CN116453341B

CN116453341B - Method, device and system for monitoring real-time violation at vehicle end of automatic driving vehicle

Info

Publication number: CN116453341B
Application number: CN202310446888.0A
Authority: CN
Inventors: 王长君; 赵成祥; 于文浩; 王伟达; 王红; 周文辉; 赵光明; 马明月
Original assignee: Road Traffic Safety Research Center Ministry Of Public Security Of People's Republic Of China; Tsinghua University; Beijing Institute of Technology BIT
Current assignee: Road Traffic Safety Research Center Ministry Of Public Security Of People's Republic Of China; Tsinghua University; Beijing Institute of Technology BIT
Priority date: 2023-04-24
Filing date: 2023-04-24
Publication date: 2024-04-09
Anticipated expiration: 2043-04-24
Also published as: CN116453341A

Abstract

The invention discloses a real-time violation monitoring method, device and system for a vehicle end of an automatic driving vehicle, belonging to the field of road traffic monitoring; firstly, only static component element information, vehicle state information and environment information of the crossroad are required to be received for monitoring the violations, and no requirement is required for whether monitoring equipment is installed at the crossroad or not, so that the violations can be monitored even if the monitoring equipment is not arranged at the crossroad; and secondly, judging whether the triggering condition of any rule is met or not, and acquiring corresponding data to judge the rule violation when the triggering condition is met, so that the data calculation amount can be greatly reduced. Finally, judging according to the self state information and static component element information of the crossroad, judging according to environment information such as other vehicle information, pedestrian information and signal lamp state, detecting simple rules such as speed limit violations, red light running violations and the like, and monitoring road right rules and pedestrian lamp rules. The method has the advantages of real-time monitoring, accurate monitoring and multiple monitoring rules.

Description

Method, device and system for monitoring real-time violation at vehicle end of automatic driving vehicle

Technical Field

The invention relates to the field of road traffic monitoring, in particular to a real-time violation monitoring method, device and system for a vehicle end of an automatic driving vehicle.

Background

With the continuous development of artificial intelligence, an automatic driving automobile is gradually landed, and in a long period of time in the future, the automatic driving and human driving share a road, so that intervention measures are needed to improve the legal awareness of the automatic driving, so that the safety of human beings on the automatic driving automobile can be improved, and dangerous behaviors caused by illegal operation and excessive reaction of the automatic driving automobile are avoided. The existing automatic driving automobile behavior decision only considers a part of simple regulation regulations, and the considered emphasis is only a simple safety layer, and whether the planned behavior is compliant or not is illegal, so that misunderstanding and danger of human beings are caused is not considered. On the other hand, legal compliance in monitoring vehicle behavior may provide strong evidence for the traceability of responsibility for traffic accidents. Whereas current road traffic regulations are written in human-oriented natural language, their ambiguity and non-digitality make it difficult for the autonomous car to understand these regulations, so that it is necessary to digitize these regulations so that the autonomous car can understand and comply with the traffic regulations correctly. Meanwhile, the behavior of the automatic driving automobile is accurately monitored, so that compliance decisions can be made, and responsibility tracing work after accidents is also facilitated.

The current rule violation monitoring only depends on road side equipment, and only can monitor partial simple and clear rules, such as speed limit violations and red light running violations. And the intersection without road-end equipment is not monitored from the past.

Disclosure of Invention

In order to overcome the defects of the prior art, the invention provides a real-time violation monitoring method, device and system for an automatic driving automobile, which are used for solving the problem that the violation monitoring of the current road regulations only depends on road side equipment, and only can monitor part of simple and clear rules, such as speed limit violations and red light running violations. And the intersection without road end equipment is more free from monitoring.

