A vehicle end real-time road right monitoring system of an RSS-based autonomous vehicle

Abstract

This invention discloses a real-time right-of-way monitoring system for autonomous vehicles based on RSS (Real-Time Sequence of Traffic Flow). Belonging to the field of autonomous driving, it constructs a real-time vehicle-side violation monitoring system for autonomous vehicles, consisting of a perception input module, a high-precision map module, a decision-making module, a CAN bus, a monitoring trigger judgment module, and a violation monitoring module. This system can monitor the right-of-way compliance of autonomous vehicles at intersections in real time. The system uses a high-precision map for assisted monitoring and constructs an auxiliary monitoring domain through region segmentation. Furthermore, based on the RSS formula, it designs dynamic thresholds for right-of-way monitoring under different right-of-way conflict conditions, reducing the input requirements of the perception system and achieving better monitoring and judgment accuracy compared to previous region segmentation methods.

Description

Technical Field

[0001] This invention belongs to the field of autonomous driving, specifically relating to a vehicle-side real-time right-of-way monitoring system for autonomous vehicles based on RSS. Background Technology

[0002] With the development of autonomous vehicles, they will share roads with human drivers for a considerable period. Therefore, it is necessary to take measures to improve autonomous vehicles' awareness of regulations to enhance human trust in their safety. Currently, the behavioral decisions of autonomous vehicles primarily focus on safety, and while some regulations are involved, they are limited to specific and simple rules. This is mainly because existing road traffic regulations are written in natural language, and the ambiguity and lack of standardization in some rule expressions make them difficult for autonomous vehicles to understand. Therefore, it is necessary to develop a system that can accurately monitor vehicle behavior, thereby helping autonomous vehicles correctly understand and comply with traffic regulations. Furthermore, monitoring vehicle behavior compliance can provide strong evidence for subsequent liability determination in traffic accidents. Among these issues, right-of-way is the most common and most likely to cause traffic accidents, and it is not sufficiently clearly defined in road regulations.

[0003] Current technologies for automatically monitoring vehicle violations rely solely on roadside equipment and can only monitor simple, explicit rules such as speed limits and running red lights. However, monitoring right-of-way rules requires post-event human judgment based on video recordings, and is impossible at intersections without roadside equipment. The Road Traffic Safety Center of the Ministry of Public Security invented a method and system for simulating traffic regulation compliance testing of autonomous vehicles in patent CN 113254336 A, published in 2021. This invention aims to achieve traffic regulation compliance simulation testing. By using a predefined formalized LTL traffic regulation library, information on traffic participants is obtained, and the truth value of atomic propositions and their logical combination relationships are determined using LTL formulas defined in the library, thereby assessing the compliance of traffic participants.

[0004] Current technologies rely on all information throughout the entire process for judgment, making them only suitable for simulation or offline testing. However, autonomous driving has limited perception capabilities, and much of the information used for judgment is difficult to obtain directly from the vehicle in real time, thus rendering it unsuitable for real-time violation detection on the vehicle side.

[0005] As one of the most common traffic scenarios, intersections are not only the locations with the highest frequency of traffic violations but also the areas with the most traffic accidents. Therefore, this patent uses intersections as its design operating domain and proposes a real-time right-of-way violation monitoring system for autonomous vehicles to address these issues. This patent constructs a real-time autonomous vehicle-side right-of-way violation monitoring system consisting of a perception input module, a high-precision map module, a decision-making module, a monitoring trigger judgment module, and a violation monitoring module. By acquiring the current intersection scene information through the perception system and combining it with high-precision map information for logical judgment, it can achieve real-time monitoring of right-of-way violations by vehicles passing through intersections. The dynamic threshold designed based on the Responsibility Sensitive Safety (RSS) model offers better flexibility and accuracy compared to previous region segmentation methods. Furthermore, combining it with a high-precision map reduces the input requirements of the perception system, and dividing the violation into trigger judgment and logical judgment parts further reduces the system's computational power requirements. Therefore, this system has advantages such as real-time monitoring, high accuracy, low perception requirements, low computational power requirements, and strong applicability. Summary of the Invention

[0006] In order to solve the technical problems existing in the background art, the present invention aims to provide a vehicle-side real-time right-of-way monitoring system for autonomous vehicles based on RSS.

[0007] To solve the technical problem, the technical solution of the present invention is as follows:

[0008] A vehicle-side real-time right-of-way monitoring system for autonomous vehicles based on RSS, the system comprising: a perception input module, a high-precision map module, a decision module, a CAN bus, a monitoring trigger judgment module, and a violation monitoring module;

[0009] The perception input module includes: an on-board camera, radar, and a vehicle status sensor. The on-board camera and radar are used to perceive the status of other vehicles in the surrounding environment, and the vehicle status sensor is used to acquire the vehicle's position, speed, acceleration, and heading angle information in the global coordinate system.

[0010] The high-precision map module is used to provide coordinate information of the static components of the intersection, including the endpoint coordinates of the stop line in the global coordinate system, the ID number of each lane and the coordinates of the boundary vertices, and the directional indicator marking information.

[0011] The decision module is used to provide behavioral decisions for the vehicle when passing through an intersection, which are divided into left turn, right turn and straight ahead. It is also used to calculate the time it takes for the vehicle to pass through a certain position.

[0012] Once the vehicle enters the preset operating domain, the high-precision map inputs static intersection information into the CAN bus. The perception input module acquires the status information of other vehicles in real time and transmits it to the CAN bus. The decision module inputs the vehicle's decision information into the CAN bus. The monitoring trigger judgment module reads the above input information from the CAN bus in real time and judges the trigger conditions. When the trigger condition of a certain rule is met, the violation monitoring module of the corresponding rule receives the above input information from the CAN bus, monitors the vehicle's right-of-way violations in real time, and outputs the monitoring results to the CAN bus.

[0013] Furthermore, the status of other vehicles includes: the position, velocity, acceleration, heading angle, turn signal status, and geometric dimensions of other vehicles in the global coordinate system.

[0014] Furthermore, the preset operating domain ODD is a polygonal region enclosed by the extension of the stop line within the intersection.

[0015] Furthermore, the system operation method includes:

[0016] S1: The high-precision map module inputs the static component information of the intersection into the CAN bus. When a vehicle enters the ODD of the right-of-way monitoring system, the high-precision map module sends the static component information required for right-of-way monitoring to the CAN bus in a predefined protocol format. The static component information includes the endpoint coordinates of the stop line in the global coordinate system, the ID number and boundary vertex coordinates of each intersection lane, and the direction indicator marking information. The high-precision map information includes the IDs of each branch intersection - R1, R2, R3, R4, with the branch intersection IDs increasing counterclockwise; and the number of entrance lanes N at each branch intersection. en and the number of exit lanes N ex Lanes entering the intersection area are defined as entrance lanes, and lanes leaving the intersection area are defined as exit lanes. Lanes are classified as front, rear, left, and right based on their direction of travel. For each branch intersection, the lane ID is defined as follows: L1 is the lane closest to the road centerline in the entrance direction, increasing sequentially to the right as L1, L2, ...; L is the lane closest to the road centerline in the exit direction. -1 Decrease L to the right -1 L -2 ...; Each lane is considered as a quadrilateral consisting of four vertices, including the coordinates of the four vertices and the lane marking information. Each lane is represented by a 1×11 array [X... lf X lr X rf X rr Y lf Y lr Y rf Y rr[TL, GS, TR], where each element represents the x-coordinate of the left front point, the x-coordinate of the left rear point, the x-coordinate of the right front point, the x-coordinate of the right rear point, the y-coordinate of the left front point, the y-coordinate of the left rear point, the y-coordinate of the right front point, the y-coordinate of the right rear point, the Boolean symbol for left turn, the Boolean symbol for straight ahead, and the Boolean symbol for right turn in the global coordinate system; the stop line s is a 1×4 array [X... l X r , Y l, Y r Each element represents the left x-coordinate, right x-coordinate, left y-coordinate, and right y-coordinate of the stop line in the global coordinate system. The left and right sides of the stop line are defined the same as the corresponding entrance lanes. Example of information representation: The number of entrance lanes with branch intersection id 1 is represented as R1.N. en The lane with branch intersection ID 2 and lane ID 1 is represented as R2.L1; the ordinate to the right of the stop line for branch intersection ID 1 is represented as R1. . s(4);

