Vehicle-road cooperative data verification method, device and storage medium for complex roads

Abstract

The present invention relates to the field of communication technology between intelligent networked vehicles and infrastructure, and discloses a vehicle-road collaborative data verification method for complex roads, including: step one, obtaining the full amount of true value data preset by the road infrastructure on the target intersection / section; step two, constructing a simplified acquisition model, and classifying the simplified acquisition models for all intersections / sections, each type of simplified acquisition model has a corresponding simplified optimization acquisition rule; step three, selecting a correspondingly classified simplified acquisition model according to the geographical features of the target intersection / section, and screening the full amount of true value data of the target intersection / section according to the simplified optimization acquisition rule to obtain the simplified true value data; step four, forming simplified test data according to the target intersection passing direction selected by the simplified optimization acquisition rule; automatically comparing the simplified test data and the simplified true value data, and generating and outputting an inspection report. The present invention can effectively improve the verification efficiency under the premise of ensuring the accuracy of verification.

Description

Technical Field

[0001] The present invention relates to the field of communication technology between intelligent connected vehicles and infrastructure, and in particular to a vehicle-road collaborative data verification method, device and storage medium for complex roads. Background Art

[0002] Intelligent and Connected Vehicle (ICV) is a high-tech product that integrates modern communication and network technologies, artificial intelligence, vehicle engineering and other fields. It can exchange information through the in-vehicle network and the external network (such as mobile communication network, Internet, etc.), and has certain functions such as environmental perception, intelligent decision-making and collaborative control.

[0003] Driven by policy guidance and market demand, the production and sales scale of intelligent connected vehicles continues to expand. Especially in the field of new energy vehicles, intelligent networking has become an important feature.

[0004] Since there are always blind spots in the autonomous driving of a single connected car, the perception of the autonomous car is insufficient. Now, in order to solve this problem, it is usually compensated by using roadside perception data, which is often called vehicle-road collaboration.

[0005] However, because the sources of roadside perception data are diverse and directly related to the installation location and transmission speed of specific road infrastructure (such as traffic lights, RSI traffic signs, roadside units RSU, roadside equipment, etc.), in order to ensure the accuracy of the roadside perception data used by intelligent connected vehicles, these data need to be verified in advance. Especially for the operation and maintenance of networked intersections in complex road environments, ensuring the accuracy of roadside data is very important in the vehicle-road collaborative system. For example, how to accurately and efficiently test and monitor the accuracy of maintenance signals has always been the focus of the vehicle-road collaborative test and maintenance team. Signal lights are characterized by a large number, different signal light groups, different countdowns, and different directions, which makes testing and maintenance very difficult.

[0006] In addition, since traffic lights now transmit C-V2X data, the long transmission link can easily cause signal light SPAT message delays, making the signal light SPAT message received by smart connected cars different from the actual one. Therefore, there is no way to use signal light SPAT messages in a timely and direct manner, and the corresponding roadside perception data cannot be effectively used, so the problem of insufficient single-vehicle perception of smart connected cars has never been effectively solved.

[0007] For this reason, it is common to conduct preliminary verification of the vehicle-road cooperative data obtained by the vehicle-road cooperative perception system in order to improve the effectiveness of the roadside perception data. However, most of the current verifications are completed on ideal roads in experimental scenarios, and the practicality of the verification results is not high. If the verification is carried out in an open scenario, that is, in an actual road environment, due to the complexity of the actual road environment, in order to ensure the accuracy of the verification results, only one target intersection or one target section can be verified one by one. Not only does it have many operating steps and low efficiency, but it also has a huge amount of data processing, and it is impossible to obtain verification results in a timely manner. It is not suitable for the current reality of rapid changes in the road environment. Summary of the invention

[0008] The present invention aims to provide a method, device and storage medium for vehicle-road collaborative data verification for complex roads, so as to solve the problem of low efficiency of vehicle-road collaborative data verification for current intelligent connected vehicles.

[0009] To solve the above problems, the present invention adopts the following technical solutions:

[0010] Solution 1: A vehicle-road cooperative data verification method for complex roads, including the following steps:

[0011] Step 1: Obtain the full amount of true value data of the road infrastructure preset at the target intersection / road section;

[0012] Step 2: construct a simplified collection model and classify the simplified collection models for all intersections / road sections. Each type of simplified collection model has corresponding simplified optimization collection rules.

[0013] Step 3: Select a simplified collection model of the corresponding classification according to the geographical features of the target intersection / road section, and screen the full amount of true value data of the target intersection / road section according to the simplified optimization collection rules to obtain the simplified true value data;

[0014] Step 4: Select the target intersection direction according to the streamlined and optimized collection rules, and the inspection vehicle connected to the vehicle-road cooperative perception system collects the infrastructure operation data at the target intersection / road section to form streamlined test data;

[0015] Step 5: Automatically compare the simplified test data and the simplified true value data through the preset automatic verification strategy, generate and output the inspection report;

[0016] The simplified test data includes simplified test data of traffic lights and simplified test data of signboards; the simplified true value data includes true value data of traffic lights and true value data of signboards.