The technical scheme adopted for solving the technical problems is as follows:

in a first aspect, a real-time violation monitoring method for a vehicle end of an automatic driving vehicle is provided, and the method is applied to an intersection, and comprises the following steps:

receiving static component element information of an intersection, vehicle state information and environment information, wherein the static component element information of the intersection comprises: each branch road intersection id, the number of entrance lanes and the number of exit lanes of each branch road intersection, marking information of each lane, stop lines and crosswalk coordinates; the own vehicle state information includes: the method comprises the steps of self-vehicle decision, the coordinates of the self-vehicle in a global coordinate system, the course angle of the self-vehicle, the lane to which the self-vehicle belongs and the longitudinal speed of the self-vehicle; the environment information comprises other vehicle information, pedestrian information and signal lamp states, and the other vehicle information comprises coordinates of the other vehicle in a global coordinate system, the length and the width of the other vehicle, the longitudinal speed of the other vehicle, course angle information and lanes to which the other vehicle belongs; the pedestrian information comprises coordinates and speed vectors of pedestrians in a global coordinate system; the signal lamp state comprises a red lamp, a green lamp and a yellow lamp;

If the static component element information of the crossroad, the vehicle state information and the environment information meet the preset triggering conditions of any regulations, the data corresponding to the regulations are obtained to judge whether the regulations are illegal.

Further, if the static component element information of the intersection, the vehicle state information and the environment information meet a trigger condition preset by any rule, acquiring data corresponding to the rule to judge whether the rule is violated, including:

when judging that the vehicle geometric center passes over the corresponding stop line according to the static component element information and the vehicle state information of the crossroad, meeting the triggering condition of traffic light passing rules;

acquiring decision information of the own vehicle, coordinates of the own vehicle under a global coordinate system, a heading angle of the own vehicle, a lane to which the own vehicle belongs and a signal lamp state, and calculating whether the own vehicle meets traffic light passing rules; the traffic light passing rule comprises: whether traffic light rules are complied with or whether the vehicle is traveling in the allowed direction of the lane marking.

When judging that the geometric center of the vehicle is positioned in an area surrounded by straight lines where four parking lines are positioned according to the static component element information and the vehicle state information of the crossroad, the rule triggering condition of the driving route is met;

generating an optimal virtual lane and an un-recommended virtual lane according to a lane to which a vehicle belongs and a vehicle decision, wherein the un-recommended virtual lane is an adjacent exit lane of the optimal virtual lane;

calculating the intersection points of the lanes to which the own vehicle belongs, the good exit lane and the non-recommended virtual lane and the outer boundary of the crosswalk;

generating an optimal virtual lane boundary and an un-recommended virtual lane boundary according to the intersection point and the self-vehicle decision;

judging whether the vehicle coordinate is in the optimal virtual lane boundary and the non-recommended virtual lane boundary or not according to the vehicle coordinate;

if yes, judging that the self-vehicle driving route is not illegal; and if not, judging that the self-vehicle driving route is illegal.

Further, whether the self-vehicle driving route is within the optimal virtual lane boundary is judged according to the self-vehicle coordinates, if not, the data are recorded for self-vehicle learning, and the self-vehicle driving route selection capability is improved.

When judging that the geometric center of the vehicle is positioned in an area surrounded by straight lines where four parking lines are positioned according to the static component element information and the vehicle state information of the crossroad, meeting the road right rule triggering condition;

generating a virtual stop line and a highway right monitoring area according to the optimal virtual lane, the vehicle state information and the static component element information of the crossroad and the signal lamp state;

when the own vehicle passes through the virtual stop line, judging whether the other vehicle exists in the highway right monitoring area;

if yes, judging that the own vehicle violates the road right rule; if the road right rule is not violated, judging that the own vehicle does not violate the road right rule.

Further, the generating the virtual stop line and the high road right monitoring area according to the optimal virtual lane and the self-vehicle state information and the static component element information of the crossroad comprises the following steps:

determining a lane of the highway right according to the signal lamp state and the self-vehicle decision, and obtaining a highway right monitoring area according to the lane of the highway right;

and generating a virtual stop line according to the lane to which the own vehicle belongs and the optimal virtual lane.

When judging that the front end of the vehicle enters a crosswalk area according to the static component element information of the crossroad and the vehicle state information, and the vehicle geometric center does not enter the crosswalk area yet, meeting the rule triggering condition of not obstructing pedestrians;

dividing the crosswalk area into a plurality of crosswalk subareas according to lanes;

when a self-vehicle enters any crosswalk subarea, judging whether pedestrians exist in the crosswalk subarea or not;

if so, judging that the own vehicle hinders the pedestrians.

