A regional security risk assessment method, system, device and storage medium

Abstract

Embodiments of the present specification provide a regional safety risk assessment method, system, device and storage medium. The method comprises: acquiring road section data of a plurality of road sections in a target region; based on the road section data, extracting risk features, the risk features including static features and dynamic features; based on at least the risk features, determining risk information of the target region through a regional safety risk assessment model.

Description

TECHNICAL FIELD

[0001] The present specification relates to the technical field of automatic driving, and in particular to a regional safety risk assessment method, system, device and storage medium. BACKGROUND

[0002] With the development of automatic driving technology, the process of automatic driving technology testing and commercial operation is promoted. The safety of automatic driving technology testing and demonstration application is controllable and orderly, and a reliable, mature and universal calculation method is needed to evaluate the safety risk of the region and the road in the region, so as to guide the selection of automatic driving testing and commercial operation road. The existing landmark evaluation system is mainly carried out in an artificial manner, and has limitations in data collection, system updating and data volume, and lacks universality, extensibility and reliability. Therefore, it is necessary to provide a regional safety risk assessment method, system, device and storage medium to improve the universality, extensibility and reliability of the regional safety risk assessment system. SUMMARY

[0003] One of the embodiments of the present specification provides a regional safety risk assessment method, the method comprising: acquiring road segment data of a plurality of road segments in a target region; based on the road segment data, extracting risk features, the risk features comprising static features and dynamic features; and based on at least the risk features, determining risk information of the target region through a regional safety risk assessment model.

[0004] One of the embodiments of the present specification provides a regional safety risk assessment system, the system comprising: an acquisition module configured to acquire road segment data of a plurality of road segments in a target region; an extraction module configured to extract risk features based on the road segment data, the risk features comprising static features and dynamic features; and a determination module configured to determine risk information of the target region through a regional safety risk assessment model based on at least the risk features.

[0005] One of the embodiments of the present specification provides a regional safety risk assessment device, the device comprising: at least one storage medium storing computer instructions; and at least one processor executing the computer instructions to implement the method of any embodiment of the present specification.

[0006] One of the embodiments of the present specification provides a computer-readable storage medium storing computer instructions, when the computer reads the computer instructions, the computer executes the method of any embodiment of the present specification.

[0007] One of the embodiments of the present specification provides a computer program product comprising a computer program or instructions, which, when executed by a processor, implements the method of any embodiment of the present specification. BRIEF DESCRIPTION OF DRAWINGS

[0008] The present specification will be further explained in the way of example embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting, in which the same numbers refer to the same structures, wherein:

[0009] Figure 1 is a schematic diagram of an application scenario of an example regional security risk assessment system according to some embodiments of the present specification;

[0010] Figure 2 is a block diagram of an example regional security risk assessment system according to some embodiments of the present specification;

[0011] Figure 3 is a flowchart of an example regional security risk assessment method according to some embodiments of the present specification;

[0012] Figure 4 is a flowchart of an example risk information determination method according to some embodiments of the present specification;

[0013] Figure 5 is a flowchart of an example road segment risk level determination method according to some embodiments of the present specification;

[0014] Figure 6 is a schematic diagram of an example important feature according to some embodiments of the present specification;

[0015] Figure 7 is a flowchart of an example road risk level determination method according to some embodiments of the present specification;

[0016] Figure 8 is a schematic diagram of an example regional security risk assessment process according to some embodiments of the present specification;

[0017] Figure 9 is a schematic diagram of an example data acquisition and processing process according to some embodiments of the present specification;

[0018] Figure 10 is a flowchart of an example feature space determination method according to some embodiments of the present specification;

[0019] Figure 11 is a schematic diagram of an example feature space construction and optimization process according to some embodiments of the present specification; and

[0020] Figure 12 is a flowchart of an example regional security risk assessment model training method according to some embodiments of the present specification. DETAILED DESCRIPTION

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present specification, the drawings needed to be used in the embodiment description will be briefly introduced as follows. Obviously, the drawings in the following description are only some examples or embodiments of the present specification, and for those skilled in the art, the present specification can also be applied to other similar scenarios without creative labor on the basis of these drawings. Unless it is clear from the language environment or otherwise stated, the same reference numbers in the drawings represent the same structures or operations.

[0022] It should be understood that the "system", "device", "unit" and / or "module" used herein is a method for distinguishing different components, elements, parts, sections or assemblies at different levels. However, if other words can achieve the same purpose, the words can be replaced by other expressions.

[0023] As shown in the specification and claims, unless the context clearly indicates otherwise, "one", "a", "an", and / or "the" do not refer to the singular, but can also include the plural. Generally speaking, the terms "comprise" and "include" only indicate the inclusion of the steps and elements explicitly identified, and these steps and elements do not constitute an exclusive list, and the method or device can also include other steps or elements.

[0024] Flowcharts are used in the present specification to illustrate the operations performed by the system according to the embodiments of the present specification. It should be understood that the preceding or subsequent operations are not necessarily performed in sequence. On the contrary, each step can be processed in reverse order or simultaneously. At the same time, other operations can be added to these processes, or one or more steps of the operation can be removed from these processes.

[0025] Figure 1 is a schematic diagram of an application scenario of an exemplary regional security risk assessment system according to some embodiments of the present specification. In some embodiments, the regional security risk assessment system 100 can be an online service platform for autonomous driving services. In some embodiments, the regional security risk assessment system 100 can be used to perform security risk assessment of regions and / or roads within the regions related to autonomous driving services.

[0026] In some embodiments, as shown in Figure 1 the regional security risk assessment system 100 can include a server 110, a network 120, a data acquisition device 130, an information source 140, and a storage device 150.

[0027] The server 110 can be a single server or a group of servers. The group of servers can be centralized or distributed (e.g., the server 110 can be a distributed system). In some embodiments, the server 110 can be local or remote. For example, the server 110 can access information and / or data stored in the data acquisition device 130, the information source 140, and / or the storage device 150 through the network 120. For another example, the server 110 can be directly connected to the data acquisition device 130, the information source 140, and / or the storage device 150 to access the stored information and / or data. In some embodiments, the server 110 can be implemented on a cloud platform. By way of example only, the cloud platform can include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, between clouds, a multiple cloud, and / or the like, or any combination of the foregoing examples.

[0028] In some embodiments, the server 110 can include a processing device 112 for implementing the example methods and / or systems described in this specification. For example, the processing device 112 can acquire road segment data of a plurality of road segments within a target region from the data acquisition device 130, and extract risk features based on the road segment data. Further, the processing device 112 can determine risk information of the target region by a regional security risk assessment model based on at least the risk features.

[0029] In some embodiments, the server 110 and / or the processing device 112 can be implemented by a computing device, for example, a computing device including a processor, a memory, a network interface, a communication interface, a display device, and / or the like.