[0017] S2: The perception input module acquires information about the vehicle itself and other vehicles, and obtains the vehicle's E-value through the vehicle positioning device combined with the high-precision map module. go The ordinate X in the global coordinate system Ego x-axis T Ego Longitudinal velocity vx Ego , vehicle heading angle θ Ego And the lane that the vehicle belongs to when it enters the intersection. Ego The longitudinal acceleration 'a' of the vehicle can be obtained through the inertial measurement unit (IMU). Ego Cameras and radar sense the surrounding environment of other vehicles (Tgt) and the status of traffic lights. The information about other vehicles includes their x-coordinate in the global coordinate system. Tgt , ordinate Y Tgt The length of his car, L Tgt Width W Tgt The longitudinal speed vc of his car Tgt Heading angle information θ Tgt Turn signal status: Light Tgt And the lane his car was in when he entered the intersection. Tgt The traffic light status is TrafficLight, which includes three states: red light 0, green light 1, and yellow light 2. The sensing input module sends the acquired information to the CAN bus using a predefined bus protocol.

[0018] S3: The decision module provides vehicle decision information. Before entering the intersection, the vehicle has already decided on the global path and made the decision. EgoThere are three types: left turn (TL), straight (GS), and right turn (TR). When a vehicle enters the right-of-way monitoring system's ODD, the decision module sends the decision information to the CAN bus using a predefined bus protocol.

[0019] S4: The right-of-way monitoring trigger judgment module reads the information from steps one to three above from the CAN bus in real time;

[0020] S5: The right-of-way monitoring trigger judgment module judges different traffic light states and decision information, which will trigger right-of-way monitoring in different situations. There are three situations in which a vehicle needs to give up its right-of-way:

[0021] 1. When a vehicle turns left, its right-of-way is lower than that of oncoming vehicles going straight;

[0022] 2. When a vehicle turns right on a green light, its right-of-way is lower than that of an oncoming vehicle turning left;

[0023] 3. When a vehicle turns right at a red light, its right-of-way is lower than that of vehicles going straight on the left;

[0024] When autonomous vehicles navigate within an intersection, they follow an optimal circular curve. First, based on the vehicle's initial lane upon entering the intersection and its own decision-making process, an optimal virtual lane is generated. The generation method is as follows:

[0025] Enter the lane your vehicle belongs to when entering the intersection. Ego And map information and vehicle decision information. Ego The optimal exit lane and the unrecommended exit lane are calculated as follows:

[0026] Let NELane Ego =Lane Ego =R i .L j The optimal exit lane is:

[0027]

[0028] After obtaining the optimal exit lane, enter the vehicle's entry lane ENLane. Ego Best Exit Lane (BEXLane) Ego With map information, the vertices of the optimal virtual lane can be obtained:

[0029] The left and right rear vertices of the optimal virtual lane are the entrance lanes (ENLane). Ego The left and right front vertices:

[0030] Left front vertex: P enlf =(ENLane) Ego .X lf ENLane Ego .Ylf );

[0031] Right front vertex: P enrf =(ENLane) Ego .X rf ENLane Ego .Y rf );

[0032] The optimal virtual lane has its left front vertex P. bexlr =(BEXLane) Ego .X lr BEXLane Ego .Y lr The right front vertex is P. bexrr =(BEXLane) Ego .X rr BEXLane Ego .Y rr After obtaining the vertices of the optimal virtual lane, input the coordinates of these vertices and decision information to generate the optimal virtual lane boundary. If the vehicle decides to turn left, then take P. enlf = (x1, y1), P enrf = (x2, y2), let the left front and right front vertices of the optimal virtual lane calculated above be (x3, y3), and substitute them into the following formula to obtain the center (x, y) of the left and right arc boundaries of the virtual lane respectively:

[0033]

[0034] If the vehicle decides to turn right, then take P. enlf = (x2, y2), P enrf = (x1, y1), the calculation method is the same as the above formula;

[0035] After obtaining the center of the circular boundary of each virtual lane, let the center be c = [x c y c ], then the radius of the circular boundary of each virtual lane is If the boundary of the circular curve is represented as [start point, (center, radius), end point], then the optimal virtual lane area can be represented as:

[0036] BVLane = [[Pn] elf , (c bl r bl ), P bexlr ], [P enrf , (c br r br ), P bexrr ]

[0037] Among them, c bl cbr These represent the left and right boundary centers of the optimal virtual lane, respectively; r bl r br These represent the left and right boundary radii of the optimal virtual lane, respectively.

[0038] If a vehicle decides to proceed straight, the virtual lane is a quadrilateral area, which can be directly represented as:

[0039] BVLane = [P enlf P bexlr P bexrr P enrf ]

[0040] After obtaining the optimal virtual lane area, K virtual stop lines (Vstopline) will be generated within the lane. k The number of K is equal to the number of other high-rights optimal virtual lanes intersecting with the optimal virtual lane; the generation method is as follows:

[0041] Still set up ENLane Ego =R i L j The branch intersection on the left side of the vehicle entrance lane is Opposite branch intersections are The right-hand branch intersection is N len =R l .N en N represents the number of lanes at the left-hand branch intersection entrance; oen =R o .N en N represents the number of lanes at the entrance of the opposite branch intersection. lex =R l .N ex This refers to the number of exit lanes at the left-hand branch intersection;

[0042] If Decision Ego =tl,P vsl_k For the opposite entrance lane R o .L k A point on the left boundary of the optimal virtual straight lane, where the distance to the left and right boundaries of the optimal virtual lane is equal; the virtual stop line can be represented as:

[0043] Vstopline k =y(c bl P vsl_k k = 1, 2...N oen

[0044] Where y(p1, p2) represents the equation of the line containing points p1 and p2;

[0045] If Decision Ego =tr, Pvsl_k is the left-side entrance lane R l .L k A point on the right boundary of the optimal virtual straight lane, where the distance to the left and right boundaries of the optimal virtual lane is equal; the virtual stop line can be represented as:

[0046] Vstopline k =y(c br P vsl_k ), k = j, j+1...N len

[0047] Calculate the virtual stop line Vstopline k Then, based on the virtual stop line equation, input the ordinate X of the vehicle in the global coordinate system. Ego x-axis Y Ego Heading angle θ Ego And the longitudinal speed vx of the vehicle Ego Decision information of the vehicle Ego In addition to the traffic light status information (TrafficLight), right-of-way monitoring can be triggered, and the judgment formula is as follows:

[0048]

[0049]

[0050]

[0051] Where T_tl, T_tr_R, and T_tr_G represent three different right-of-way monitoring scenarios; overlap(ele1, ele2) is an atomic proposition indicating whether two elements have an overlapping region. If ele1 and ele2 overlap, the output is 1; otherwise, the output is 0. ele1 and ele2 can be line segments or planar regions. The formula for calculating the atomic proposition is: segment(p1, p2) represents a line segment formed by connecting points p2 and p2;

[0052] L represents the coordinates of the midpoint of the front end of the vehicle; Ego The length of the vehicle; Crdn Ego =(X Ego Y Ego () represents the coordinates of the vehicle's geometric center;

[0053] If the calculation result of the monitoring trigger condition is 1, it means that this situation needs to be monitored at this time, and the corresponding right-of-way violation monitoring module starts working, proceeding to step six; if the calculation result of the monitoring trigger condition is 0, it means that this regulation does not need to be monitored at this time, and step five is repeated.