Further, if no pedestrian exists in the crosswalk subarea, judging whether a pedestrian travelling towards the subarea where the own vehicle is located exists in an adjacent subarea of the crosswalk subarea;

if the vehicle is present, judging that the vehicle is in the way of pedestrians; if the vehicle is not present, the vehicle is judged not to obstruct the pedestrian.

In a second aspect, there is provided a real-time violation monitoring device at a vehicle end of an automatically driven vehicle, for use at an intersection, the device comprising:

the information acquisition unit is used for receiving the static component element information of the crossroad, the self-vehicle state information and the environment information, and the static component element information of the crossroad comprises: each branch road intersection id, the number of entrance lanes and the number of exit lanes of each branch road intersection, marking information of each lane, stop lines and crosswalk coordinates; the own vehicle state information includes: the method comprises the steps of self-vehicle decision, the coordinates of the self-vehicle in a global coordinate system, the course angle of the self-vehicle, the lane to which the self-vehicle belongs and the longitudinal speed of the self-vehicle; the environment information comprises other vehicle information, pedestrian information and signal lamp states, and the other vehicle information comprises coordinates of the other vehicle in a global coordinate system, the length and the width of the other vehicle, the longitudinal speed of the other vehicle, course angle information and lanes to which the other vehicle belongs; the pedestrian information comprises coordinates and speed vectors of pedestrians in a global coordinate system; the signal lamp state comprises a red lamp, a green lamp and a yellow lamp;

And the violation judging unit is used for acquiring data corresponding to any regulations and judging whether the rule is violated if the static component element information of the crossroad, the vehicle state information and the environment information meet the preset triggering conditions of any regulations.

In a third aspect, a real-time violation monitoring system for a vehicle end of an autopilot vehicle is provided, comprising:

a processor;

a memory for storing the processor-executable instructions;

the processor is configured for performing the method of any one of the technical solutions provided in the first aspect.

The beneficial effects are that:

the technical scheme of the application provides a real-time violation monitoring method, device and system for a vehicle end of an automatic driving vehicle, firstly, the violation monitoring only needs to receive static component element information, vehicle state information and environment information of an intersection, and no requirement is made on whether monitoring equipment is installed at the intersection, so that the violation monitoring can be realized even if the monitoring equipment is not arranged at the intersection; and secondly, judging whether the triggering condition of any rule is met or not, and acquiring corresponding data to judge the rule violation when the triggering condition is met, so that the data calculation amount can be greatly reduced. Finally, judging according to the self state information and static component element information of the crossroad, judging according to environment information such as other vehicle information, pedestrian information and signal lamp state, detecting simple rules such as speed limit violations, red light running violations and the like, and monitoring road right rules and pedestrian lamp rules. The method has the advantages of real-time monitoring, accurate monitoring and multiple monitoring rules.

Drawings

In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

FIG. 1 is a flow chart of a real-time violation monitoring method at the vehicle end of an automatic driving vehicle provided by an embodiment of the invention;

fig. 2 is a schematic structural diagram of a real-time violation monitoring device at a vehicle end of an automatic driving vehicle according to an embodiment of the present invention;

FIG. 3 is a block diagram of a real-time violation monitoring system at the vehicle end of an automatic driving vehicle according to an embodiment of the present invention;

fig. 4 is a schematic diagram of map information of an intersection according to an embodiment of the present invention.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the present application more apparent, the technical solutions of the present invention will be described in detail with reference to the accompanying drawings and examples. It will be apparent that the described embodiments are only some, but not all, of the embodiments of the present application. All other embodiments, based on the examples herein, which are within the scope of the protection sought by those of ordinary skill in the art without undue effort, are intended to be encompassed by the present application.