[0030] The network 120 can facilitate the exchange of information and / or data. In some embodiments, components in the regional security risk assessment system 100 (e.g., the server 110, the data acquisition device 130, the information source 140, and / or the storage device 150) can send information and / or data to other components in the regional security risk assessment system 100 through the network 120. For example, the server 110 can acquire road segment data from the data acquisition device 130, the information source 140, and / or the storage device 150 through the network 120. In some embodiments, the network 120 can be any one or a combination of a wired network or a wireless network. For example, the network 120 can include a cable network, a wired network, a fiber optic network, a telecommunication network, an intranet, the Internet, a local area network (LAN), a wide area network (WAN), a wireless local area network (WLAN), a metropolitan area network (MAN), a public switched telephone network (PSTN), a Bluetooth network, a ZigBee network, a near-field communication (NFC) network, and / or the like, or any combination of the aforementioned examples. In some embodiments, the network 120 can include one or more network access points. For example, the network 120 can include wired or wireless network access points, such as base stations and / or Internet exchange points 120-1, 120-2, and / or the like. Through the network access points, the components of the regional security risk assessment system 100 can connect to the network 120 to exchange data and / or information.

[0031] The data acquisition device 130 can acquire data related to traffic participants. In some embodiments, the data acquisition device 130 can acquire vehicle travel data (e.g., vehicle travel trajectories (e.g., GPS trajectories), vehicle travel speeds, vehicle travel accelerations, vehicle travel directions, travel road segments, vehicle traffic volumes, and / or the like). In some embodiments, the data acquisition device 130 can also acquire vehicle crash-related data (e.g., whether a crash occurred, the number of crashes, the severity of crashes, and / or the like) and / or traffic participant data (e.g., motor vehicle traffic volumes, non-motor vehicle traffic volumes, pedestrian traffic volumes, roadside parking conditions, and / or the like). In some embodiments, the data acquisition device 130 can acquire the related data in real time. In some embodiments, the data acquisition device 130 can acquire the related data periodically (e.g., at certain time intervals).

[0032] In some embodiments, the data acquisition device 130 can include, but is not limited to, a mobile device 130-1, an in-vehicle built-in device 130-2, a notebook computer 130-3, a desktop computer 130-4, etc., or any combination thereof. In some embodiments, the mobile device 130-1 can include, but is not limited to, a smartphone, a personal digital assistance (PDA), a tablet computer, a palm game console, smart glasses, a smart watch, a wearable device, a virtual display device, a display enhancement device, etc., or any combination thereof. In some embodiments, the in-vehicle built-in device 130-2 can include, but is not limited to, a driving recorder (e.g., a dashcam), an in-vehicle navigation positioning device, a vehicle sensor (e.g., a speed sensor, an acceleration sensor, a distance sensor), an in-vehicle computer, an in-vehicle head-up display (HUD), an on-board diagnostic system (OBD), etc., or any combination thereof.

[0033] In some embodiments, the data acquisition device 130 can include and / or communicate with a positioning device. In some embodiments, the positioning device can include a global positioning system (GPS), a GLONASS satellite navigation system, a BeiDou satellite navigation system, a Galileo satellite navigation system, a quasi-zenith satellite system (QZSS), a wireless fidelity (WiFi), etc., or any combination thereof.

[0034] The information source 140 can provide reference information for the regional safety risk assessment system 100. In some embodiments, the information source 140 can include and / or communicate with a road network data source to provide road network data (e.g., ground objects, speed management information, road segment names, road segment numbers, road segment flatness, road segment slope, road segment curvature, road segment length, road segment width, road segment direction, signs and markings, road surface skid resistance coefficient, intersection attributes, traffic light timing, etc.). In some embodiments, the information source 140 can provide environmental data related to regional safety risk assessment (e.g., weather information, wind information, light, sight distance, tree shade blocking information, etc.). In some embodiments, the information source 140 can provide other information related to regional safety risk assessment, such as assessment time, assessment region, legal and regulatory information, news information, etc. In some embodiments, the data in the information source 140 can be dynamically updated (e.g., real-time update, periodic update, etc.). For example, if there is a new road segment, a road segment type change, etc., the road network data can be dynamically updated.

[0035] The storage device 150 can store data and / or instructions. In some embodiments, the storage device 150 can store data acquired from the server 110, the data acquisition device 130, and / or the information source 140 (e.g., vehicle travel data, road network data, vehicle collision related data, environmental data, traffic actor data, etc.). In some embodiments, the storage device 150 can partition the data acquired from the server 110, the data acquisition device 130, and / or the information source 140 into a plurality of data warehouses, such as a vehicle travel data warehouse, a road network data warehouse, a vehicle collision data warehouse, an environmental data warehouse, a traffic actor data warehouse, etc. In some embodiments, the storage device 150 can store data and / or instructions for the server 110 to execute or use (e.g., data and / or instructions for implementing the example methods described in this specification).

[0036] In some embodiments, the storage device 150 can be connected with the network 120 to enable communication between the storage device 150 and other components in the regional safety risk assessment system 100 (e.g., the server 110, the data acquisition device 130, the information source 140, etc.). In some embodiments, the storage device 150 can be directly connected or in communication with other components in the regional safety risk assessment system 100 (e.g., the server 110, the data acquisition device 130, the information source 140, etc.).

[0037] In some embodiments, the information source 140 and / or the storage device 150 can be part of the server 110. In some embodiments, the information source 140 can be part of the storage device 150. In some embodiments, the storage device 150 can be part of the information source 140. In some embodiments, the information source 140 and the storage device 150 can share information or storage space.

[0038] It is noted that the above description of the regional safety risk assessment system 100 is for convenience of description only and should not limit the scope of this specification to the described embodiments. It is understood that one of ordinary skill in the art, upon understanding the principles of the system, can make various changes to the system and its components without departing from the principles.

[0039] Figure 2 is a block diagram of an example regional safety risk assessment system according to some embodiments of the present specification. In some embodiments, the regional safety risk assessment system 200 can be implemented by the processing device 112. In some embodiments, the regional safety risk assessment system 200 can include an acquisition module 210, an extraction module 220, and a determination module 230.

[0040] The acquisition module 210 can be used to acquire road segment data for multiple road segments within a target area. For more information on acquiring road segment data for multiple road segments within a target area, please refer to step 310 and its related description, which will not be repeated here.

[0041] The extraction module 220 can be used to extract risk features based on road segment data. In some embodiments, risk features may include static features and dynamic features. For more information on extracting risk features based on road segment data, please refer to step 320 and its related description, which will not be repeated here.

[0042] The determination module 230 can be used to determine the risk information of a target area based at least on risk characteristics, using a regional security risk assessment model. More details on determining the risk information of the target area can be found in step 330 and its related description, and will not be repeated here.

[0043] In some embodiments, the regional security risk assessment system 200 may further include a model training module. Figure 2 (Not shown in the text). In some embodiments, the model training module can be used to train a regional security risk assessment model. More information on training regional security risk assessment models can be found in [link to relevant documentation]. Figure 12 The details and related descriptions will not be repeated here.

[0044] It should be understood that Figure 2 The regional security risk assessment system 200 and its modules shown can be implemented in various ways, such as through hardware, software, or a combination of both. The system and its modules described in this specification can be implemented not only with hardware circuits such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips or transistors, or programmable hardware devices such as field-programmable gate arrays or programmable logic devices, but also with software, for example, executed by various types of processors, or with a combination of the aforementioned hardware circuits and software (e.g., firmware).