[0017] The advantages of this solution are:

[0018] By building a streamlined collection model, the complex target intersection / road section environment in the actual road environment can be quickly classified according to its terrain characteristics, and a suitable streamlined collection model can be selected for data collection and verification judgment. This can greatly reduce the processing steps and effectively improve the verification efficiency.

[0019] In addition, by streamlining the collected test data, the amount of streamlined test data and streamlined true value data is greatly reduced while saving comparison steps, which reduces the amount of data processing and calculation, and further improves the verification efficiency.

[0020] Preferably, the target intersection / road section includes a target intersection section and a target road section; the target intersection section refers to the target intersection and the connecting road section connected thereto, a target area with a diameter of 50-320 meters with the center point of the target intersection as the center of the circle; the target road section is a target road section area with a length of less than 300 meters excluding the intersection.

[0021] By limiting the target intersection / road section area, data collection and comparison can be completed more quickly and accurately.

[0022] Preferably, the road infrastructure includes traffic lights, RSI signs, roadside equipment and roadside units; the inspection vehicle is provided with an on-board communication unit; the on-board communication unit communicates with the roadside equipment and the roadside unit respectively to obtain traffic light test data reflecting the operation of the traffic lights, and signboard test data reflecting the operation of the RSI signs.

[0023] Preferably, the simplified signal light test data is signal light test data obtained by screening signal light test data according to the simplified optimization collection rule;

[0024] The simplified signboard test data is the signboard test data collected by the inspection vehicle according to the simplified optimization collection rules for the target intersection section or the limited area of ​​the target section.

[0025] Road infrastructure generally refers to various existing functional equipment including traffic lights and / or roadside real-time traffic information (RSI), as well as roadside equipment and roadside units used to access the vehicle-road collaborative perception system and serve intelligent connected vehicles. Roadside real-time traffic information (RSI) includes traffic event messages and traffic sign information issued by the roadside unit to the surrounding vehicle-mounted units. It is a combination of sensors installed next to the road infrastructure to detect the operation of the road infrastructure. Among them, RSI signs belong to roadside real-time traffic information (RSI). Roadside perception data contains a lot of content. This solution mainly verifies the corresponding roadside perception data of traffic lights and RSI signs.

[0026] Preferably, the simplified and optimized collection rule is to select the direction where the traffic rules conflict to collect the signal light test data perceived by the vehicle-road cooperative system to form the signal light test data. The amount of signal light test data can be reduced by limiting the collection area.

[0027] Preferably, the simplified signal light test data is obtained by screening the signal light test data in a ratio of 50%-75%; the simplified signal light test data is not the full data, so as to further reduce the amount of data that needs to be compared and verified.

[0028] Preferably, when the target intersection is a crossroads with four directions, two directions of east to west / west to east and south to north / north to south are collected; when the target intersection is a T-shaped, Y-shaped or herringbone intersection with three directions, two randomly selected directions are collected.

[0029] Preferably, the streamlined optimization collection rules include: the inspection vehicle collects the signal light test data of the direction of the intersection that the current driving route will pass, and the signal light test data that may be passed after the route is changed as the signal light simplified test data; the inspection vehicle collects the signboard test data of the direction of the intersection that the current driving route will pass, and the signboard test data that may be passed after the route is changed as the signboard simplified test data; the signal light simplified test data and the signboard simplified test data are all full data. Starting from the inspection route of the inspection, a limited area is selected for test data collection to form simplified test data, which not only achieves the purpose of reducing the amount of data, but also can cooperate with the driving trajectory of the inspection vehicle, making the collection and verification operation more convenient.

[0030] Solution 2: The present invention also provides a vehicle-road cooperative data verification device for complex roads, which is used to implement the aforementioned vehicle-road cooperative data verification method for complex roads, including a central processor connected to the vehicle-road cooperative perception system, and an on-board communication unit and an automatic verification module respectively connected to the central processor for communication;

[0031] The central processor is preset with a plurality of simplified acquisition models, each of which includes at least one simplified optimization acquisition rule;

[0032] The vehicle-mounted communication unit is arranged on the inspection vehicle, and the central processor selects the corresponding simplified collection model according to the terrain characteristics of the target intersection / road section and sends it to the inspection vehicle; the inspection vehicle collects the operation data of the adjacent road infrastructure along the driving track of the inspection vehicle according to the corresponding simplified optimization collection rules, processes it into simplified test data and sends it to the automated verification module;

[0033] The central processor processes the preset true value data of the corresponding target intersection / road section according to the simplified optimization collection rules, obtains the simplified true value data, and sends it to the automated verification module;

[0034] The automated verification module is used to automatically compare the simplified test data with the simplified true value data, determine whether the two are consistent, and send the determination result to the central processing unit;

[0035] The central processing unit generates and outputs an inspection report based on the judgment result.

[0036] Solution three: The present invention also provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, it implements the aforementioned vehicle-road collaborative data verification method for complex roads.

[0037] The advantages of the present invention are:

[0038] First, by streamlining the collection model, all intersections / road sections are classified according to the different terrain features of the intersections / road sections. Intersections / road sections of the same type can be tested in the same way to form a unified test method and unified judgment criteria, which can not only effectively reduce the workload of test verification, but also make the test results more credible.