[0054] S6: The right-of-way violation monitoring module receives data. When the calculation result of the monitoring trigger condition for a certain situation in step five is 1, the right-of-way violation monitoring module for that situation starts to receive the information from step one to step three from the CAN bus.

[0055] S7: The violation monitoring module performs right-of-way violation judgments. It inputs the vehicle's decision information, map information, traffic light status information, and optimal virtual lane information. Based on different situations, it first generates a judgment area (JudgArea). k Each virtual stop line corresponds to a judgment area, and is generated as follows:

[0056] If Decision Ego =tl, take the opposite entrance lane R o .L k The intersection point of the right boundary of the optimal virtual straight lane and the right boundary of the optimal virtual lane for the vehicle is P. jarf_k , through P jarf_k To R o .L k The optimal virtual straight lane has a perpendicular line drawn from its left boundary, and the point where the perpendicular line intersects it is set as P. jalf_k Then determine the region JudgArea k It can be represented as R o .L k The area enclosed by four points within the optimal virtual straight lane range, namely:

[0057] JudgArea k =[P jarf_k P jalf_k , (R o .L k .X lf R o .L k .Y lf ), (R o .L k .X rf R o .L k .Y rf k = 1, 2...N oen

[0058] The corresponding right-of-way determination line is called JudgLine. k =(P jarf_k P jalf_k );

[0059] If Decision Ego =tr∧TrafficLight=1, take the left entrance lane R l .L k The intersection point of the right boundary of the optimal virtual straight lane and the left boundary of the optimal virtual lane for the vehicle is P. jarf_k , through P jarf_k To R l .L k The optimal virtual straight lane has a perpendicular line drawn from its left boundary, and the point where the perpendicular line intersects it is set as P. jalf_k Then determine the region JudgArea k It can be represented as R l .L k The area enclosed by four points within the optimal virtual straight lane range, namely:

[0060] JudgArea k =[P jarf_k P jalf_k , (R l .L k .X lf R l .L k .Y lf (R) l .L k .X rf R l .L k .Y rf )}, k=j, j+1...N len

[0061] The corresponding right-of-way determination line is called JudgLine. k =(Pj arf_k P jalf_k );

[0062] If Decision Ego =tr∧TrafficLight=2, take the opposite entrance lane R o .L k The intersection point of the right boundary of the optimal virtual left-turn lane and the left boundary of the optimal virtual lane for the vehicle is P. jarf_k , through P jarf_k To R o .L k The optimal virtual left-turn lane has a perpendicular line drawn from its left boundary, and the intersection point is set as R. jalf_k Then determine the region JudgArea k It can be represented as R o .L k The area enclosed by four points within the optimal virtual left-turn lane range, namely:

[0063] JudgArea k =[P jarf_k P jalf_k , (R o .L k .X lf R o .L k .Y lf ), (R o .L k .X rf .R o .L k .Y rf )], k = j, j+1...N oen

[0064] The corresponding right-of-way determination line is called JudgLine. k =(P jarf_k P jalf_k );

[0065] After generating the judgment area and right-of-way judgment line, input the vehicle status information and other vehicle status information, and calculate whether the vehicle obstructs the passage of vehicles with high right-of-way. The calculation method is as follows:

[0066] If Decision Ego =tl, triggering the judgment within the time limit, the vehicle is perpendicular to Vstopline. k When they intersect, let Vstopline be the line from the midpoint of the rear end of the vehicle along the optimal virtual lane to the next virtual lane line. k+1 The distance is L c If k = N oen Then, it is the distance from the midpoint of the rear of the vehicle along the optimal virtual lane to the opposite entrance lane R. o .L Noen The distance to the right boundary; then, based on the longitudinal dynamics model, the distance to L in the current state can be calculated. c The required time is t c A violation of right-of-way can be represented as:

[0067]

[0068]

[0069] Where distance(p1, y1) represents the distance from point p1 along the optimal virtual lane to line y1, ρ′=max(ρ-t, 0) is the remaining reaction time, ρ is the reaction time (taken as ρ=0.2s), and t is the time from the start of triggering to the current moment. acc This represents the maximum acceleration during travel at the intersection.

[0070] If DecisionEgo =tr∧TrafficLight=1, triggering the judgment within the time limit, the vehicle is perpendicular to Vstopline k Intersection, similarly defined as Vstopline, which extends from the midpoint of the rear end of the vehicle along the optimal virtual lane to the next virtual lane line. k+1 The distance is L c (If k = N) len Then, it is the distance from the midpoint of the rear of the vehicle along the optimal virtual lane to the optimal exit lane (BEXLane). Ego The distance to the rear boundary); then, according to the longitudinal dynamics model, the distance to the current state of the vehicle passing L is calculated. c The required time is t c A violation of right-of-way can be represented as:

[0071]

[0072] Where k = j indicates that after passing the intersection, the vehicle will be in the same lane as the vehicle with the higher right-of-way, while k ≠ j indicates that the vehicle's trajectory intersects with the vehicle with the higher right-of-way, and they will not merge into the same lane after passing the intersection; d gap JudgLine j To the best exit lane BEXLane Ego The distance from the back boundary; d RSS This indicates that after passing through an intersection, a reasonable RSS distance must still be maintained from vehicles with high right-of-way. The specific expression is:

[0073]

[0074] Among them, vx Ego_c Based on the longitudinal dynamics model, the current state of the vehicle passing through L is obtained. c The speed at time a max，brake For the vehicle's automatic emergency braking (AEB) deceleration, a min，brake For vehicles with high right-of-way, take the minimum emergency braking deceleration.

[0075] If Decision Ego =tr∧TrafficLight=2, triggering the judgment within the time limit, the vehicle is perpendicular to Vstopline k Intersection, similarly defined as Vstopline, which extends from the midpoint of the rear end of the vehicle along the optimal virtual lane to the next virtual lane line. k+1 The distance is L c If k = N len Then, it is the distance from the midpoint of the rear of the vehicle along the optimal virtual lane to the optimal exit lane (BEXLane). Ego The distance to the rear boundary; then, according to the longitudinal dynamics model, the distance to the current state of the vehicle passing L is calculated. c The required time is tc A violation of right-of-way can be represented as:

[0076]

[0077] Based on the right-of-way violation calculation formulas for the three scenarios, calculate the right-of-way violation result. If the output is 1, it indicates that the right-of-way rule has been violated at the current moment; if the output is 0, it indicates that the right-of-way rule has not been violated at the current moment.