[0045] It should be noted that the above description of the regional security risk assessment system 200 and its modules is for the convenience of description, and does not limit the scope of the present specification to the embodiments described. It can be understood that, for those skilled in the art, after understanding the principles of the system, any combination of the modules can be made, or the subsystems can be connected to other modules without departing from the principles. For example, in some embodiments, the acquisition module 210, the extraction module 220, and the determination module 230 can be different modules in a system, or one module can implement the functions of two or more modules described above. For another example, each module can share a storage module, and each module can also have its own storage module. For another example, the model training module can also be an independent module independent of the regional security risk assessment system 200. Variations such as this are within the scope of protection of the present application.

[0046] Figure 3 is a flowchart of an exemplary regional security risk assessment method according to some embodiments of the present specification. In some embodiments, the flow 300 can be executed by the processing device 112 or the regional security risk assessment system 200. For example, the flow 300 can be stored in the form of a program or instructions in a storage device (for example, the storage device 150, the storage unit of the processing device 112), and when the processor or Figure 2 the modules shown execute the program or instructions, the flow 300 can be implemented. In some embodiments, the flow 300 can utilize one or more additional operations not described below, and / or be completed without one or more operations discussed below. In addition, the order of the operations shown is not limiting. Figure 3

[0047] At step 310, road segment data of a plurality of road segments in a target region is acquired. In some embodiments, step 310 can be executed by the acquisition module 210.

[0048] In some embodiments, the target region can be a region that needs to be subjected to regional security risk assessment. In some embodiments, the target region can include a plurality of road segments. In some embodiments, the plurality of road segments can be independent of or connected to each other. In some embodiments, the processing device 112 can acquire road segment data in units of road segments.

[0049] ​In some embodiments, the road segment data can include vehicle travel data, road network data, or vehicle collision related data on the road segment. In some embodiments, the vehicle travel data can include vehicle travel trajectory (e.g., GPS trajectory), vehicle travel speed, vehicle travel acceleration, vehicle travel direction, travel road segment, vehicle flow, etc., or any combination thereof. In some embodiments, the road network data can include ground objects, speed management information, road segment name, road segment number, road segment flatness, road segment slope, road segment curvature, road segment length, road segment width, road segment direction, sign and marking, road surface skid resistance coefficient, intersection attribute, traffic light timing, etc., or any combination thereof. In some embodiments, the vehicle collision related data can include whether a collision occurs, collision times, collision severity, etc., or any combination thereof.

[0050] In some embodiments, the road segment data can further include environmental data, traffic participant data, etc. of the road segment. In some embodiments, the environmental data can include weather information, wind information, light, sight distance, tree shade blocking information, etc., or any combination thereof. In some embodiments, the traffic participant data can include motor vehicle flow, non-motor vehicle flow, pedestrian flow, roadside parking situation, etc., or any combination thereof.

[0051] In some embodiments, the processing device 112 can obtain the vehicle travel data, vehicle collision related data, and / or traffic participant data from the data acquisition device 130. In some embodiments, the processing device 112 can obtain the road network data and / or environmental data from the information source 140. In some embodiments, the processing device 112 can obtain the above-mentioned data from the storage device 150.

[0052] In some embodiments, the processing device 112 can obtain the above-mentioned data within a certain time period (e.g., 1 day, 1 week, 1 month, 3 months, etc.). In some embodiments, the processing device 112 can obtain the above-mentioned data based on time intervals. For example, the processing device 112 can obtain the above-mentioned data corresponding to the morning peak period, the flat peak period, the evening peak period, and the night period, respectively, within 1 month.

[0053] At step 320, risk features are extracted based on the road segment data. In some embodiments, step 320 can be performed by the extraction module 220.

[0054] In some embodiments, the risk features can be features that have an impact on the regional safety risk assessment result. In some embodiments, the risk features can include static features and dynamic features. In some embodiments, the risk features can be determined by a feature space. In some embodiments, the feature space can be determined by a large number of sample features through screening, processing, expert participation, etc. More details about the feature space can be found in the following description.Figure 10 The related description of the foregoing is not repeated here.

[0055] In some embodiments, the static features can reflect the basic attributes of the road. In some embodiments, the static features can be divided according to roads and lanes. For example, Table 1 below exemplarily shows static feature examples.

[0056] Table 1 Static feature examples

[0057]

[0058]

[0059] In some embodiments, the dynamic features can reflect the dynamic attributes of the road. In some embodiments, the dynamic features can be divided according to behaviors, types, and time periods, and are freely combined.

[0060] Table 2 Dynamic feature examples

[0061]

[0062] As can be seen from Table 2, the dynamic features in Table 2 include different time periods. In some embodiments, in the process of regional safety risk assessment, the risk information of the road section of the target region can be evaluated in different time periods based on different time periods.

[0063] In some embodiments, taking a certain road section as an example, the processing device 112 can divide the road according to the dimensions of type, attribute, function level, etc. based on the road network data, so as to determine each static feature. In some embodiments, the processing device 112 can analyze and count the vehicle trajectories passing through the road section, so as to determine the longitudinal behavior and lateral behavior of the vehicle, etc. In some embodiments, the processing device 112 can statistically analyze the vehicles passing through the road section, so as to determine the vehicle types passing through the road section, the number of each vehicle type, etc.

[0064] In some embodiments, the processing device 112 can extract risk features by data extraction, calculation, integration, etc. on the road section data. In some embodiments, the processing device 112 can automatically extract risk features based on the road section data through a machine learning model.

[0065] Step 330, based at least on the risk features, determining the risk information of the target region through a regional safety risk assessment model. In some embodiments, step 330 can be performed by the determination module 230.

[0066] In some embodiments, the regional safety risk assessment model can be determined through an offline training manner. The training process of the regional safety risk assessment model can be referred to the description of the foregoing Figure 12The relevant descriptions will not be repeated here.

[0067] In some embodiments, the risk information for the target area may include the risk level of each road segment or road within the target area. In some embodiments, the risk level may include at least one of low risk, moderate risk, relatively high risk, and high risk.

[0068] In some embodiments, the processing device 112 may at least input risk characteristics into the regional security risk assessment model, and determine the risk information of the target area based on the output of the regional security risk assessment model.

[0069] In some embodiments, the processing device 112 can further determine collision information corresponding to multiple road segments based on road segment data; and determine risk information of a target area based on risk characteristics and collision information through a regional safety risk assessment model. Specifically, the processing device 112 can input risk characteristics and collision information into the regional safety risk assessment model, and determine the risk information of the target area based on the output of the regional safety risk assessment model.

[0070] In some embodiments, the output of the regional safety risk assessment model may include risk values ​​corresponding to multiple road segments and at least one important feature. Accordingly, the processing device 112 can determine a reference risk level for a road segment based on the at least one important feature and the threshold corresponding to that feature; and determine the risk level of the road segment based on the risk values ​​and the reference risk level. For more information on determining risk information for a target area based on the output of the regional safety risk assessment model, please refer to [link to relevant documentation]. Figure 4 and Figure 5 The relevant content will not be elaborated here.