[0039] By streamlining and optimizing the collection rules, the preset true value data is streamlined to obtain streamlined true value data; at the same time, by streamlining and optimizing the collection rules, the collection test area is narrowed and limited, and the roadside data reflecting the operation of road infrastructure collected from the roadside units by patrol vehicles in these limited areas are used as streamlined test data; and the consistency of the two is determined through automatic comparison.

[0040] The present invention reduces the amount of comparison data, making the comparison of simplified true value data and simplified test data faster, and because of the setting of simplified optimization rules, it can ensure the accuracy of the verification judgment results while ensuring automatic and rapid comparison and judgment. The accuracy and timeliness of the data transmitted by the roadside unit are judged by the judgment results, as well as whether the road infrastructure is operating normally, and an inspection report is output.

[0041] The present invention can not only simultaneously verify whether the road infrastructure is operating normally and whether the roadside unit and the inspection vehicle are communicating normally, but also verify whether the detection information of the roadside unit is accurate or has a delay. It can effectively reduce the amount of calculation and workload while ensuring the accuracy of verification. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 This is a flow chart of Embodiment 1 of the present invention.

[0043] Figure 2 This is a schematic diagram for comparing the simplified true value data of a traffic light and the simplified test data of a traffic light in the first embodiment.

[0044] Figure 3This is a schematic diagram of the screening area for collecting target intersections / road sections according to the simplified and optimized collection rules in Example 1.

[0045] Figure 4 This is a schematic diagram of comparing the true value data of the signboard and the test data of the signboard in Example 1. DETAILED DESCRIPTION

[0046] The following is further described in detail through specific implementation methods:

[0047] Embodiment 1

[0048] Due to the physical structure limitations of smart connected vehicles, there are always blind spots when driving autonomously. There is no way to accurately perceive the vehicle's operating conditions, including vehicle conditions and road conditions, and thus ensure the smooth progress of autonomous driving only through the various vehicle-side devices on smart connected vehicles.

[0049] To this end, the mainstream research direction in the industry is to obtain roadside perception data by adding roadside units (RSUs) and roadside equipment on road infrastructure, and to make up for and solve the problem of insufficient perception of single-vehicle intelligent networked vehicles through vehicle-side data and real-time received and confirmed roadside perception data, so that intelligent networked vehicles can accurately perceive the vehicle's operating status in real time, and then successfully complete the automatic driving of the corresponding road section. This is the essence of the current general vehicle-road collaboration.

[0050] Intelligent connected vehicles, through various vehicle-side devices installed on their own vehicles, receive and reasonably use the roadside perception data obtained from roadside units installed on various road infrastructures and roadside units installed on roadside equipment in real time, plus the vehicle-side data collected by the vehicle-side devices themselves, to finally form comprehensive vehicle-side perception data.

[0051] However, due to the limitations of complex urban roads, the installation of roadside equipment and roadside units in different sections and intersections is quite different. The installation of roadside equipment and roadside units directly affects the collection and application of roadside perception data. At different intersections / sections, the quality of collected roadside perception data cannot be unified, which greatly reduces the quality of vehicle-road collaboration. More importantly, some roadside equipment, such as traffic lights, use C-V2X data for data transmission. Due to the long transmission link, the signal light SPAT (Signal Phase and Timing Message) message delay causes the signal light SPAT message received by the vehicle to be different from the actual one; causing the vehicle to distrust this part of the roadside perception data, further reducing the credibility of the roadside perception data.

[0052] In order to improve the credibility of roadside perception data, it is necessary to verify the roadside perception data. At present, for different intersections / road sections, the roadside perception data sent by the roadside unit received by the vehicle is compared with the actual data of changes in the actual intersection / road section signal lights found through observation and comparison to verify whether the roadside perception data is accurate. This operation has great disadvantages. There is no unified deployment mode for the installation of roadside equipment and roadside units at each intersection / road section. It is time-consuming and labor-intensive to verify one by one. In addition, in order to make more accurate judgments, the inspection vehicle needs to stop and go back and forth at the designated intersection / road section. And the method of obtaining actual data is very rough. Basically, the actual signal light data representing the changes in the signal light and the actual RSI sign data representing the changes in the sign are obtained through human eye observation. Therefore, the verification of vehicle-road collaborative data, especially the verification of roadside data, is cumbersome and has a large gap with actual driving applications.

[0053] In addition, since the test is conducted on an actual open road, a closed loop cannot be formed, which increases the difficulty of automated testing and requires manual testing and comparison. As a result, the efficiency is low, the possibility of human misjudgment during testing is high, the acceptance cost after large-scale deployment is high, and the maintenance after acceptance is difficult.

[0054] To this end, this solution proposes a vehicle-road collaborative data verification method for complex roads to solve the above problems.

[0055] like Figure 1 As shown, the vehicle-road cooperative data verification method for complex roads in this embodiment includes the following contents:

[0056] The first step is to obtain the preset true value data of the target intersection / road section in advance, and the true value data includes the true value data of the traffic light and RSI.