[0078] S8: The monitoring trigger module determines whether the monitoring has ended; after step seven outputs a set of results, the monitoring trigger module will make another judgment to determine whether the monitoring of the regulation should end at the current moment. If the exit condition is not met, it means that the regulation still needs to be monitored, and steps four to seven need to be repeated; if the exit condition is met, it means that the monitoring of the regulation has ended and the right-of-way violation monitoring module stops working.

[0079] Furthermore, in step one, the left-turn Boolean symbol is: 1 represents that the lane can turn left, and 0 represents that it cannot turn left; the straight-ahead Boolean symbol is: 1 represents that the lane can go straight, and 0 represents that it cannot go straight; the right-turn Boolean symbol is: 1 represents that the lane can turn right, and 0 represents that it cannot turn right.

[0080] Compared with the prior art, the advantages of the present invention are as follows:

[0081] 1. The technical solution of this invention proposes a vehicle-side real-time right-of-way monitoring system for autonomous vehicles based on RSS in intersection scenarios. By digitizing the regulations related to right-of-way and setting dynamic monitoring thresholds, the system enables autonomous vehicles to better understand and comply with traffic regulations. At the same time, the system can accurately monitor the compliance of autonomous vehicle behavior.

[0082] 2. This invention can guide autonomous vehicles to make decisions that comply with right-of-way monitoring results, thereby improving their legal compliance and making their behavior more in line with human driving patterns. This improvement can increase people's sense of security regarding autonomous vehicles and reduce overreactions caused by violations by autonomous vehicles, thus reducing the occurrence of dangerous behaviors.

[0083] 3. The monitoring system can record and provide detailed data on the behavior of autonomous vehicles, which can be used for accident investigations and liability analysis to help determine liability and clarify attribution. Attached Figure Description

[0084] Figure 1 A block diagram of the real-time right-of-way violation monitoring system for autonomous vehicles.

[0085] Figure 2 Flowchart for real-time violation monitoring at intersections;

[0086] Figure 3 Example diagram of the map information structure of a crossroads. Detailed Implementation

[0087] The specific implementation of the present invention is described below with reference to embodiments:

[0088] It should be noted that the structures, proportions, sizes, etc. shown in this specification are only used to complement the content disclosed in the specification for those skilled in the art to understand and read, and are not intended to limit the conditions under which the present invention can be implemented. Any modifications to the structure, changes in the proportions, or adjustments to the size, without affecting the effects and objectives that the present invention can produce, should still fall within the scope of the technical content disclosed in the present invention.

[0089] Furthermore, the terms such as "upper," "lower," "left," "right," "middle," and "one" used in this specification are merely for clarity of description and are not intended to limit the scope of the invention. Any changes or adjustments to their relative relationships, without substantially altering the technical content, should also be considered within the scope of the invention.

[0090] Example 1:

[0091] This embodiment relates to a real-time right-of-way violation monitoring system for autonomous vehicles in intersection scenarios. For example... Figure 1The diagram shows the structural block diagram of a real-time right-of-way violation monitoring system for autonomous vehicles. It mainly consists of a perception input module, a high-precision map module, a decision module, a CAN bus, a monitoring trigger judgment module, and a violation monitoring module. The perception input module includes an onboard camera, radar, and vehicle status sensors. The camera and radar are used to perceive the status of other vehicles in the surrounding environment, including their position, speed, acceleration, heading angle, turn signal status, and geometric dimensions in the global coordinate system. The vehicle status sensors are used to acquire the vehicle's position, speed, acceleration, and heading angle information in the global coordinate system. The high-precision map module provides the coordinate information of the static components of the intersection, including the endpoint coordinates of the stop lines in the global coordinate system, the ID numbers and boundary vertex coordinates of each lane at the intersection, and directional marking information. The decision module provides behavioral decisions for the vehicle when passing through the intersection, categorized as left turn, right turn, and straight ahead. It also calculates the time it takes for the vehicle to pass a certain location. The design operating domain (ODD) of this system is a polygonal area enclosed by the extension of the stop lines within the intersection. After the vehicle enters the designed operating domain, the high-precision map inputs static intersection information into the CAN bus. The perception input module acquires the status information of other vehicles in real time and transmits it to the CAN bus. The decision module inputs the vehicle's decision information into the CAN bus. The monitoring trigger judgment module reads the above input information from the CAN bus in real time and judges the trigger conditions. When the trigger condition of a certain rule is met, the right-of-way logic judgment module of the corresponding rule receives the above input information from the CAN bus, performs right-of-way violation monitoring of the vehicle in real time, and outputs the monitoring results to the CAN bus.

[0092] The proposed solution in this embodiment abstracts a dynamic threshold from the RSS and makes a real-time judgment based on the actual distance to vehicles with high right-of-way within the corresponding area. Alternatively, a fixed right-of-way judgment area can be obtained by segmenting intersection areas based on a high-precision map, thereby achieving right-of-way monitoring.

[0093] Example 2:

[0094] The specific steps for real-time monitoring of right-of-way violations at intersections are as follows: Figure 2 As shown.

[0095] Step 1: The high-precision map inputs the static component information of the intersection into the CAN bus. When a vehicle enters the right-of-way monitoring system's ODD (Operational Domain Name), the high-precision map sends the static component information required by the system to the CAN bus in a predefined protocol format. This static component information includes the endpoint coordinates of the stop line in the global coordinate system, the lane IDs and boundary vertex coordinates of each intersection, and directional marking information. Specific information is as follows... Figure 3As shown, the high-precision map information includes the ID of each branch intersection (R1, R2, R3, R4), with the branch intersection IDs increasing counterclockwise; and the number of entrance lanes (N) for each branch intersection. en ) and the number of exit lanes (N) ex (The lanes entering the intersection area are defined as entrance lanes, and the lanes leaving the intersection area are defined as exit lanes); lanes are divided into front, back, left, and right based on their direction of travel. For each branch intersection, the lane ID is defined as follows: L1 is the lane closest to the road centerline in the entrance direction, increasing sequentially to the right (L1, L2...); L is the lane closest to the road centerline in the exit direction. -1 Decreasing sequentially to the right (L) -1 L -2 …); Each lane is considered as a quadrilateral consisting of four vertices, including the coordinates of the four vertices and the lane marking information. Each lane is represented by a 1×11 array ([X…). lf X lr X rf X rr Y lf Y lr Y rf Y rr [TL, GS, TR]), where each element represents the x-coordinate of the left front point, the x-coordinate of the left rear point, the x-coordinate of the right front point, the x-coordinate of the right rear point, the y-coordinate of the left front point, the y-coordinate of the left rear point, the y-coordinate of the right front point, and the y-coordinate of the right rear point in the global coordinate system, a left turn boolean symbol (1 means the lane can turn left, 0 means it cannot), a straight go boolean symbol (1 means the lane can go straight, 0 means it cannot), and a right turn boolean symbol (1 means the lane can turn right, 0 means it cannot); the stop line (s) is a 1×4 array of the coordinates of the two vertices ([X l X r Y l Y r Each element represents the left x-coordinate, right x-coordinate, left y-coordinate, and right y-coordinate of the stop line in the global coordinate system. The left and right sides of the stop line are defined the same as the corresponding entrance lanes. Example of information representation: The number of entrance lanes with branch intersection id 1 is represented as R1.N. en The lane with branch intersection ID 2 and lane ID 1 is represented as R2.L1; the right ordinate of the stop line with branch intersection ID 1 is represented as R1·s(4).