[0071] In some embodiments, the processing device 112 can determine the risk levels corresponding to multiple road segments using a regional safety risk assessment model; and based on the risk levels corresponding to the multiple road segments, determine the road risk levels of roads within a target area. For more information on determining the road risk levels of roads within a target area based on the risk levels corresponding to multiple road segments, please refer to... Figure 7 The relevant content will not be elaborated here.

[0072] It should be noted that the above description of flow 300 is merely for example and illustration, and does not limit the scope of the present specification. Various modifications and changes can be made to flow 300 under the guidance of the present specification by those skilled in the art. However, these modifications and changes still fall within the scope of the present specification. For example, the time periods of peak, flat and night of the dynamic features in Table 2 can be divided into different time periods according to different target areas or different seasons. For example, the time period of peak can be 07:00-10:00 and 17:00-21:00; the time period of flat can be 06:00-07:00 and 10:00-17:00; the time period of night can be 21:00-06:00. Such variations are within the scope of the present application.

[0073] Figure 4 is a flowchart of an exemplary risk information determination method according to some embodiments of the present specification. In some embodiments, flow 400 can be performed by processing device 112 or area security risk assessment system 200. For example, flow 400 can be stored in the form of programs or instructions in a storage device (e.g., storage device 150, storage unit of processing device 112), and when a processor or Figure 2 the modules shown in the figure, flow 400 can be implemented. In some embodiments, flow 400 can utilize one or more additional operations not described below, and / or be completed without one or more of the operations discussed below. In addition, the order of the operations shown is not limiting. Figure 4 The order of the operations shown is not limiting.

[0074] At step 410, collision information corresponding to each of the plurality of road segments is determined based on the road segment data. In some embodiments, step 410 can be performed by determination module 230.

[0075] In some embodiments, collision information can embody information related to vehicle collisions. Taking a certain road segment as an example, the collision information of the road segment can embody the overall degree of collisions of vehicles passing through the road segment. In some embodiments, collision information can include a collision label and a collision level. In some embodiments, the collision label can embody whether a vehicle has collided (corresponding marks are 1 or 0). In some embodiments, the collision level can include minor collision, general collision, moderate collision and serious collision. In some embodiments, the collision level can be embodied in the form of text, numerical value, vector, matrix, picture, etc. For example, the collision level can be embodied by 0-1 or 0-100, etc. The larger the numerical value, the higher the collision level.

[0076] In some embodiments, taking a certain road segment as an example, the processing device 112 can determine the collision information based on the vehicle collision related data of the road segment. For example, taking a certain time period (e.g., 1 day) as an example, there are 10 collisions on the road segment in the time period, and 8 of them are serious collisions. Accordingly, the processing device 112 can determine that the collision label of the road segment is 1, and the collision level is 80 (calculated in the value interval of 0-100) or a serious collision.

[0077] Step 420, input the risk features and the collision information into the regional safety risk assessment model. In some embodiments, step 420 can be performed by the determination module 230.

[0078] In some embodiments, the processing device 112 can perform cleaning, vectorization, normalization, discretization, data transformation, etc. on the risk features and the collision information, and input the processed risk features and the collision information into the regional safety risk assessment model.

[0079] Step 430, determine the risk information of the target region based on the output of the regional safety risk assessment model. In some embodiments, step 430 can be performed by the determination module 230.

[0080] In some embodiments, the output of the regional safety risk assessment model can include a risk value corresponding to each road segment and at least one important feature.

[0081] In some embodiments, taking a certain road segment as an example, the risk value of the road segment can comprehensively reflect the risk level that can occur on the road segment. In some embodiments, the risk value can be in the form of text, numerical value, vector, matrix, picture, etc. For example, the risk value can reflect the risk level by any value in the similar numerical value interval of 0-1 or 1-100, and the larger the value, the higher the risk level.

[0082] In some embodiments, taking a certain road segment as an example, the important feature corresponding to the road segment can be a feature that has a greater impact on the risk situation of the road segment. In some embodiments, the important features corresponding to different road segments, different roads or different regions can be different. In some embodiments, the important features corresponding to different time or different time periods can be different. For example, taking a certain road segment or region as an example, as shown in FIG. 2, the at least one important feature can include traffic flow, vehicle speed, large vehicle proportion and motorcycle proportion. Figure 6

[0083] In some embodiments, each important feature can correspond to one or more threshold values. Taking a certain important feature as an example, the threshold value can reflect the risk critical situation of the important feature. Based on the important feature and the threshold value corresponding thereto, the reference risk situation of the road segment can be determined.

[0084] ​In some embodiments, taking a certain road segment as an example, the processing device 112 can determine the risk level of the road segment based on the risk value output by the regional safety risk assessment model and the reference risk situation corresponding to the at least one important feature. For more information about determining the risk level of the road segment based on the risk value and the reference risk situation, please refer to the relevant description of Figure 5 , which will not be repeated here.

[0085] It should be noted that the above description of the flow 400 is only for example and illustration, and does not limit the scope of the present specification. Those skilled in the art can make various modifications and changes to the flow 400 under the guidance of the present specification. However, these modifications and changes are still within the scope of the present specification. For example, the important features can also include traffic flow, vehicle speed, non-motor vehicle flow, time period, etc. or any combination thereof. For another example, the collision information can be extracted together with or as part of the risk features. Such variations are within the scope of the present application.

[0086] Figure 5 is a flowchart of an exemplary road segment risk level determination method according to some embodiments of the present specification. In some embodiments, the flow 500 can be performed by the processing device 112 or the regional safety risk assessment system 200. For example, the flow 500 can be stored in the form of programs or instructions in a storage device (for example, the storage device 150, the storage unit of the processing device 112), and when the processor or Figure 2 The modules shown in the figure execute the programs or instructions, the flow 500 can be implemented. In some embodiments, the flow 500 can utilize one or more additional operations not described below, and / or not be completed by one or more operations discussed below. In addition, the order of the operations shown in Figure 5 is not limiting.

[0087] In some embodiments, the risk level of each of the plurality of road segments can be determined. Specifically, it can include the following steps:

[0088] Step 510, determining the reference risk situation of the road segment based on the at least one important feature and the threshold value corresponding to the at least one important feature respectively. In some embodiments, step 510 can be performed by the determination module 230.

[0089] In some embodiments, taking a certain important feature as an example, the threshold value corresponding to the important feature can represent the risk critical situation of the important feature. For example, "traffic flow" can correspond to three threshold values A, B, C, where A > B > C. When the traffic flow is greater than A, the reference risk situation is high; when the traffic flow is between A and B, the reference risk situation is relatively high; when the traffic flow is between B and C, the reference risk situation is general; when the traffic flow is less than C, the reference risk situation is relatively low.