[0057] Among them, RSI data (Road Side Information) is road side information, including traffic event messages and traffic sign information issued by the road side unit to surrounding vehicle-mounted units.

[0058] Before maintaining the vehicle-road collaborative traffic light data, the traffic light and RSI data designed in the original design parameter table, or the full data of the traffic lights at the target intersection / section are obtained through WIFI as the true value data of the traffic lights, including the light color, countdown, cycle, and UTC time of the traffic lights in each direction.

[0059] Through the existing RSI sign collection tools, the location of the RSI signs at the target intersection / section is collected through the positioning equipment, and the name and type of the sign are recorded to form the RSI sign true value data.

[0060] Specifically, in this embodiment, before operation and maintenance, the RSI sign true value data is formed according to the true value data stored in the original design content, that is, according to the sign position, sign name and type marked in the intersection map information, and the signal light true value data is obtained according to the signal light control content. The acquisition of true value data is an existing technology, and other existing means can be used to obtain and replace it.

[0061] An on-board communication unit is installed on the inspection vehicle to receive the actual traffic light and RSI data perceived by the vehicle-road cooperative system at the inspection target intersection / road section as the corresponding test data. That is, the inspection vehicle equipped with the on-board communication unit obtains the actual traffic light data and actual RSI data from the roadside equipment installed on both sides of the corresponding road through the on-board communication unit when driving in the target intersection, target road section or the entire area including the target road section and the target intersection.

[0062] The actual RSI test data includes RSI sign test data, including the name, type and location of RSI.

[0063] This solution can obtain true value data from the original design parameter table at the design source, which can ensure that the real data is completely correct and not distorted, making it easier to compare and judge the vehicle-side data collected by the inspection vehicle and the comprehensive vehicle perception data integrated with the roadside perception data, providing a real comparison basis for data verification, making the comparison results more real and reliable.

[0064] The second step is to build a simplified collection model for the target intersection / road section; and classify the simplified collection models of all intersections / road sections.

[0065] The specific classification refers to the terrain features, such as the shape of the specified intersection or the length and slope of the road section. The classification category can be determined based on the actual situation, which will not be expanded here.

[0066] Abstract the simplified collection model of target intersections including crossroads and T-junctions and corresponding road sections. In this way, all intersections / road sections can be classified into models. When verifying, the verification method and steps of the same model can be unified. There is no need to test and verify each direction of the target intersection / road section one by one. As long as the road section area between two sets of traffic lights is selected, that is, one direction of the inspection car passing through the intersection and the corresponding area are selected for collection and verification, the workload can be greatly reduced without affecting the accuracy of verification.

[0067] Among them, the streamlined model is carried out according to the preset streamlined optimization collection rules, and the streamlined optimization collection rules are classified according to the geographical characteristics of the target intersection / road section, especially the intersection shape is prioritized for judgment and classification.

[0068] The third step is to select a simplified collection model of the corresponding classification according to the geographical features of the target intersection / road section, and filter the full true value data of the target intersection / road section according to the simplified optimization collection rules of the simplified collection model to obtain the simplified true value data.

[0069] In the fourth step, according to the target intersection passing direction selected by the streamlined optimization collection rules, the inspection vehicle collects test data transmitted from the roadside units on the infrastructure operation and roadside equipment at the target intersection and section to form streamlined test data; through the preset automatic verification strategy, the streamlined test data and the streamlined true value data are automatically compared to generate and output the inspection report.

[0070] Among them, according to the preset streamlined optimization collection rules, the full amount of true value data is streamlined and screened to obtain streamlined test data, and according to the streamlined optimization collection rules, the test data collected by the inspection vehicle on the roadside equipment and roadside units in the specific areas and directions specified by the streamlined optimization collection rules are used as streamlined test data.

[0071] The simplified test data includes simplified test data of traffic lights and simplified test data of signboards; the simplified true value data includes true value data of traffic lights and true value data of signboards.

[0072] The simplified true value data and the corresponding simplified test data in this embodiment are as follows: Figure 2 The simplified true value data is used to characterize the theoretical operation of the target intersection / road section and the road infrastructure set up in this area. The simplified test data is used to characterize the roadside perception data corresponding to the traffic lights and signboards in the screening area obtained in real time by the patrol vehicle from the roadside equipment of the target intersection / road section. The roadside perception data received by the patrol vehicle is used as the test data, and is highly extracted and simplified after passing through the specified collection area to form the simplified test data.

[0073] By streamlining and optimizing the collection rules, according to the terrain features of the target intersection / road section, the streamlined collection model classification corresponding to the target intersection / road section is selected to extract and screen out streamlined true value data and streamlined test data from the true value data and test data respectively, so that a set of streamlined data formed by the streamlined true value data and the streamlined test data can not only basically represent the operation status of the road infrastructure of the target intersection / road section, but also effectively reduce the amount of data, so that when performing comparison and verification, the comparison between the streamlined test data and the streamlined true value data can be completed more quickly and accurately, and on the premise of ensuring the accuracy of the comparison results, the amount of calculation is effectively reduced and the efficiency of comparison and verification is improved.