[0096] Step Two: The sensing input module acquires information about the vehicle and other vehicles. By combining the vehicle positioning device with a high-precision map, the ordinate (X) of the vehicle (Ego) in the global coordinate system can be obtained. Ego x-axis Y Ego Longitudinal velocity vx Ego , vehicle heading angle θ EgoAnd the lane that the vehicle belongs to when it enters the intersection. Ego The longitudinal acceleration a of the vehicle can be obtained through the inertial measurement unit (IMU). Ego Cameras and radar sense information about other vehicles (Tgt) in the surrounding environment, as well as the status of traffic lights. This information includes the vehicle's x-coordinate in the global coordinate system. Tgt , ordinate Y Tgt The length of his car, L Tgt Width W Tgt The longitudinal speed vx of his car Tgt Heading angle information θ Tgt Turn signal status: Light Tgt (A 1×2 Boolean array [L, R], where 1 represents on and 0 represents off) and the lane the other car belongs to when it enters the intersection. Tgt The traffic light status is TrafficLight, which includes three states: red (0), green (1), and yellow (2). The sensing input module will send the acquired information to the CAN bus according to the defined bus protocol.

[0097] Step 3: The decision-making module provides vehicle decision information. Before entering the intersection, the vehicle has already decided on its overall path. Ego There are three types: left turn (TL), straight (GS), and right turn (TR). When a vehicle enters the right-of-way monitoring system's ODD, the decision module sends the decision information to the CAN bus using a predefined bus protocol.

[0098] Step 4: The right-of-way monitoring trigger judgment module receives data. The right-of-way monitoring trigger judgment module reads the information from Steps 1 to 3 from the CAN bus in real time.

[0099] Step 5: Right-of-way monitoring triggers different traffic light states and decision information, which will trigger right-of-way monitoring in different situations. There are three situations in which a vehicle needs to give up its right-of-way:

[0100] 1. When a vehicle turns left, its right-of-way is lower than that of oncoming vehicles going straight;

[0101] 2. When a vehicle turns right on a green light, its right-of-way is lower than that of an oncoming vehicle turning left;

[0102] 3. When a vehicle turns right at a red light, its right-of-way is lower than that of vehicles going straight on the left.

[0103] When autonomous vehicles navigate intersections, they will follow optimal circular curves, such as... Figure 3 As shown, the optimal virtual lane is first generated based on the initial lane when the vehicle enters the intersection and the vehicle's decision. The generation method is as follows:

[0104] Enter the lane your vehicle belongs to when entering the intersection. Ego And map information and vehicle decision information. Ego The optimal exit lane and the unrecommended exit lane are calculated as follows:

[0105] Let ENLane Ego =Lane Ego =R i .L j The optimal exit lane is:

[0106]

[0107] After obtaining the optimal exit lane, enter the vehicle's entry lane ENLane. Ego Best Exit Lane (BEXLane) Ego With map information, the vertices of the optimal virtual lane can be obtained:

[0108] The left and right rear vertices of the optimal virtual lane are the entrance lanes (ENLane). Ego The left and right front vertices;

[0109] Left front vertex P enlf =(ENLane) Ego .X lf ENLane Ego .Y lf );

[0110] Right front vertex P enrf =(ENLane) Ego .X rf ENLane Ego .Y rf The optimal virtual lane's left front vertex is P. bexlr =(BEXLane) Ego .X lr BEXLane Ego .Y lr );

[0111] The right front vertex is P bexrr =(BEXLane) Ego .X rr BEXLane Ego .Y rr After obtaining the vertices of the optimal virtual lane, input the coordinates of these vertices and decision information to generate the optimal virtual lane boundary. If the vehicle decides to turn left, then take P. enlf = (x1, y1), P enrf= (x2, y2), let the left front and right front vertices of the optimal virtual lane calculated above be (x3, y3), and substitute them into the following formula to obtain the center (x, y) of the left and right arc boundaries of the virtual lane respectively:

[0112]

[0113]

[0114] If the vehicle decides to turn right, then take P. enlf = (x2, y2), P enrf = (x1, y1), and the calculation method is the same as the above formula.

[0115] After obtaining the center of the circular arc boundary of each virtual lane, let the center be c = [xc, yc]. Then the radius of the circular arc boundary of each virtual lane is... If the boundary of the circular curve is represented as [start point, (center, radius), end point], then the optimal virtual lane area can be represented as:

[0116] BVLane = [[P enlf , (c bl r bl )P bexlr ], [P enrf , (c br r br ), P bexrr ]]

[0117] Among them, c bl c br These represent the left and right boundary centers of the optimal virtual lane, respectively; r bl r br These represent the left and right boundary radii of the optimal virtual lane, respectively.

[0118] If a vehicle decides to proceed straight, the virtual lane is a quadrilateral area, which can be directly represented as:

[0119] BVLane = [P enlf P bexlr P bexrr P enrf ]

[0120] After obtaining the optimal virtual lane area, K virtual stop lines (Vstopline) will be generated within the lane. k The number of K lanes is equal to the number of other high-rights optimal virtual lanes that intersect with the optimal virtual lane. The generation method is as follows:

[0121] Still set up ENLane Ego =R i L jThe branch intersection on the left side of the vehicle entrance lane is Opposite branch intersections are The right-hand branch intersection is N len =R l .N en R represents the number of lanes at the left-hand branch intersection entrance; oen =R o .N en N represents the number of lanes at the entrance of the opposite branch intersection. lex =R l .N ex This represents the number of exit lanes at the left-hand branch intersection.

[0122] If Decision Ego =tl,P vsl_k For the opposite entrance lane R o .L k A point on the left boundary of the optimal virtual straight lane, where the distance to the left and right boundaries of the optimal virtual lane for the vehicle is equal. The virtual stop line can be represented as:

[0123] Vstopline k =y(c bl P vsl_k k = 1, 2...N oen

[0124] Where y(p1, p2) represents the equation of the line containing points p1 and p2.

[0125] If Decision Ego =tr,P vsl_k For the left-side entrance lane R l .L k A point on the right boundary of the optimal virtual straight lane, where the distance to the left and right boundaries of the optimal virtual lane is equal. The virtual stop line can be represented as:

[0126] Vstopline k =y(c br P vsl_k ), k = j, j+1...N len

[0127] Calculate the virtual stop line Vstopline k Then, based on the virtual stop line equation, input the ordinate X of the vehicle in the global coordinate system. Ego x-axis Y Ego , heading angle θ Ego And the longitudinal speed vx of the vehicle Ego Decision information of the vehicle EgoIn addition to the traffic light status information (TrafficLight), right-of-way monitoring can be triggered, and the judgment formula is as follows:

[0128]

[0129]

[0130]

[0131] Where T_tl, T_tr_R, and T_tr_G represent three different right-of-way monitoring scenarios; overlap(ele1, ele2) is an atomic proposition indicating whether two elements have an overlapping region. If ele1 and ele2 overlap, the output is 1; otherwise, the output is 0. ele1 and ele2 can be line segments or planar regions. The formula for calculating the atomic proposition is: segment(p1, p2) represents a line segment formed by connecting points p2 and p2;

[0132] L represents the coordinates of the midpoint of the front end of the vehicle; Ego The length of the vehicle; Crdn Ego =(X Ego Y Ego ) represents the coordinates of the vehicle's geometric center.