Referring to fig. 1, the embodiment of the invention provides a real-time monitoring method for real-time violations at the vehicle end of an automatic driving vehicle, which is applied to an intersection, and comprises the following steps:

s11: receiving static component element information of an intersection, vehicle state information and environment information, wherein the static component element information of the intersection comprises: each branch road intersection id, the number of entrance lanes and the number of exit lanes of each branch road intersection, marking information of each lane, stop lines and crosswalk coordinates; the own vehicle state information includes: the method comprises the steps of self-vehicle decision, the coordinates of the self-vehicle in a global coordinate system, the course angle of the self-vehicle, the lane to which the self-vehicle belongs and the longitudinal speed of the self-vehicle; the environment information comprises other vehicle information, pedestrian information and signal lamp states, wherein the other vehicle information comprises coordinates of the other vehicle in a global coordinate system, the length and the width of the other vehicle, the longitudinal speed of the other vehicle, course angle information and lanes to which the other vehicle belongs; the pedestrian information comprises coordinates and speed vectors of pedestrians in a global coordinate system; the signal lamp state comprises a red lamp, a green lamp and a yellow lamp;

s12: if the static component element information of the crossroad, the vehicle state information and the environment information meet the preset triggering conditions of any regulations, the data corresponding to the regulations are obtained to judge whether the regulations are violated.

According to the real-time violation monitoring method for the vehicle end of the automatic driving vehicle, which is provided by the embodiment of the invention, firstly, the violation monitoring only needs to receive the static component element information, the vehicle state information and the environment information of the crossroad, and no requirement is required for whether monitoring equipment is installed at the crossroad or not, so that the violation monitoring can be realized even if the monitoring equipment is not arranged at the crossroad; and secondly, judging whether the triggering condition of any rule is met or not, and acquiring corresponding data to judge the rule violation when the triggering condition is met, so that the data calculation amount can be greatly reduced. Finally, judging according to the self state information and static component element information of the crossroad, judging according to environment information such as other vehicle information, pedestrian information and signal lamp state, detecting simple rules such as speed limit violations, red light running violations and the like, and monitoring road right rules and pedestrian lamp rules. The method has the advantages of real-time monitoring, accurate monitoring and multiple monitoring rules.

In a second embodiment, as a supplementary explanation of the above embodiment, if the static component element information of the intersection, the vehicle state information and the environment information meet a trigger condition preset by any rule, the acquiring data corresponding to the rule to determine whether the rule is violated includes: when judging that the vehicle geometric center passes over the corresponding stop line according to the static component element information and the vehicle state information of the crossroad, the traffic light passing rule triggering condition is met; acquiring decision information of the own vehicle, coordinates of the own vehicle under a global coordinate system, a heading angle of the own vehicle, a lane to which the own vehicle belongs and a signal lamp state, and calculating whether the own vehicle meets traffic light passing rules; the traffic light passing rule comprises: whether traffic light rules are complied with or whether the vehicle is traveling in the allowed direction of the lane marking.

Optionally, if the static component element information of the intersection, the vehicle state information and the environment information meet a trigger condition preset by any rule, the data corresponding to the rule is obtained to judge whether the rule is violated, including: when judging that the geometric center of the vehicle is positioned in an area surrounded by straight lines where four parking lines are positioned according to static component element information and vehicle state information of the crossroad, the driving route rule triggering condition is met; generating an optimal virtual lane and an un-recommended virtual lane according to the lane to which the own vehicle belongs and an own vehicle decision, wherein the un-recommended virtual lane is an adjacent exit lane of the optimal virtual lane; calculating the intersection points of the lanes to which the own vehicle belongs, the good exit lane and the non-recommended virtual lane and the outer boundary of the crosswalk; generating an optimal virtual lane boundary and a non-recommended virtual lane boundary according to the intersection point and the vehicle decision; judging whether the vehicle coordinate is in the optimal virtual lane boundary and the non-recommended virtual lane boundary or not according to the vehicle coordinate; if yes, judging that the self-vehicle driving route is not illegal; if not, the self-vehicle driving route violation is judged. And judging whether the self-vehicle driving route is within the optimal virtual lane boundary according to the self-vehicle coordinates, and if not, recording data for self-vehicle learning for improving the self-vehicle driving route selection capability.