[0074] When forming streamlined test data, the operation and maintenance personnel drive the inspection vehicle, classify it according to the streamlined collection model, drive the inspection vehicle to the target intersection and select the direction, receive the traffic light data and RSI data perceived by the vehicle-road cooperative system on the inspection vehicle, and form the corresponding streamlined test data.

[0075] like Figure 3 As shown, in this embodiment, the simplified collection model classification selected according to the terrain characteristics of the target intersection / road section, and its simplified optimization collection rules are specifically as follows: the collection area is demarcated according to the target intersection / road section, and the road section area radiating outward within a certain distance with the target intersection as the center belongs to the target intersection / road section, and the distance radiating outward can be determined according to actual needs. Generally, the radiating road section within 50-320 meters from the center of the target intersection can be counted as the target road section corresponding to the target intersection. In this embodiment, the target intersection / road section refers to the location area containing the intersection and / or road section within the range of 300 meters with the center point of the target intersection as the center of the circle. Among them, when there are intersections and connecting sections, the areas of 50 meters, 100 meters, 200 meters, 230 meters, 250 meters, 280 meters, 300 meters and 320 meters from the center of the intersection are all for different road complexities, and can be selected to ensure the accuracy of collection verification and reduce the length value of the amount of data.

[0076] For a target road section consisting of a single road section, the collection direction is fixed, and the length is preferably in the range of 230 meters, 250 meters, 280 meters, 300 meters and 320 meters.

[0077] According to the characteristics of the target intersection, determine the number of target sections radiating outward from the target intersection, that is, determine the shape of the target intersection to see whether the target intersection is a T-junction connecting 3 target sections, a crossroads connecting 4 target sections, a plum blossom intersection connecting 5 target sections, a M-shaped intersection connecting 6 target sections, or a multi-section intersection connecting more sections. According to the rule that the lateral lane is contrary to the longitudinal lane, and according to the principle that the inspection vehicle is closest to the section, collect the traffic light data of the intersection closest to the inspection vehicle and the intersection in the adjacent direction, that is, collect the traffic light data of the intersection in the direction that the inspection vehicle will pass through in its current driving path, and the traffic light data of the intersection that may be passed by when the path is changed; simplify the full amount of data perceived by the vehicle-road collaboration to reduce the amount of comparison calculation of the simplified data. Figure 3 As shown, in the target intersection / road section of this example, the target intersection is a crossroad, and there are target road sections in four directions connecting the target intersection. Figure 3In the target intersection / road section shown, mark 800 is the center position of the target intersection, and marks 94, 795, 3485 and mark 798 are respectively the marking points on the target road section connected to the target intersection mark 800. Therefore, around the target intersection mark 800, there are 4 detectable intersection directions, namely, mark 94-800, mark 795-800, 3485-800, and 798-800. According to the intersection traffic rules, as well as the current driving path of the inspection vehicle and the possible change of the driving path, the two directions of mark 94-800 and mark 798-800 are selected as the streamlined collection directions. It should be noted that the test of each intersection direction is bidirectional. For example, driving from mark 94 to mark 800 and driving from mark 800 to mark 94 are both indicated by the traffic light in the direction of mark 94-800.

[0078] There is no requirement for the periodicity of data collection for the light color, countdown, cycle, and UTC time of the traffic lights in the selected two directions. If there are only T-shaped or herringbone intersections in less than or equal to three directions at the intersection, it is only necessary to select the light color, countdown, cycle, and UTC time of the traffic lights in one direction of the intersection according to the route that the inspection vehicle is about to pass through to perform streamlined data collection on the traffic light data. There is no requirement for the periodicity of data collection.

[0079] As for the reception and collection of RSI data, no simplification is made. When the inspection vehicle passes any intersection, it receives the full amount of RSI data from the roadside equipment installed at the intersection / road end. Figure 4 As shown, taking the RSI data of the signboard as an example, the roadside equipment on the intersection direction and road section actually passed by the patrol vehicle is the communication object, and the full RSI data received by the patrol vehicle is the RSI test data used for comparison with the RSI true value data.

[0080] It should be noted that MAP data, that is, map information, is local area map information broadcast by roadside equipment to vehicles, including intersection information, road section information, lane information, etc., as well as the connection relationship between roads. Therefore, when the inspection vehicle randomly checks the direction of the intersection, the correlation between the driving direction in the direction of the randomly checked intersection and the phase information in SPAT can be obtained from the MAP data.

[0081] After determining the direction and location area of ​​the intersection / section where the road signs are collected, collect the traffic lights in the direction of the selected target intersection / section: select the key anchor points of the traffic lights at the target intersection / section: the countdown and timestamps of the two key nodes of the red light turning green and the green light turning red, and then record the countdown and timestamps within the green light and red light cycles respectively. According to the countdown and timestamp of the traffic light change, combined with the current position coordinates of the patrol vehicle, adjust the speed of the patrol vehicle so that the patrol vehicle can pass the target intersection / section smoothly without stopping and waiting, saving patrol time while enabling the patrol vehicle to simulate the normal passage of intelligent connected vehicles to complete the verification and comparison of vehicle-road collaborative data, making the verification results more accurate and true.

[0082] When receiving RSI data, the inspection vehicle receives and records all RSI data of the intersection when passing through the intersection / road section.