[0133] If the calculation result of the monitoring trigger condition is 1, it means that this situation needs to be monitored at this time, and the corresponding right-of-way violation monitoring module starts working and proceeds to step six; if the calculation result of the monitoring trigger condition is 0, it means that this regulation does not need to be monitored at this time, and step five is repeated.

[0134] Step Six: Right-of-way violation monitoring module receives data. When the calculation result of the monitoring trigger condition for a certain situation in Step Five is 1, the right-of-way violation monitoring module for that situation begins to receive the information from Step One to Step Three from the CAN bus.

[0135] Step 7: The violation monitoring module performs right-of-way violation judgment. Input the vehicle's decision information, map information, traffic light status information, and optimal virtual lane information. Based on different situations, it first generates the judgment area (JudgArea). k Each virtual stop line corresponds to a judgment area, and is generated as follows:

[0136] If Decision Ego =tl, take the opposite entrance lane R o .L k The intersection point of the right boundary of the optimal virtual straight lane and the right boundary of the optimal virtual lane for the vehicle is P. jarf_k , through P j xrf_k To R o .L k The optimal virtual straight lane has a perpendicular line drawn from its left boundary, and the point where the perpendicular line intersects it is set as P. jalf_k Then determine the region JudgArea k It can be represented as R o .L k The area enclosed by four points within the optimal virtual straight lane range, namely:

[0137] JudgArea k =[P jarf_k P jalf_k , (R o .L k .X lf ,R o .L k .Y lf ), (R o .L k .X rf ,R o .L k .Y rf )], k = 1, 2…N oen

[0138] The corresponding right-of-way determination line is called JudgLine. k =(P jarf_k P jalf_k ).

[0139] If Decisin Ego =tr∧TrafficLight=1, take the left entrance lane R l ·L k The intersection point of the right boundary of the optimal virtual straight lane and the left boundary of the optimal virtual lane for the vehicle is P. jarf_k , through P jarf_k To R l .L k The optimal virtual straight lane has a perpendicular line drawn from its left boundary, and the point where the perpendicular line intersects it is set as P. jalf_k Then determine the region JudgArea k It can be represented as R l .L k The area enclosed by four points within the optimal virtual straight lane range, namely:

[0140] JudgArea k =[P jarf_k P jalf_k , (R l .L k .X lf .R l .L k .Ylf ), (R l .L k .X rf .R l .L k .Y rf )], k = j, j+1...N len

[0141] The corresponding right-of-way determination line is called JudgLine. k =(P jarf_k P jalf_k ).

[0142] If Decision Ego =tr∧TrafficLight=2, take the opposite entrance lane R o .L k The intersection point of the right boundary of the optimal virtual left-turn lane and the left boundary of the optimal virtual lane for the vehicle is P. jarf_k , through P jarf_k To R o .L k The optimal virtual left-turn lane has a perpendicular line drawn from its left boundary, and the point where the perpendicular line intersects it is set as P. jalf_k Then determine the region JudgArea k It can be represented as R o .L k The area enclosed by four points within the optimal virtual left-turn lane range, namely:

[0143] JudgArea k =[P jarf_k P jalf_k , (R o .L k .X lf R o .L k .Y lf ), (R o .L k .X rf R o .L k .Y rf )], k = j, j+1...N oen

[0144] The corresponding right-of-way determination line is called JudgLine. k =(P jarf_k P jalf_k ).

[0145] After generating the judgment area and right-of-way judgment line, input the vehicle status information and other vehicle status information, and calculate whether the vehicle obstructs the passage of vehicles with high right-of-way. The calculation method is as follows:

[0146] If Decision Ego =tl, triggering the judgment within the time limit, the vehicle is perpendicular to Vstopline. k When they intersect, let Vstopline be the line from the midpoint of the rear end of the vehicle along the optimal virtual lane to the next virtual lane line. k+1 The distance is L c (If k = N) oen This means that the distance from the midpoint of the rear of the vehicle along the optimal virtual lane to the opposite entrance lane is... (Distance to the right boundary). Then, based on the longitudinal dynamics model, the distance to the current state after passing L can be calculated. c The required time is t c A violation of right-of-way can then be represented as:

[0147]

[0148]

[0149] Where distance(p1, y1) represents the distance from point p1 along the optimal virtual lane to line y1, ρ′=max(ρ-t, 0) is the remaining reaction time, ρ is the reaction time (taken as ρ=0.2s), and t is the time from the start of triggering to the current moment. acc This represents the maximum acceleration during travel at the intersection.

[0150] If Decision Ego =tr∧TrafficLight=1, triggering the judgment within the time limit, the vehicle is perpendicular to Vstopline k Intersection, similarly defined as Vstopline, which extends from the midpoint of the rear end of the vehicle along the optimal virtual lane to the next virtual lane line. k+1 The distance is L c (If k = N) len Then, it is the distance from the midpoint of the rear of the vehicle along the optimal virtual lane to the optimal exit lane (BEXLane). Ego (The distance to the rear boundary). Then, according to the longitudinal dynamics model, the distance to the current state after passing L is calculated. c The required time is t c A violation of right-of-way can then be represented as:

[0151]

[0152] Where k = j indicates that after passing the intersection, the vehicle will be in the same lane as the vehicle with the higher right-of-way, while k ≠ j indicates that the vehicle's trajectory intersects with the vehicle with the higher right-of-way, and they will not merge into the same lane after passing the intersection. gap JudgLine j To the best exit lane BEXLane Ego The distance from the back boundary; dRSS This indicates that after passing through an intersection, a reasonable RSS distance must still be maintained from vehicles with high right-of-way. The specific expression is:

[0153]

[0154] Among them, vx Ego_c Based on the longitudinal dynamics model, the current state of the vehicle passing through L is obtained. c The speed at time a max，brake To reduce the vehicle's automatic emergency braking (AEB) speed, a min，brake For vehicles with high right-of-way, take the minimum emergency braking deceleration.

[0155] If Decision Ego =tr∧TrafficLight=2, triggering the judgment within the time limit, the vehicle is perpendicular to Vstopline k Intersection, similarly defined as Vstopline, which extends from the midpoint of the rear end of the vehicle along the optimal virtual lane to the next virtual lane line. k+1 The distance is L c (If k = N) len Then, it is the distance from the midpoint of the rear of the vehicle along the optimal virtual lane to the optimal exit lane (BEXLane). Ego (The distance to the rear boundary). Then, according to the longitudinal dynamics model, the distance to the current state after passing L is calculated. c The required time is t c A violation of right-of-way can then be represented as:

[0156]

[0157] Based on the right-of-way violation calculation formulas for the three scenarios, calculate the right-of-way violation result. If the output is 1, it indicates that the right-of-way rule has been violated at the current moment. If the output is 0, it indicates that the right-of-way rule has not been violated at the current moment.

[0158] Step 8: The monitoring trigger module determines whether monitoring has ended. After Step 7 outputs a set of results, the monitoring trigger module will again determine whether the monitoring of this regulation should end at the current moment. If the exit condition is not met, it means that the regulation still needs to be monitored, and Steps 4 to 7 need to be repeated; if the exit condition is met, it means that the monitoring of this regulation has ended, and the right-of-way violation monitoring module stops working.

