multi-source heterogeneous data, including autonomous driving line-of-sight information, tunnel section design information, and weather environment information, as model input, and employing a support vector machine model suitable for this type of data, the safe speed of autonomous driving in tunnel sections is calculated. This provides an effective technical means for the actual safe operation of autonomous vehicles in tunnel sections. The method designed in this invention makes the calculation results more realistic and effective, filling a gap in this field. It avoids the drawback of theoretical calculations leading to overly idealized results compared to actual conditions, while also being more cost-effective and ensuring test safety compared to field testing. Furthermore, it overcomes the shortcomings of existing technologies that only specify the design and operating conditions for a single level of autonomous driving vehicles, making it difficult to propose convenient management measures from a road traffic management perspective.

[0087] 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.

[0088] 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 1 One or more processes and / or boxes Figure 1A device that provides the functions specified in one or more boxes.

[0089] 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.

[0090] 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.

[0091] 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.

Claims

1. A method for determining the safe speed of automated driving in tunnel sections based on support vector regression, characterized in that, Includes the following steps: Step S1: Obtain information related to the line-of-sight distance for autonomous driving, tunnel section design information, and weather and environmental information; Step S2: Using the information obtained in step S1, a virtual testing method is used to obtain the effective line-of-sight distance for autonomous driving in the tunnel section and to construct an obtainable line-of-sight database. Step S3: Based on the available line-of-sight database constructed in step S2, principal component analysis is used to extract the available line-of-sight feature parameters for autonomous driving. Step S4: Using the autonomous driving obtainable line-of-sight feature parameters extracted in step S3, establish an autonomous driving obtainable line-of-sight model based on the support vector machine regression method. Step S5: Based on the autonomous driving obtainable line-of-sight model established in step S4, calculate the safe speed for autonomous driving in the tunnel section according to the line-of-sight safety principle. In step S1, the autonomous driving line-of-sight related information includes at least: autonomous vehicle perception sensor technical parameters and installation information, autonomous vehicle perception function information, autonomous vehicle driver takeover reaction time, autonomous vehicle preset braking deceleration, autonomous vehicle braking deceleration after autonomous driving system or driver takeover, and autonomous driving system perception reaction time. The tunnel section design information includes at least: tunnel alignment design information, tunnel cross section design information, tunnel cross passage and parallel passage design information, as well as the location and shape information of traffic signs, lighting, and traffic monitoring facilities that may obstruct the view; The weather environment information includes at least: weather type, wherein when the weather type is fog, it includes at least the fog intensity level and visibility; The technical parameters of the autonomous vehicle's perception sensors include at least: perception sensor type, detection distance, horizontal field of view, vertical field of view, horizontal angular resolution, and vertical angular resolution. The installation information perceived by the autonomous vehicle includes at least: installation height and number of installations; The autonomous vehicle's perception function information includes at least: a static obstacle perception algorithm; The implementation method for step S2 is as follows: Step S21: Based on the acquired autonomous driving line-of-sight related information, tunnel section design information and weather environment information, a virtual tunnel and environment model and autonomous driving test scenario are built in the software environment using a virtual testing method to map the real tunnel section autonomous driving scenario. Step S22: Conduct virtual testing to obtain the effective line-of-sight distance for autonomous driving in the tunnel section; Step S23: Using autonomous driving line-of-sight related information, tunnel section design information and weather environment information as the data link input parameter set, and using the effective line-of-sight distance of autonomous driving in the tunnel section as the data link output parameter, match the data link input-output parameters to construct an obtainable line-of-sight database. The implementation method for step S3 is as follows: Step S31: For the data chain input parameter set in the available database, examine the correlation between parameters, and obtain the Pearson, Spearman, or Kendall correlation results and significance according to the parameter value type; Step S32: Perform KMO test and Bartlett's test of sphericity on the parameters to ensure the structural validity of the parameter data; Step S33: Perform principal component analysis on the parameters. Using the component eigenvalue greater than 1 as the extraction criterion, extract the principal components that reflect the line-of-sight information of autonomous driving, the tunnel section design information and the weather environment information as the line-of-sight feature parameters that autonomous driving can acquire, and obtain the component score coefficient matrix. The implementation method for step S4 is as follows: Step S41: Based on the component score coefficient matrix obtained in step S33, update the data chain input parameters in the available line-of-sight database, add the corresponding available line-of-sight feature parameter data, and use it as an input variable; Step S42: Set the effective line-of-sight distance of the tunnel section that matches the available line-of-sight feature parameter data to obtain the data link output parameters in the line-of-sight database as output variables; Step S43: Use radial basis functions as the kernel function of the support vector machine, and use cross-validation to optimize the parameters of the support vector machine to establish the support vector machine model; Step S44: Normalize the samples composed of input and output variables, randomly divide the samples according to a 7:3 ratio, and use them as the model training set and test set respectively. Then, substitute them into the support vector machine model for training and obtain the optimal model as the autonomous driving line-of-sight model.

Images