[0083] The types of RSI data can be increased according to actual conditions. Currently, the main RSI data include RSI data of signboards and RSI data of traffic events. Both are collected according to the following contents:

[0084] RSI message attributes - signboard: signboard type, installation location, affected area or road section;

[0085] RSI message attributes - traffic events: event type, event center point, event coverage area, and event-affected road sections.

[0086] In this embodiment, the verification of the RSI data of a signboard is taken as an example.

[0087] When comparing the simplified test data and the simplified true value data, the accuracy of the two sets of data is automatically verified according to the automatic verification strategy preset in the automatic verification module.

[0088] Specifically, the accuracy of the target direction traffic light countdown data at the target intersection is automatically verified, and compared with the preset standard range value, to determine whether the traffic light frequency, end-to-end delay, and packet loss rate meet the standard requirements, verify the accuracy of the traffic light data and RSI data in the vehicle-road cooperative system, and generate an inspection report. Among them, the preset standard range value is a set of numerical ranges set according to the current standard requirements to determine whether the traffic light frequency, end-to-end delay, and packet loss rate meet the corresponding requirements.

[0089] like Figure 2 As shown, the automatic verification strategy compares the traffic light simplified data and the traffic light test data to compare and judge the accuracy of the traffic light vehicle-road collaboration data.

[0090] During the comparative verification process, the two sets of data completed the following comparisons:

[0091] (1) The accuracy of the countdown, timestamp, and true value when the red light switches to the green light; the accuracy of the countdown, timestamp, and true value of the signal light during the red light cycle;

[0092] (2) The accuracy of the difference between the countdown, timestamp and the true value within the green light cycle.

[0093] Through the automatic verification strategy setting of this embodiment, the true value data and simplified data of the traffic light can be automatically compared, and the comparative verification can be quickly completed while ensuring the accuracy of the verification result.

[0094] like Figure 4 As shown, the automatic verification strategy, when performing RSI data comparison, compares the RSI test data received by the inspection vehicle with the RSI true value data pre-collected in the first step to verify the accuracy of sending and receiving. This embodiment takes the RSI data of the sign as an example. When performing the comparison, first determine whether the type, ID and message attributes are consistent, and then compare the installation position. When the positioning error and the error of the affected section are both less than 1 meter, and the overlap error within the affected radius of 0.1 km is within 10%, it is determined that the comparison of the two sets of data meets the requirements, that is, the corresponding sign RSI data verification is passed.

[0095] It should be noted that in order to ensure the accuracy of verification, all timestamps in this embodiment are uniformly converted into UTC time for comparison.

[0096] The present invention provides a vehicle-road cooperative signal light and RSI data operation and maintenance method suitable for complex urban road scenes, which is used to verify the accuracy of signal lights and RSI and carry out real-time operation and maintenance work.

[0097] Compared with other currently used methods, the advantages of this method are:

[0098] 1. Applicable to complex urban road scenes

[0099] By streamlining and optimizing the collection rules, the target intersection / road section is divided into reasonable areas, especially the division of the collection direction of the target intersection signal light, which can divide the complex urban roads into collection areas that are easy to collect and have limited comparative calculation.

[0100] More importantly, the patrol car simulates the characteristics of intelligent connected vehicles but can adjust the driving speed according to the specific test section. In complex urban road scenarios, it can avoid the situation where large vehicles pass through more intersections due to their size and shape. The autonomous driving camera is blocked by the vehicle itself or other reasons, or the intelligent connected vehicle cannot receive data because of the obstruction of road side equipment due to tree branches. The patrol car can slow down the traffic according to the complexity of the target intersection / section to supplement the corresponding test data.

[0101] Compared with the current intelligent connected vehicles, the perception ability of the self-driving single vehicle has certain errors due to environmental interference. There are great safety hazards in installing cameras on the self-driving vehicle to obtain the true value of the traffic light on the open and complex urban roads. This paper proposes that the operation and maintenance personnel drive the inspection vehicle to collect the target traffic light data perceived by the vehicle-road coordination system and compare it with the true value. The verification method is safer and more suitable for complex urban road scenes. Compared with the situation where the average perception error on general roads is not much different, in complex urban road scenes, due to the information interference emitted by various infrastructures themselves, and the obstruction and interference of various buildings on urban roads on the transmission of communication information, the error on complex urban roads according to the existing method is more than 20%, and this solution can reduce the error to 0.

[0102] 2. Reliable ground truth data

[0103] Cameras are installed in autonomous vehicles to obtain the true value of traffic lights through image recognition. Due to the uneven road surface and differences in intersections, there are errors in camera calibration. Due to the continuity of traffic light timing, the periodicity and accuracy of image recognition of traffic lights have certain limitations. This article directly reads traffic light data from traffic lights through original design data or mature WIFI acquisition methods, and directly obtains RSI-road sign data from traffic police as the true value, which is more reliable without time delay.

[0104] 3. Automated comparison and verification results are more accurate

[0105] When the human eye compares the traffic lights on site with the target traffic light data received by the vehicle-road cooperative system, it is easy to make judgment errors. This embodiment automatically outputs a comparison list through automated comparison according to the corresponding threshold parameters on the preset comparison table, and the comparison result is more accurate.