[0159] The preferred embodiments of the present invention have been described in detail above. However, the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the present invention.

[0160] Many other changes and modifications can be made without departing from the concept and scope of this invention. It should be understood that this invention is not limited to the specific embodiments, and the scope of this invention is defined by the appended claims.

Claims

1. A vehicle-side real-time right-of-way monitoring system for autonomous vehicles based on RSS, characterized in that, The system includes: a perception input module, a high-precision map module, a decision module, a CAN bus, a monitoring trigger judgment module, and a violation monitoring module; The perception input module includes: an on-board camera, radar, and a vehicle status sensor. The on-board camera and radar are used to perceive the status of other vehicles in the surrounding environment, and the vehicle status sensor is used to acquire the vehicle's position, speed, acceleration, and heading angle information in the global coordinate system. The high-precision map module is used to provide coordinate information of the static components of the intersection, including the endpoint coordinates of the stop line in the global coordinate system, the ID number of each lane and the coordinates of the boundary vertices, and the directional indicator marking information. The decision module is used to provide behavioral decisions for the vehicle when passing through an intersection, which are divided into left turn, right turn, and straight ahead. It is also used to calculate the time it takes for the vehicle to pass through a certain position. Once the vehicle enters the preset operating domain, the high-precision map inputs static intersection information into the CAN bus. The perception input module acquires the status information of other vehicles in real time and transmits it to the CAN bus. The decision module inputs the vehicle's decision information into the CAN bus. The monitoring trigger judgment module reads the input information from the CAN bus in real time and judges the trigger conditions. When the trigger condition of a certain rule is met, the violation monitoring module of the corresponding rule receives the input information from the CAN bus, monitors the vehicle's right-of-way violation in real time, and outputs the monitoring results to the CAN bus. The preset operating domain ODD is a polygonal area enclosed by the extension of the stop line within the intersection. The system operation method includes: S1: The high-precision map module inputs the static component information of the intersection into the CAN bus. When a vehicle enters the ODD of the right-of-way monitoring system, the high-precision map module sends the static component information required for right-of-way monitoring to the CAN bus in a predefined protocol format. The static component information includes the endpoint coordinates of the stop line in the global coordinate system, the ID number and boundary vertex coordinates of each intersection lane, and the direction indicator marking information; the high-precision map information includes the ID of each branch intersection - The branch intersection IDs increment counter-clockwise; the number of entrance lanes at each branch intersection. and number of exit lanes Lanes entering the intersection area are defined as entrance lanes, and lanes leaving the intersection area are defined as exit lanes; lanes are classified as front, rear, left, and right based on their direction of travel, and the lane ID for each branch intersection is defined as the lane closest to the road centerline in the entrance direction. Increasing sequentially to the right The lane closest to the center line of the road in the exit direction is Decreasing sequentially to the right Each lane is considered as a quadrilateral consisting of four vertices, including the coordinates of the four vertices and the lane marking information. Each lane's information is a 1×11 array. Each digit represents, in the global coordinate system, the x-coordinate of the left front point, the x-coordinate of the left rear point, the x-coordinate of the right front point, the x-coordinate of the right rear point, the y-coordinate of the left front point, the y-coordinate of the left rear point, the y-coordinate of the right front point, the y-coordinate of the right rear point, the Boolean symbol for left turn, the Boolean symbol for straight ahead, and the Boolean symbol for right turn; stop line A 1×4 array consisting of the coordinates of two vertices. Each element represents the left and right x-coordinates, left and right y-coordinates of the stop line in the global coordinate system, respectively. The left and right sides of the stop line are defined the same as the corresponding entrance lanes. Example of information representation: The number of entrance lanes with branch intersection id 1 is represented as... The lane with branch intersection ID 2 and lane ID 1 is represented as follows: The ordinate to the right of the stop line at the branch intersection with id 1 is represented as: ; S2: The perception input module acquires information about the vehicle itself and other vehicles, and obtains the vehicle's location information through the vehicle positioning device combined with the high-precision map module. The ordinate in the global coordinate system x-axis Longitudinal velocity Self-driving heading angle and the lane that the vehicle belongs to when entering the intersection. The longitudinal acceleration of the vehicle can be obtained through the inertial measurement unit (IMU). Cameras and radar detect other vehicles in the environment surrounding the vehicle. Information and traffic light status, other vehicle information including the horizontal coordinate of the other vehicle in the global coordinate system. y-axis The length of his car ,width The longitudinal speed of his vehicle Heading angle information Turn signal status and the lane his vehicle was in when it entered the intersection. The traffic light status is It includes three states: red light 0, green light 1, and yellow light 2; the sensing input module sends the acquired information to the CAN bus using a predefined bus protocol; S3: The decision-making module provides vehicle decision-making information. Before entering the intersection, the vehicle has already decided on the global path. There are three types: left turn ,straight and turn right When a vehicle enters the right-of-way monitoring system's ODD, the decision module sends the decision information to the CAN bus using a predefined bus protocol. S4: The right-of-way monitoring trigger judgment module reads the information from steps one to three above from the CAN bus in real time; S5: The right-of-way monitoring trigger judgment module judges different traffic light states and decision information, which will trigger right-of-way monitoring in different situations. There are three situations in which a vehicle needs to give up its right-of-way:

1. When a vehicle turns left, its right-of-way is lower than that of oncoming vehicles going straight; 2. When a vehicle turns right on a green light, its right-of-way is lower than that of an oncoming vehicle turning left; 3. When a vehicle turns right at a red light, its right-of-way is lower than that of vehicles going straight on the left; When autonomous vehicles navigate within an intersection, they follow an optimal circular curve. First, based on the vehicle's initial lane upon entering the intersection and its own decision-making process, an optimal virtual lane is generated. The generation method is as follows: Enter the lane your vehicle belongs to when entering the intersection. And map information and vehicle decision-making information The optimal exit lane and the unrecommended exit lane are calculated as follows: set up The optimal exit lane is: ； After obtaining the optimal exit lane, enter the vehicle's entry lane. Optimal Exit Lane With map information, the vertices of the optimal virtual lane can be obtained: The left and right rear vertices of the optimal virtual lane are the entrance lanes. The left and right front vertices: Left front vertex: ; Right front vertex: ; The optimal virtual lane's left front vertex is The right front vertex is After obtaining the vertices of the optimal virtual lane, input the coordinates of these vertices and decision information to generate the optimal virtual lane boundary. If the vehicle decides to turn left, then... , Let the left and right front vertices of the optimal virtual lane calculated above be respectively... Substituting these values ​​into the following formula, we can obtain the centers of the left and right circular boundaries of the virtual lane. : ； If the vehicle decides to turn right, then take... , The calculation method is the same as the above formula; After obtaining the center of the circular boundary of each virtual lane, let the center be... Then the radius of the circular boundary of each virtual lane is If the boundary of the circular curve is represented as [start point, (center, radius), end point], then the optimal virtual lane area can be represented as: ； in, , These represent the left and right boundary centers of the optimal virtual lane, respectively. , These represent the left and right boundary radii of the optimal virtual lane, respectively. If a vehicle decides to proceed straight, the virtual lane is a quadrilateral area, which can be directly represented as: ； After obtaining the optimal virtual lane area, K virtual stop lines will be generated within the lane. The number of K is equal to the number of other high-rights optimal virtual lanes intersecting with the optimal virtual lane; the generation method is as follows: Still set The branch intersection on the left side of the vehicle entrance lane is The opposite branch intersection is The right-hand branch intersection is ; This refers to the number of lanes at the entrance of the left-hand branch intersection; This refers to the number of lanes at the entrance of the opposite branch intersection; This refers to the number of exit lanes at the left-hand branch intersection; like , For the opposite entrance lane A point on the left boundary of the optimal virtual straight lane, where the distance to the left and right boundaries of the optimal virtual lane is equal; the virtual stop line can be represented as: ； in, Point With point The equation of the line in question; like , Left-side entrance lane A point on the right boundary of the optimal virtual straight lane, where the distance to the left and right boundaries of the optimal virtual lane is equal; the virtual stop line can be represented as: ； Calculate the virtual stop line Then, based on the virtual stop line equation, input the ordinate of the vehicle in the global coordinate system. x-axis Heading angle and the longitudinal speed of the vehicle Decision information of the vehicle and traffic light status information If this occurs, right-of-way monitoring will be triggered, and the judgment formula is as follows: ； in, , and These represent three different road right-of-way monitoring situations; An atomic proposition states whether two elements have an overlapping region. and If there is overlap, output 1; otherwise, output 0. and It can be a line segment or a planar region. The formula for calculating atomic propositions is: ; Indicates the origin of the point and A line segment formed by connecting lines; Indicates the coordinates of the midpoint of the front of the vehicle; The length of the vehicle; Represents the coordinates of the vehicle's geometric center; If the calculation result of the monitoring trigger condition is 1, it means that this situation needs to be monitored at this time, and the corresponding right-of-way violation monitoring module starts working, proceeding to step six; if the calculation result of the monitoring trigger condition is 0, it means that this regulation does not need to be monitored at this time, and step five is repeated. S6: The right-of-way violation monitoring module receives data. When the calculation result of the monitoring trigger condition for a certain situation in step five is 1, the right-of-way violation monitoring module for that situation starts to receive the information from step one to step three from the CAN bus. S7: The violation monitoring module performs right-of-way violation judgments. It inputs the vehicle's decision information, map information, traffic light status information, and optimal virtual lane information. Based on different situations, it first generates a judgment area. Each virtual stop line corresponds to a judgment area, and is generated as follows: like Take the opposite entrance lane The intersection of the right boundary of the optimal virtual straight lane and the right boundary of the vehicle's optimal virtual lane is ,Pass Towards The optimal virtual straight lane's left boundary is perpendicular to the intersection point of the line and set as... Then determine the region It can be represented as The area enclosed by four points within the optimal virtual straight lane range, namely: ； The corresponding right-of-way determination line is ; like Take the left-hand entrance lane The intersection of the right boundary of the optimal virtual straight lane and the left boundary of the optimal virtual lane for the vehicle is ,Pass Towards The optimal virtual straight lane's left boundary is perpendicular to the intersection point of the line and set as... Then determine the region It can be represented as The area enclosed by four points within the optimal virtual straight lane range, namely: ； The corresponding right-of-way determination line is ; like Take the opposite entrance lane The intersection of the right boundary of the optimal virtual left-turn lane and the left boundary of the optimal virtual lane for the vehicle is ,Pass Towards The optimal virtual left-turn lane has a perpendicular line drawn from its left boundary, and the point where the perpendicular line intersects it is set as... Then determine the region It can be represented as The area enclosed by four points within the optimal virtual left-turn lane range, namely: ； The corresponding right-of-way determination line is ; After generating the judgment area and right-of-way judgment line, input the vehicle status information and other vehicle status information, and calculate whether the vehicle obstructs the passage of vehicles with high right-of-way. The calculation method is as follows: like Within the time limit for triggering the judgment, the vehicle is currently in contact with... When they intersect, let the midpoint of the rear end of the vehicle be the line from the optimal virtual lane to the next virtual lane. The distance is ,like This means that the distance from the midpoint of the rear of the vehicle along the optimal virtual lane to the opposite entrance lane is... The distance to the right boundary; then, based on the longitudinal dynamics model, the distance to the current state of travel can be calculated. The required time is A violation of right-of-way can be represented as: ； in, Point From the best virtual lane to a straight line distance, The remaining reaction time, For the reaction time, take , To trigger the time from the start to the current moment, This represents the maximum acceleration during travel at the intersection. like Within the time limit for triggering the judgment, the vehicle is currently in contact with... Intersecting, similarly, the line from the midpoint of the rear end of the vehicle along the optimal virtual lane to the next virtual lane is defined. The distance is ,like This means that the distance from the midpoint of the rear of the vehicle along the optimal virtual lane to the optimal exit lane is... The distance to the rear boundary; then, according to the longitudinal dynamics model, the distance to the current state of the vehicle is calculated. The required time is A violation of right-of-way can be represented as: ； in, This indicates that after passing the intersection, the vehicle will be in the same lane as vehicles with higher right-of-way, and This indicates that the vehicle's trajectory intersects with the high right-of-way trajectory, and it will not merge into the same lane after passing the intersection; express To the best exit lane The distance from the back boundary; This indicates that after passing through an intersection, a reasonable RSS distance must still be maintained from vehicles with high right-of-way. The specific expression is: ； in, Based on the longitudinal dynamics model, the current state of the vehicle is obtained. The speed of time, The automatic emergency braking (AEB) deceleration of the vehicle. For vehicles with high right-of-way, take the minimum emergency braking deceleration. ; like Within the time limit for triggering the judgment, the vehicle is currently in contact with... Intersecting, similarly, the line from the midpoint of the rear end of the vehicle along the optimal virtual lane to the next virtual lane is defined. The distance is ,like This means that the distance from the midpoint of the rear of the vehicle along the optimal virtual lane to the optimal exit lane is... The distance to the rear boundary; then, according to the longitudinal dynamics model, the distance to the current state of the vehicle is calculated. The required time is A violation of right-of-way can be represented as: ； Based on the right-of-way violation calculation formulas for the three scenarios, calculate the right-of-way violation result. If the output is 1, it indicates that the right-of-way rule has been violated at the current moment; if the output is 0, it indicates that the right-of-way rule has not been violated at the current moment. S8: The monitoring trigger module determines whether the monitoring has ended; after step seven outputs a set of results, the monitoring trigger module will make another judgment to determine whether the monitoring of the regulation should end at the current moment. If the exit condition is not met, it means that the regulation still needs to be monitored, and steps four to seven need to be repeated; if the exit condition is met, it means that the monitoring of the regulation has ended and the right-of-way violation monitoring module stops working.

2. The vehicle-side real-time right-of-way monitoring system for autonomous vehicles based on RSS according to claim 1, characterized in that, The status of other vehicles includes: their position, velocity, acceleration, heading angle, turn signal status, and geometric dimensions in the global coordinate system.

3. The vehicle-side real-time right-of-way monitoring system for autonomous vehicles based on RSS according to claim 1, characterized in that, In step one, the left-turn Boolean symbol is: 1 means that the lane can turn left, and 0 means that it cannot turn left; the straight-ahead Boolean symbol is: 1 means that the lane can go straight, and 0 means that it cannot go straight; the right-turn Boolean symbol is: 1 means that the lane can turn right, and 0 means that it cannot turn right.

Images