[0090] In some embodiments, the threshold value corresponding to the important feature can be a system default value (e.g., an empirical value), or can be dynamically adjusted for different situations. In some embodiments, the threshold value corresponding to different road sections, different roads, or different regions can be different. In some embodiments, the threshold value corresponding to different time instants or different time periods can be different.

[0091] In some embodiments, in determining the reference risk situation of a road section based on the important features and the threshold values corresponding thereto, expert opinions can also be combined.

[0092] In some embodiments, the expert opinions can be professional adjustment opinions of experts on the important features and / or the threshold values corresponding thereto. In some embodiments, taking a specific road section as an example, according to the expert opinions, the categories and / or quantities of the important features can be adjusted. In some embodiments, according to the expert opinions, the categories and / or quantities of the important features can be increased, decreased, or changed. For example, if the important feature of a specific road section is “large vehicle proportion”, according to the expert opinions, the important feature can be adjusted to “traffic flow” and / or “vehicle speed”. For another example, if the important features of a specific road section are “large vehicle proportion” and “motorcycle proportion”, according to the expert opinions, the important features can be adjusted to “vehicle speed”. For another example, if the important features of a specific road section are “large vehicle proportion” and “motorcycle proportion”, according to the expert opinions, the important features can be adjusted to “large vehicle proportion”, “traffic flow”, and “vehicle speed”. In some embodiments, taking a specific road section as an example, according to the expert opinions, the threshold values corresponding to the important features can be adjusted. For example, if the important feature of a specific road section is “large vehicle proportion”, the threshold value corresponding thereto is D, and when D is greater, the reference risk situation is high, and when D is smaller, the reference risk situation is low. According to the expert opinions, the specific value of the threshold value D can be adjusted in combination with the specific situation of the road section (e.g., vehicle driving data, road network data, etc.) or the specific situation of the region where the road section is located.

[0093] In step 520, the risk level of the road section is determined based on the risk value and the reference risk situation. In some embodiments, step 520 can be performed by the determination module 230.

[0094] In combination with the above Figure 3 or Figure 4 It is stated that the risk value of the road section is output by the regional safety risk assessment model. Based on the risk value output by the model, the final risk level of the road section can be determined in combination with the reference risk situation determined by the important features.

[0095] In some embodiments, the processing device 112 can determine the final risk level of the road segment based on the matching of the risk value and the reference risk situation. The matching can indicate whether the risk level embodied by the risk value is consistent or substantially consistent with the risk level embodied by the reference risk situation. For example, if the risk value indicates a high or higher risk level, and the reference risk situation indicates a low or general risk level, the two do not match. For another example, if the risk value indicates a high risk level, and the reference risk situation also indicates a high risk level, the two match.

[0096] In some embodiments, if the risk value matches the reference risk situation, the processing device 112 can determine the risk level of the road segment directly based on the risk value or the reference risk situation. For example, assuming the risk value is 30, and the reference risk situation is general, i.e., the risk value matches the reference risk situation, the processing device 112 can determine that the risk level of the road segment is general risk.

[0097] In some embodiments, if the risk value does not match the reference risk situation, the processing device 112 can determine the risk level of the road segment in combination with expert opinions. For example, if the risk value is 30, and the reference risk situation is high, i.e., the risk value does not match the reference risk situation, the final risk level of the road segment can be determined in combination with expert opinions.

[0098] By means of different important features and corresponding threshold values of different road segments, and in combination with expert opinions, the reliability of the regional safety risk assessment system can be improved, and the result bias caused by excessive reliance on the reliability of input data when determining the risk level can be eliminated to a certain extent.

[0099] It should be noted that the above description of the process 500 is merely for example and illustration, and does not limit the scope of the present specification. Various modifications and changes can be made to the process 500 under the guidance of the present specification. However, these modifications and changes are still within the scope of the present specification. For example, when adjusting the important features and / or the threshold values corresponding to the important features according to the expert opinions, it is not limited to the four listed in step 520, and other adjustment methods can also be included. Variations such as this are within the scope of protection of the present application.

[0100] Figure 7 is a flowchart of an exemplary road risk level determination method according to some embodiments of the present specification. In some embodiments, the process 700 can be performed by the processing device 112 or the regional safety risk assessment system 200. For example, the process 700 can be stored in the form of a program or instructions in a storage device (e.g., the storage device 150, the storage unit of the processing device 112), and when the processor or Figure 2The modules shown can implement the flow 700 when executing programs or instructions. In some embodiments, the flow 700 can utilize one or more additional operations not described below, and / or be completed without one or more of the operations discussed below. Additionally, the order of the operations shown is not limiting. Figure 7 The order of the operations shown is not limiting.

[0101] At step 710, the risk levels corresponding to the plurality of road segments respectively are determined based on at least the risk features by the regional safety risk assessment model. In some embodiments, step 710 can be performed by the determining module 230.

[0102] In some embodiments, the risk features and / or the collision information corresponding to the plurality of road segments respectively can be input into the regional safety risk assessment model, and the risk levels corresponding to the plurality of road segments respectively are determined. More details can be referred to the description of the regional safety risk assessment model, which is not repeated here. Figure 3-6

[0103] At step 720, the road risk level of the roads in the target region is determined based on the risk levels corresponding to the plurality of road segments respectively, wherein each road comprises at least one road segment. In some embodiments, step 720 can be performed by the determining module 230.

[0104] In some embodiments, the processing device 112 can comprehensively process the risk levels of the plurality of road segments included in a road to determine the road risk level of the road.

[0105] In some embodiments, the processing device 112 can determine the road risk level by the following formula (1):

[0106]

[0107] wherein R r represents the road risk level, represents the road segment risk level of the i-th road segment, represents the road segment length of the i-th road segment, and L represents the total length of the road.

[0108] In some embodiments, the processing device 112 can calculate the average or weighted average of the risk levels of the plurality of road segments to obtain the road risk level. In some embodiments, the processing device 112 can determine the road risk level by the following formula (2):

[0109]

[0110] wherein, ​where W represents the weight of the ith road segment, and N represents the number of road segments. In some embodiments, the weight of a road segment can be related to at least one of vehicle travel data, road network data, vehicle collision-related data, or environmental data of the road segment. For example, the longer the length of a road segment, the greater the weight of the road segment. For another example, the greater the collision level of a road segment, the greater the weight of the road segment. For yet another example, the darker the light or the more tree shade blocking of a road segment, the greater the weight of the road segment. For yet another example, the more the motor vehicle traffic or non-motor vehicle traffic of a road segment, the greater the weight of the road segment. In some embodiments, the weight of a road segment can be adjusted in combination with expert opinions.

[0111] It should be noted that the above description of the flow 700 is merely for example and illustration, and does not limit the scope of the present specification. Various modifications and changes can be made to the flow 700 by those skilled in the art under the guidance of the present specification. However, these modifications and changes still fall within the scope of the present specification. For example, the processing device 112 can also calculate a geometric mean value according to the risk level of at least one road segment in the road to obtain the road risk level. Variations such as this are within the scope of protection of the present application.

[0112] Figure 8 is a schematic diagram of an exemplary regional safety risk assessment process according to some embodiments of the present specification.