Optionally, if the static component element information of the intersection, the vehicle state information and the environment information meet a trigger condition preset by any rule, the data corresponding to the rule is obtained to judge whether the rule is violated, including: when judging that the geometric center of the vehicle is positioned in an area surrounded by straight lines where four parking lines are positioned according to static component element information and vehicle state information of the crossroad, meeting the road right rule triggering condition; generating a virtual stop line and a highway right monitoring area according to the optimal virtual lane, the vehicle state information and the static component element information of the crossroad and the signal lamp state; when the own vehicle passes through the virtual stop line, judging whether other vehicles exist in the highway right monitoring area; if yes, judging that the own vehicle violates the road right rule; if the road right rule is not violated, judging that the own vehicle does not violate the road right rule.

Wherein, generating virtual stop line and high road right monitoring area according to the best virtual lane and self-vehicle state information and static component element information of the crossroad, comprising: determining a lane of the highway right according to the signal lamp state and the self-vehicle decision, and obtaining a highway right monitoring area according to the lane of the highway right; and generating a virtual stop line according to the lane to which the own vehicle belongs and the optimal virtual lane.

Optionally, if the static component element information of the intersection, the vehicle state information and the environment information meet a trigger condition preset by any rule, the data corresponding to the rule is obtained to judge whether the rule is violated, including: when judging that the front end of the vehicle enters a pedestrian crosswalk area according to static component element information and vehicle state information of the crosswalk, and the geometric center of the vehicle does not enter the pedestrian crosswalk area yet, meeting the rule triggering condition of not obstructing pedestrians; dividing the crosswalk area into a plurality of crosswalk subareas according to lanes; when the self-vehicle enters any crosswalk subarea, judging whether pedestrians exist in the crosswalk subarea or not; if so, judging that the own vehicle hinders the pedestrians. If no pedestrians exist in the pedestrian crossing subarea, judging whether pedestrians which move towards the subarea where the self-vehicle is located exist in the adjacent subareas of the pedestrian crossing subarea; if the vehicle is present, judging that the vehicle is in the way of pedestrians; if the vehicle is not present, the vehicle is judged not to obstruct the pedestrian.

In a third embodiment, the present invention provides a real-time violation monitoring device for an automatic driving automobile, which is applied to an intersection, as shown in fig. 2, and the device includes:

an information acquisition unit 21 for receiving intersection static constituent element information including: each branch road intersection id, the number of entrance lanes and the number of exit lanes of each branch road intersection, marking information of each lane, stop lines and crosswalk coordinates; the own vehicle state information includes: the method comprises the steps of self-vehicle decision, the coordinates of the self-vehicle in a global coordinate system, the course angle of the self-vehicle, the lane to which the self-vehicle belongs and the longitudinal speed of the self-vehicle; the environment information comprises other vehicle information, pedestrian information and signal lamp states, wherein the other vehicle information comprises coordinates of the other vehicle in a global coordinate system, the length and the width of the other vehicle, the longitudinal speed of the other vehicle, course angle information and lanes to which the other vehicle belongs; the pedestrian information comprises coordinates and speed vectors of pedestrians in a global coordinate system; the signal lamp state comprises a red lamp, a green lamp and a yellow lamp;

And the rule violation judging unit 22 is configured to obtain data corresponding to any rule and judge whether the rule is violated if the static component element information, the vehicle state information and the environment information of the intersection meet the trigger conditions preset by any rule.

In a fourth embodiment, the present invention provides a real-time violation monitoring system for a vehicle end of an automatic driving vehicle, including:

a processor;

a memory for storing processor-executable instructions;

the processor is configured to perform the real-time violation monitoring method for the vehicle end of the autonomous vehicle provided in the first embodiment or the second embodiment.