[0106] 4. Streamline data collection to improve operation and maintenance efficiency

[0107] According to the principle of releasing opposite lanes with traffic lights at intersections, at an intersection with four directions, only two directions need to be sampled, and at a T-shaped intersection with three directions or less, only one direction needs to be sampled to verify the accuracy of traffic lights in all directions of the intersection, thereby improving operation and maintenance efficiency. Compared with existing inspections, work efficiency is improved by 39.39% (30 minutes for one intersection and 7.5 minutes for one direction. Assuming that four-direction and three-direction intersections each account for 50%, full coverage verification of one four-direction intersection and one three-direction intersection requires 7.5*7=49.5 minutes, and streamlined coverage verification requires 7.5*4=30 minutes, saving 19.5 minutes and improving 19.5 / 49.5=39.39%).

[0108] Compared with the existing signal light verification method of the vehicle-road cooperative system, the first matching result of the target signal light information identified by the autonomous driving vehicle and the target signal light information perceived by the vehicle-road cooperative system is obtained; the running status of the vehicle in the area controlled by the target signal light perceived by the vehicle-road cooperative system is collected; the second matching result is obtained according to the running status of the vehicle and the target signal light information perceived by the vehicle-road cooperative system; and the accuracy of the signal light information in the vehicle-road cooperative system is verified according to the first matching result and the second matching result. The existing method has two shortcomings. First, the use of autonomous driving technology is not applicable in complex urban road scenes and there are certain safety hazards. Second, cameras are installed at the road end to perceive the running status of the vehicle as the second matching result. In the large-scale deployment of vehicle-road cooperation, the intersections where cameras are deployed only account for 27.63% of all intersections, and the scope of verification is very limited.

[0109] This solution collects the real value data of the infrastructure at the nearest intersection through the running track of the inspection car and filters it to obtain streamlined data. It can not only quickly complete the collection, but also quickly determine the connection between the intelligent connected car and the road infrastructure, and achieve the effect of fast and accurate verification. It can effectively avoid roadside data quality problems caused by the camera installation location, including verification delays and low accuracy. This solution can effectively improve the quality and credibility of roadside data by improving the efficiency and accuracy of verification data, providing a prerequisite for improving the perception of intelligent connected cars.

[0110] Embodiment 2

[0111] What is different from the first embodiment is that the streamlined and optimized collection rules in this embodiment are as follows: the direction where the traffic rules conflict is selected to collect the target traffic light data perceived by the vehicle-road cooperative system. For example, at an intersection with four directions, when the light for straight driving in the east-west direction is green, the light for straight driving in the north-south direction must be red, and when the light for straight driving in the east-west direction is red, the light for straight driving in the north-south direction must be green. That is, at the intersection, it is sufficient to collect data from two directions: east to west / west to east, and south to north / north to south. Similarly, for T-shaped, Y-shaped, and herringbone intersections with three directions, two directions are randomly sampled. This streamlined and optimized collection rule can be applied to most classified streamlined collection models, to most existing intersections and sections, has universality, and can effectively reduce the amount of calculation while ensuring the accuracy of verification.

[0112] Embodiment 3

[0113] Different from the first embodiment, after the second step, the simplified test data of the traffic lights is obtained by screening the traffic light test data according to the ratio of 50%-75%; the simplified test data of the traffic lights in this embodiment is not the full data. On the basis of collecting the traffic light test data in the limited area, the simplified test data of the traffic lights is further simplified to effectively reduce the amount of data calculation without affecting the accuracy of the verification judgment.

[0114] Embodiment 4

[0115] The vehicle-road cooperative data verification device for complex roads in this embodiment is used to implement the vehicle-road cooperative data verification method for complex roads, including a central processor connected to the vehicle-road cooperative perception system, and an on-board communication unit and an automatic verification module respectively connected to the central processor for communication;

[0116] The central processor is preset with a plurality of simplified acquisition models, each of which includes at least one simplified optimization acquisition rule;

[0117] The vehicle-mounted communication unit is arranged on the inspection vehicle, and the central processor selects the corresponding simplified collection model according to the terrain characteristics of the target intersection / road section and sends it to the inspection vehicle; the inspection vehicle collects the operation data of the adjacent road infrastructure along the driving track of the inspection vehicle according to the corresponding simplified optimization collection rules, processes it into simplified test data and sends it to the automated verification module;

[0118] The central processor processes the preset true value data of the corresponding target intersection / road section according to the simplified optimization collection rules, obtains the simplified true value data, and sends it to the automated verification module;

[0119] The automatic verification module is used to automatically compare the simplified test data and the simplified true value data according to the preset comparison table and threshold parameters, under the premise of meeting the threshold range, to determine whether the two are consistent, and send the judgment result to the central processing unit;

[0120] The central processing unit generates and outputs an inspection report based on the judgment result.

[0121] The comparison table and threshold parameters in this embodiment are set with reference to current regulations.

[0122] Through this device, the vehicle-road cooperative data, especially the roadside perception data corresponding to traffic lights and signboards, can be verified quickly and conveniently.