[0113] As Figure 8 shown, for a certain road segment, the processing device 112 can input the risk features and collision information into the regional safety risk assessment model, and the regional safety risk assessment model outputs the risk value and at least one important feature of the road segment. The processing device 112 can determine the risk level of the road segment based on the risk value and at least one important feature of the road segment in combination with expert opinions. Specifically, the processing device 112 can determine the reference risk situation of the road segment based on at least one important feature and the threshold value corresponding to at least one important feature, in combination with expert opinions. Further, the processing device 112 can determine the risk level of the road segment based on the matching of the risk value and the reference risk situation, in combination with expert opinions. In some embodiments, the road usually includes a plurality of road segments. The processing device 112 can determine the road risk level of each road in the target region based on the risk level of each road segment.

[0114] Figure 9 is a schematic diagram of an exemplary data acquisition and processing process according to some embodiments of the present specification.

[0115] In the present specification, the root of the regional security risk assessment model training is to obtain basic data by means of big data and the like. The basic data involves multiple dimensions, so that the model can realize comprehensive security risk assessment of the region. In order to ensure the reliability of the model training process, a series of processing needs to be performed on the obtained basic data to determine the sample road section data (training data set) used for model training.

[0116] In some embodiments, the processing device 112 can extract the corresponding basic data from a plurality of data warehouses (for example, a road network data data warehouse, a vehicle driving data data warehouse, a vehicle collision data data warehouse, an environment data data warehouse, a traffic participant data data warehouse, etc.).

[0117] In some embodiments, the processing device 112 can perform data fusion processing on the basic data in the plurality of data warehouses. In some embodiments, the data can be fusion processed in a road section as a basic unit. In some embodiments, the processing device 112 can divide the road to determine a plurality of road sections, and perform preliminary division processing on the basic data in units of road sections. In some embodiments, the processing device 112 can obtain the road section division result from the data source 140 and / or communicate with other interfaces to obtain the road section division result. In some embodiments, the basic data itself is stored in the form of road sections, so there is no need to perform road section division again.

[0118] In some embodiments, after determining the basic data in units of road sections, the processing device 112 can perform feature space construction to determine the training features used for subsequent model training. For more information about the original feature space construction, please refer to the related content of Figure 10 , which will not be repeated here.

[0119] In some embodiments, the processing device 112 can perform risk state labeling on the data. The risk state can reflect the collision risk of the road section. In some embodiments, the processing device 112 can perform risk state labeling based on the vehicle collision related data of the road section. For example, taking a certain time period as an example, 10 collisions occurred on the road section in the time period, and 8 of them were serious collisions. Accordingly, the processing device 112 can determine that the collision label of the road section is 1, and the collision level is 80 (calculated in the numerical interval of 0-100) or serious collision. For another example, taking a certain road section as an example, no collision has occurred on the road section. Accordingly, the processing device 112 can determine that the collision label of the road section is 0.

[0120] In some embodiments, after data fusion, the processing device 112 can divide the data into a dataset of collisions and a dataset of non-collisions, which will serve as sample datasets for subsequent model training. In some embodiments, the sample dataset may include multiple training samples. In some embodiments, each training sample may include sample risk features of the sample road segment and / or sample collision information of the sample road segment (e.g., sample collision label, sample collision level). In some embodiments, a training label corresponding to each training sample may also be determined. In some embodiments, the training label may include the sample risk value and sample importance features corresponding to the sample road segment. In some embodiments, the training label may be manually labeled or automatically labeled by the processing device 112. Further description of model training can be found here. Figure 12 The relevant content will not be elaborated here.

[0121] Figure 10 This is a flowchart illustrating an exemplary feature space determination method according to some embodiments of this specification; Figure 11 This is a schematic diagram illustrating an exemplary feature space determination method according to some embodiments of this specification. In some embodiments, process 1000 may be executed by processing device 112, regional security risk assessment system 200, or model training module. For example, process 1000 may be stored in a storage device (e.g., storage device 150, storage unit of processing device 112) in the form of a program or instructions, and executed by the processor, Figure 2 When the module or model training module shown executes the program or instructions, it can implement process 1000. In some embodiments, process 1000 may be completed using one or more additional operations not described below, and / or not through one or more operations discussed below. Additionally, as Figure 10 The order of operations shown is not restrictive. The following section combines... Figure 10 and Figure 11 The process of determining the feature space is described in detail.

[0122] Step 1010: Based on the sample road segment data, determine the sample static feature set and the sample dynamic feature set. In some embodiments, step 1010 may be performed by the extraction module 220 or the model training module.

[0123] In some embodiments, the sample road segment data can include sample data related to vehicle driving or road network on the road segment. In some embodiments, the sample road segment data can include at least one of sample vehicle driving data, sample road network data, or sample vehicle collision related data of the plurality of road segments. In some embodiments, the sample road segment data can further include sample environment data, sample traffic participant data, etc. of the plurality of road segments. In some embodiments, the processing device 112 can extract the respective sample road segment data from a plurality of warehouses (e.g., a road network data warehouse, a vehicle driving data warehouse, a vehicle collision data warehouse, an environment data warehouse, a traffic participant data warehouse, etc.). In some embodiments, the sample road segment data can be the same as or similar to the data types contained in the road segment data. For more on the sample road segment data, please refer to the related description of the road segment data in Figure 3 , which will not be repeated here.

[0124] In some embodiments, the sample static feature set can reflect the basic attributes of the plurality of road segments. In some embodiments, the sample static feature set can include a set of static features of the plurality of road segments. In some embodiments, the sample static feature set can be the same as or similar to the feature types contained in the static features. For more on the sample static feature set, please refer to the related description of the static features in Figure 3 , which will not be repeated here.

[0125] In some embodiments, as shown in Figure 11 , the processing device 112 can determine the sample static feature set based on the road network data in the sample road segment data, in combination with expert knowledge (e.g., static scene composition indicators, traffic participant behaviors, driving common sense, etc.). In some embodiments, in determining the sample static feature set, the processing device 112 can remove feature elements in the scene that are less relevant to the safety risk assessment (e.g., lawns, skies, etc. in the scene).

[0126] In some embodiments, the sample dynamic feature set can reflect the dynamic attributes of the plurality of road segments. In some embodiments, the sample dynamic feature set can include a set of dynamic features of the plurality of road segments. In some embodiments, the sample dynamic feature set can be the same as or similar to the feature types contained in the dynamic features. For more on the sample dynamic feature set, please refer to the related description of the dynamic features in Figure 3 , which will not be repeated here.

[0127] In some embodiments, as shown in Figure 11 , the processing device 112 can perform data integration on the sample vehicle driving data, sample vehicle collision related data, sample environment data, and / or sample traffic participant data in the sample road segment data to determine the sample dynamic feature set.

[0128] In some embodiments, as shown inFigure 11 As shown, according to the above processing, data containing the dynamic feature set and the static feature set can be obtained, and the construction of the original feature space is completed.

[0129] At step 1020, the feature processing model is used to determine the importance of each sample feature in the sample static feature set and the sample dynamic feature set. In some embodiments, step 1020 can be performed by the extraction module 220 or the model training module.