In order to further illustrate the scheme of the embodiment of the present invention, the embodiment of the present invention provides a specific system structure composition, as shown in fig. 3: the system mainly comprises a high-precision map, a perception input module, a decision module, a CAN bus and a violation monitoring module. The high-precision map is used for providing coordinate information of static constituent elements of the crossroad, and comprises endpoint coordinates of a stop line, boundary vertex coordinates of a crosswalk, id numbers of lanes of each crossroad, boundary vertex coordinates and direction indication marking information under a global coordinate system. The perception input module comprises a radar, a camera and other vehicle-mounted sensors and is used for acquiring other vehicles, pedestrians, traffic sign marks, signal lamp states and vehicle state information in the surrounding environment. The self-vehicle state information comprises the position of the self-vehicle in a global coordinate system, the longitudinal speed of the self-vehicle and course angle information; the information of the other vehicle comprises the position of the other vehicle in the global coordinate system, the shape and the geometry of the other vehicle and the speed and course angle information of the other vehicle; the pedestrian information includes the position and speed of the pedestrian. The decision module is used for providing behavior decision information of the own vehicle when passing through the intersection and is divided into left turn, right turn and straight run. The design operation domain (ODD) of the system is a road section range of 5m around the intersection. After the self-vehicle enters the designed operation domain, the high-precision map inputs intersection information into the CAN bus, the sensing input module acquires surrounding environment information in real time and transmits the information into the CAN bus, and the decision module inputs self-vehicle decision information into the CAN bus. The violation monitoring module is divided into a monitoring triggering module and a rule logic judging module, and each rule has an independent violation monitoring module. And the rule logic judgment module of the corresponding rule starts to receive the input information from the CAN bus, carries out rule violation monitoring on vehicle behaviors in real time, and outputs rule violation monitoring results at the current moment to the CAN bus when the trigger conditions of the certain rule are met.

The general steps of real-time violation monitoring of relevant road regulations in an intersection scene are as follows:

step one: and acquiring coordinate information of static constituent elements of the crossroad through the high-precision map. After a vehicle enters a design operation domain of the system, the high-precision map CAN send the coordinate information of the static constituent elements of the current intersection to the CAN bus by using a defined bus protocol. The static map information comprises end point coordinates of a stop line, boundary vertex coordinates of a crosswalk, id numbers and boundary vertex coordinates of each intersection lane and direction indication marking information under a global coordinate system.

Step two: the perception input module acquires surrounding environment information. The camera and the radar sense states of other vehicles, pedestrians, traffic sign marks and signal lamps in the surrounding environment of the own vehicle, and the other vehicle information comprises the position of the other vehicle in a global coordinate system, the shape and the geometry of the other vehicle and the speed and course angle information of the other vehicle; the pedestrian information includes the position and speed of the pedestrian. Other onboard sensors are responsible for obtaining the position, longitudinal speed and heading angle information of the own vehicle under the global coordinate system. The sensing input module sends the acquired information to the CAN bus according to a defined bus protocol.

Step three: the decision module provides vehicle decision information. Before the self-vehicle enters the intersection, the global path is decided in advance, the decision is divided into three types, namely left turn, straight turn and right turn, and after the decision is made by the decision module, the decision information is sent to the CAN bus by a defined bus protocol.

Step four: and monitoring the data receiving of the trigger module. The monitoring triggering module in each independent violation monitoring module reads information required by triggering judgment from the CAN bus in real time.

Step five: and (5) monitoring triggering condition judgment. The monitoring triggering module judges the monitoring triggering condition according to the information read in the step four, if the monitoring triggering condition of a certain rule is judged to be true, the rule is required to be monitored at the moment, an enabling signal is generated to be 1, and the corresponding rule logic judging module starts to work and enters the step six; if the monitoring trigger condition is judged to be false, an enabling signal is generated to be 0, and the fact that the rule is not required to be monitored at the moment is indicated, and the fifth step is repeated.

Step six: the legislation logic determination module receives the data. After the monitoring triggering condition of a certain rule outputs an enabling signal, a rule logic judgment module of the violation monitoring module starts to read information needed by logic judgment from the CAN bus.

Step seven: the rule logic determination module performs rule violation monitoring. The rule logic judging module carries out logic judgment according to the received information to judge whether the behavior of the own vehicle at the current moment meets the rule requirement, if so, a corresponding code of compliance of the rule is output, and the fact that the own vehicle does not violate the rule currently is indicated; if not, outputting the corresponding code of the rule violation, which indicates that the own vehicle is currently violating the rule.

Step eight: the monitoring triggering module judges whether the monitoring is finished or not. Step seven, after outputting a group of results, judging by the monitoring triggering module again to judge whether the monitoring of the regulation at the current moment should be ended or not, if the exiting condition is not met, the regulation still needs to be continuously monitored, and then the step four to the step seven need to be repeated; if the exit condition is met, the monitoring of the rule is finished, and the rule logic judging module stops working.