[0123] The above is only an embodiment of the present invention, and the common knowledge such as the known specific technical solutions and / or characteristics in the solution is not described in detail here. It should be pointed out that for those skilled in the art, without departing from the technical solution of the present invention, several modifications and improvements can be made, which should also be regarded as the protection scope of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A vehicle-road cooperative data verification method for complex roads, characterized in that: The following steps are involved: Step 1: Obtain the full amount of true value data of the road infrastructure preset at the target intersection / road section; Step 2: construct a simplified collection model and classify the simplified collection models for all intersections / road sections. Each type of simplified collection model has corresponding simplified optimization collection rules. Step 3: Select a simplified collection model of the corresponding classification according to the geographical features of the target intersection / road section, and screen the full amount of true value data of the target intersection / road section according to the simplified optimization collection rules to obtain the simplified true value data; Step 4: Select the target intersection direction according to the streamlined and optimized collection rules, and the inspection vehicle connected to the vehicle-road cooperative perception system collects the infrastructure operation data at the target intersection / road section to form streamlined test data; Step 5: Automatically compare the simplified test data and the simplified true value data through the preset automatic verification strategy, generate and output the inspection report; The simplified test data includes simplified test data of traffic lights and simplified test data of signboards; the simplified true value data includes true value data of traffic lights and true value data of signboards.

2. The vehicle-road cooperative data verification method for complex roads according to claim 1, characterized in that: The target intersection / road section includes the target intersection section and the target road section; the target intersection section refers to the target intersection and the connecting road section connected to it, with the target intersection center point as the center of the circle and a target area with a diameter of 50-320 meters; the target road section is a road section target area with a length of less than 300 meters excluding the intersection.

3. The vehicle-road cooperative data verification method for complex roads according to claim 1, characterized in that: The road infrastructure includes traffic lights, RSI signs, roadside equipment and roadside units; the inspection vehicle is provided with an on-board communication unit; the on-board communication unit communicates with the roadside equipment and the roadside unit respectively to obtain traffic light test data used to reflect the operation of the traffic lights, and signboard test data used to reflect the operation of the RSI signs.

4. The vehicle-road cooperative data verification method for complex roads according to claim 3, characterized in that: The simplified signal light test data is the signal light test data obtained by screening the signal light test data according to the simplified optimization collection rule; The simplified signboard test data is the signboard test data collected by the inspection vehicle according to the simplified optimization collection rules for the target intersection section or the limited area of ​​the target section.

5. The vehicle-road cooperative data verification method for complex roads according to claim 1, characterized in that: The simplified and optimized collection rule is to select a direction where traffic rules conflict and collect the signal light test data perceived by the vehicle-road cooperative system to form signal light test data.

6. The vehicle-road cooperative data verification method for complex roads according to claim 5, characterized in that: The simplified signal light test data is obtained by screening the signal light test data in a ratio of 50%-75%; the simplified signal light test data is not the full data.

7. The vehicle-road cooperative data verification method for complex roads according to claim 5, characterized in that: When the target intersection is a crossroad with four directions, data is collected from two directions: east to west / west to east, and south to north / north to south. When the target intersection is a T-shaped, Y-shaped or herringbone intersection with three directions, data is collected from two randomly selected directions.

8. The vehicle-road cooperative data verification method for complex roads according to claim 1, characterized in that: The simplified optimization collection rule includes that the inspection vehicle collects the signal light test data of the direction of the intersection that the current driving route will pass, and the signal light test data that may be passed after the route is changed as the signal light simplified test data; The inspection vehicle collects the sign test data of the direction of the intersection that the current driving route will pass, as well as the sign test data of the intersection that may be passed after the route is changed as the simplified sign test data; the traffic light simplified test data and the sign simplified test data are both full data.

9. A vehicle-road cooperative data verification device for complex roads, characterized in that: Used to implement the vehicle-road cooperative data verification method for complex roads as claimed in claim 1, comprising a central processor connected to the vehicle-road cooperative perception system, and an on-board communication unit and an automatic verification module respectively connected to the central processor for communication; The central processor is preset with a plurality of simplified acquisition models, each of which includes at least one simplified optimization acquisition rule; The vehicle-mounted communication unit is arranged on the inspection vehicle, and the central processor selects the corresponding simplified acquisition model according to the terrain characteristics of the target intersection / road section and sends it to the inspection vehicle; The inspection vehicle collects the operation data of the adjacent road infrastructure along its driving trajectory according to the corresponding streamlined and optimized collection rules, processes it into streamlined test data and sends it to the automated verification module; The central processor processes the preset true value data of the corresponding target intersection / road section according to the simplified optimization collection rules, obtains the simplified true value data, and sends it to the automated verification module; The automated verification module is used to automatically compare the simplified test data with the simplified true value data, determine whether the two are consistent, and send the determination result to the central processing unit; The central processing unit generates and outputs an inspection report based on the judgment result.

10. A computer-readable storage medium, characterized in that: It stores a computer program, which, when executed by a processor, implements the vehicle-road cooperative data verification method for complex roads as claimed in claim 1.

Images