[0130] In some embodiments, the feature processing model can be determined by offline training. In some embodiments, an initial feature processing model can be trained based on historical data samples to obtain a trained feature processing model. In some embodiments, the historical data samples can include historical sample features, historical importance corresponding to the historical sample features, and the like. In some embodiments, the historical sample features can be used as the input of the initial feature processing model, and the historical importance corresponding to the historical sample features can be used as the label of the initial feature processing model to train the feature processing model.

[0131] In some embodiments, the feature processing model can include a Random Forests (RF) model, a Gradient Boosting Decision Tree (GBDT) model, an Extreme Gradient Boosting (XGBoost) model, an Extremely Randomized Trees (ET) model, and an Adaptive Boosting (Adaboost) model. In some embodiments, the feature processing model can be implemented by correlation analysis and Gaussian mapping.

[0132] In some embodiments, the importance of the sample features can represent the degree of influence of the sample features on the risk level of the road section. In some embodiments, the importance of the sample features can be represented in the form of text, numerical value, vector, matrix, picture, and the like. For example, the importance can represent the degree of influence on the risk level by any value in a numerical value interval such as 0-1 or 1-100, and the greater the numerical value, the higher the degree of influence.

[0133] In some embodiments, the processing device 112 can input the sample features in the sample static feature set and the sample dynamic feature set into the feature processing model, and determine the importance of each sample feature based on the output of the feature processing model.

[0134] At step 1030, sample features with importance satisfying a preset requirement are screened to determine the feature space. In some embodiments, step 1030 can be performed by the extraction module 220 or the model training module.

[0135] In some embodiments, the preset requirement can be a requirement related to importance. In some embodiments, the preset requirement can include an importance threshold. In some embodiments, the importance threshold can be preset in advance. For example, the importance threshold can be 50, 80, 90 (calculated in the numerical interval of 1-100), etc.

[0136] In some embodiments, the preset requirement can be a requirement related to importance ranking. In some embodiments, the preset requirement can include that the importance ranking is in the top E positions. In some embodiments, the value of E can be preset in advance. For example, the importance ranking in the top E positions can be the importance ranking in the top 100 positions, 1000 positions, 10000 positions, etc.

[0137] In some embodiments, the processing device 112 can screen sample features with importance greater than or equal to the importance threshold, and the set of screened sample features constitutes the feature space. For example, the importance threshold is 80, and accordingly, the processing device 112 can screen sample features with importance greater than or equal to 80, and construct the set of screened sample features into the feature space.

[0138] In some embodiments, the processing device 112 can screen sample features with importance ranking in the top E positions, and the set of screened sample features constitutes the feature space. For example, the value of E is 1000, and accordingly, the processing device 112 can screen sample features with importance ranking in the top 1000 positions, and construct the set of screened sample features into the feature space.

[0139] It should be noted that the above description of the flow 1000 is merely for example and illustration, and does not limit the scope of the present specification. Those skilled in the art can make various modifications and changes to the flow 1000 under the guidance of the present specification. However, these modifications and changes are still within the scope of the present specification. For example, the importance threshold is not limited to that listed in step 1030, and the importance threshold can also be 0.6, 0.7, 0.8, or 0.9 (calculated in the numerical interval of 0-1). Such variations are within the scope of protection of the present application.

[0140] Figure 12 is a schematic diagram of an exemplary regional security risk assessment model training method according to some embodiments of the present specification.

[0141] In some embodiments, the regional safety risk assessment model 1210 can be a Convolutional Neural Network (CNN), a Deep Neural Network (DNN), a Recurrent Neural Network (RNN), a Graph Neural Network (GNN), a Generative Adversarial Network (GAN), or the like, or any combination thereof.

[0142] In some embodiments, the regional safety risk assessment model 1210 can be determined based on a plurality of training samples 1220. In some embodiments, the plurality of training samples 1220 can include training samples corresponding to different road segments, different roads, and / or different regions. In some embodiments, the plurality of training samples 1220 can include training samples corresponding to different time instants and / or different time periods.

[0143] In some embodiments, each training sample 1220 can include sample risk features 1221 of a sample road segment, a corresponding sample risk value 1223, and sample importance features 1224, where the sample risk features 1221 are training data, and the corresponding sample risk value 1223 and the sample importance features 1224 are training labels.

[0144] In some embodiments, each training sample 1220 can include sample risk features 1221 of a sample road segment, sample collision information 1222 of the sample road segment, a corresponding sample risk value 1223, and sample importance features 1224, where the sample risk features 1221 and the sample collision information 1222 are training data, and the corresponding sample risk value 1223 and the sample importance features 1224 are training labels.

[0145] In some embodiments, the sample risk features 1221 are similar to the risk features, and more specific descriptions can be found in Figure 3 ; the sample collision information 1222 is similar to the collision information, and more specific descriptions can be found in Figure 4 .

[0146] In some embodiments, the sample risk value 1223 and the sample importance features 1224 can be manually labeled by a user or automatically labeled by the regional safety risk assessment system 100.

[0147] In some embodiments, the regional safety risk assessment model 1210 can include two sub-models for determining the risk value and the importance features of a road segment, respectively. In some embodiments, the two sub-models can be jointly trained or independently trained.

[0148] In some embodiments, the regional safety risk assessment model 1210 can also be implemented by a general model.

[0149] The training process of the regional safety risk assessment model 1210 is described below by taking a general model as an example: taking the sample risk features 1221 and / or the sample collision information 1222 as input, taking the corresponding sample risk values 1223 and sample important features 1224 as supervision, training the regional safety risk assessment model 1210, updating the parameters of the regional safety risk assessment model 1210 through a machine learning algorithm (for example, a stochastic gradient descent method), and minimizing the loss function until the model training is completed; or stopping training after a certain number of iteration training times.

[0150] In some embodiments, the loss function can be a perception loss function. In some embodiments, the loss function can also be other loss functions, such as a square loss function, a logistic regression loss function, etc.

[0151] In some embodiments, the road section data can be dynamically updated (for example, real-time update, periodic update, etc.). Accordingly, the plurality of training samples 1220 can also be dynamically updated (for example, every hour, every day, every week, every month). In turn, the parameters of the regional safety risk assessment model 1210 can be dynamically updated based on the updated plurality of training samples 1220. By dynamically updating the training samples and thus the regional safety risk assessment model, the comprehensive learning ability of the regional safety risk assessment model can be improved, and the reliability of the regional safety risk assessment system can be improved.

[0152] The beneficial effects that can be brought by the embodiments of the present application include but are not limited to: (1) collecting vehicle driving data, road network data, vehicle collision related data, environmental data, traffic participant data, etc. road section data in a big data manner, improving the reliability of the model from the data acquisition method and the data order; (2) real-time, continuous and dynamic updating of data, so that the model can be dynamically and periodically updated, thereby dynamically evaluating the regional safety risk; (3) automatic big data acquisition method, so that the model training data sources are extensive, ensuring the scalability of the model for multiple locations; (4) combination of dynamic and static features to meet the requirements of various test scenarios; (5) introducing expert opinions in the process of data processing and risk assessment, further improving the reliability of the regional safety risk assessment system; (6) determining the risk level of the road section based on the road section as the basic unit, and then determining the risk level of the road and the target region, improving the universality and scalability of the regional safety risk assessment system. It should be noted that different embodiments can have different beneficial effects, and in different embodiments, the beneficial effects that can be produced can be any one or a combination of several of the above, or any other beneficial effects that can be obtained.

[0153] The foregoing detailed description has set forth various embodiments of the application via the use of specific terminology. As such, it is to be understood that whenever a particular embodiment is described, that embodiment is intended to serve as a representative example, and is not intended to limit the scope of the disclosure. As described, modifications, improvements and variations of the specific examples described herein can occur to those skilled in the art. Such modifications, improvements and variations are intended to be within the spirit and scope of the examples described herein, and are intended to be encompassed by the claims.

[0154] In addition, the use of "one embodiment," "an embodiment," or "some embodiments" throughout this specification is not a limitation on the scope of the disclosure. Rather, these phrases are used to describe a particular embodiment, and do not mean the same embodiment unless otherwise indicated. Furthermore, the use of the term "a" or "an" to describe a particular element is not meant to be limiting. Thus, the terms "a" or "an" are intended to mean "one or more" unless otherwise indicated.

[0155] In addition, the order of presentation of the method steps of the present disclosure is not meant to be limiting. The order of presentation of the steps of the method can be different unless otherwise indicated. Furthermore, the use of numbering or letters in the claims to identify various elements in the description is not meant to be limiting. The use of such numbering or letters is simply to help identify various elements of the disclosure.

[0156] Similarly, it is to be noticed that the term "comprising", used in the description, is not meant to be construed as a limitation on the scope of the disclosure. Rather, it is to be understood that the term "comprising" is intended to cover the embodiments of the disclosure. In addition, the use of the term "comprising" is not meant to be construed as a statistical requirement. That is, the use of the term "comprising" is not meant to require that every element listed in the claim be present in the example of the disclosure.

[0157] In some embodiments, numbers describing the quantity of components and attributes are used. It should be understood that such numbers used in the description of embodiments are modified in some examples with the terms "approximately," "approximately," or "generally." Unless otherwise stated, "approximately," "approximately," or "generally" indicates that the numbers are allowed to vary by ±20%. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximate values, which may be changed depending on the characteristics required by individual embodiments. In some embodiments, numerical parameters should take into account specified significant digits and employ a general method of digit reservation. Although the numerical ranges and parameters used to confirm their breadth of range in some embodiments of this specification are approximate values, in specific embodiments, such values ​​are set as precisely as feasible.

[0158] For each patent, patent application, patent application publication, and other material, such as articles, books, specifications, publications, and documents, referenced in this specification, the entire contents of which are incorporated herein by reference. This excludes historical application documents that are inconsistent with or conflict with the content of this specification, as well as documents that limit the broadest scope of the claims in this specification (currently or subsequently appended to this specification). It should be noted that in the event of any inconsistency or conflict between the descriptions, definitions, and / or terminology used in the supplementary materials to this specification and the content of this specification, the descriptions, definitions, and / or terminology used in this specification shall prevail.

[0159] Finally, it should be understood that the embodiments described in this specification are merely illustrative of the principles of the embodiments described herein. Other variations may also fall within the scope of this specification. Therefore, alternative configurations of the embodiments described herein are intended to be illustrative rather than limiting, and should be considered consistent with the teachings of this specification. Accordingly, the embodiments described herein are not limited to those explicitly introduced and described herein.

Claims

1. A method for regional security risk assessment, characterized in that, The method includes: Obtain road segment data for multiple road segments within a target area, which is associated with the autonomous driving service; Based on the road segment data, risk features are extracted, including static features and dynamic features; Based at least on the aforementioned risk characteristics, a regional safety risk assessment model is used to determine the risk information of the target area. The risk information of the target area indicates the road risk level of target roads within the target area. The target roads include multiple target road segments, and the road risk level is determined based on multiple road segment risk levels corresponding to the multiple target road segments. The multiple road segment risk levels are determined based on risk values ​​and reference risk conditions corresponding to the multiple target road segments respectively. The reference risk conditions are determined based on at least one important feature of each target road segment and a threshold corresponding to the at least one important feature. The output of the regional safety risk assessment model includes the risk values ​​and at least one important feature corresponding to the multiple target road segments respectively.

2. The method as described in claim 1, characterized in that, The road segment data includes at least one of the following: vehicle driving data, road network data, or vehicle collision-related data for the multiple road segments.

3. The method as described in claim 1, characterized in that, The risk characteristics are determined through a feature space, which is determined in the following way: Based on the sample road segment data, determine the sample static feature set and the sample dynamic feature set; The importance of sample features in the static feature set and the dynamic feature set is determined by the feature processing model. The feature space is determined by selecting sample features whose importance meets preset requirements.

4. The method as described in claim 1, characterized in that, The determination of risk information for the target area based at least on the risk characteristics, using a regional security risk assessment model, includes: Based on the road segment data, the collision information corresponding to each of the multiple road segments is determined; The risk characteristics and collision information are input into the regional safety risk assessment model; Based on the output of the regional security risk assessment model, the risk information of the target area is determined.

5. The method as described in claim 1, characterized in that, The determination of the reference risk status of the road segment based on the at least one important feature and the threshold corresponding to the at least one important feature includes: Based on the at least one important feature and the threshold corresponding to the at least one important feature, and in conjunction with expert opinions, the reference risk status of the road segment is determined.

6. The method as described in claim 1, characterized in that, Determining the risk level of the road segment based on the risk value and the reference risk situation includes: Based on the risk value and the reference risk situation, and in conjunction with expert opinions, the risk level of the road segment is determined.

7. A regional security risk assessment system, characterized in that, The system includes: The acquisition module is used to acquire road segment data of multiple road segments within a target area, wherein the target area is associated with the autonomous driving service; The extraction module is used to extract risk features based on the road segment data, the risk features including static features and dynamic features; A determination module is configured to determine risk information of a target area based at least on the risk characteristics using a regional safety risk assessment model. The risk information of the target area indicates the road risk level of target roads within the target area. The target roads include multiple target road segments, and the road risk level is determined based on multiple road segment risk levels corresponding to the multiple target road segments. The multiple road segment risk levels are determined based on risk values ​​and reference risk conditions corresponding to the multiple target road segments, respectively. The reference risk conditions are determined based on at least one important feature of each target road segment and a threshold corresponding to the at least one important feature. The output of the regional safety risk assessment model includes the risk values ​​and at least one important feature corresponding to the multiple target road segments.

8. A regional security risk assessment device, characterized in that, The device includes: At least one storage medium that stores computer instructions; At least one processor executes the computer instructions to implement the method of any one of claims 1 to 6.

9. A computer-readable storage medium storing computer instructions that, when read by a computer, execute the method as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by the processor, they implement the method of any one of claims 1 to 6.

Images