In the implementation of this patent, metric Temporal Logic (MTL) is used to express and determine logical relationships, where the MTL includes boolean operators and timing operators to express boolean and timing logical relationships between atomic propositions. The boolean operators used in this patent include: and (, V), not Equivalent->The timing operators include: once (ot) until (U), where ^t Indicating a certain moment in time or a certain time interval.

The traffic light passing rule, the driving route rule, the road right rule and the pedestrian rule related to the crossroad are used for describing the real-time violation monitoring system at the vehicle end in detail. Fig. 4 is a diagram showing an example of the map information structure of an intersection.

Step one: and acquiring coordinate information of static constituent elements of the cross road through the high-precision map. After the vehicle enters the designed operation domain of the system, the high-precision map CAN send the coordinate information of the static constituent elements of the current cross road to the CAN bus. The information transmitted therein is as shown in FIG. 3, including each of the branch road ids (R ₁ ，R ₂ ，R ₃ ，R ₄ ) The method comprises the steps of carrying out a first treatment on the surface of the Number of entrance lanes per branch road junction (N _en ) And number of exit lanes (N _ex ) (defining a lane entering an intersection area as an entrance lane and a lane leaving the intersection area as an exit lane); dividing the road travelling direction into front, back, left and right, defining the lane id of each branch road junction as the lane closest to the central line of the road in the entrance direction as L ₁ Sequentially increases to the right (L ₁ ，L ₂ …); the lane closest to the road center line in the exit direction is L _-1 Decreasing in order to the right (L _-1 ，L _-2 …)。

Each lane is regarded as a quadrangle consisting of four vertexes including four vertex coordinates and the marking information of the lane, and each lane is a 1×11 array ([ X) _lf ，X _lr ，X _rf ，X _rr ，Y _lf ，Y _lr ，Y _rf ，Y _rr ，TL，GS，TR]) Each bit is sequentially a left front point abscissa, a left rear point abscissa, a right front point abscissa, a right rear point abscissa, a left front point ordinate, a left rear point ordinate, a right front point ordinate, a right rear point ordinate, a left turn Boolean (1 is that the lane can turn left, 0 is that the lane cannot turn left), a straight-going Boolean (1 is that the lane can go straight, 0 is that the lane cannot go straight), a right turn Boolean (1 is that the lane can turn right, 0 is that the lane cannot turn right);

the left and right sides of the entrance lane are divided by the traveling direction of the stop line and the crosswalk, and the stop line(s) is a 1X 4 array ([ X) composed of two vertex coordinates _l ，X _r ，Y _l ，Y _r ]) Each bit is in turnThe left abscissa, the right abscissa, the left ordinate and the right ordinate of the stop line in the global coordinate system are defined as the same as the corresponding entrance lane; the left and right sides of the crosswalk are defined to be the same as the corresponding entrance lane, the boundary of the area of the crosswalk near the stop line is regarded as the outer boundary (zout), the boundary of the area parallel to the outer boundary is regarded as the inner boundary (zin), and each boundary is a 1×4 array ([ X) _l ，X _r ，Y _l ，Y _r ]) Each bit is in turn the left-hand abscissa, the right-hand abscissa, the left-hand ordinate and the right-hand ordinate of the outer or inner boundary of the crosswalk under the global coordinate system. Information represents an example: the number of entrance lanes with a branch road id of 1 is denoted as R ₁ .N _en The method comprises the steps of carrying out a first treatment on the surface of the The lane with the branch crossing id of 1 and the lane id of 2 is denoted as R ₁ .L ₂ The method comprises the steps of carrying out a first treatment on the surface of the The ordinate on the right side of the stop line with the branch intersection id of 2 is expressed as R ₂ ·s(4)。

Step two: the perception input module acquires surrounding environment information. The vertical coordinate X of the own vehicle (Ego) under the global coordinate system can be obtained by combining the GPS and the INS of the own vehicle with a high-precision map _Ego Y of abscissa _Ego Heading angle theta of bicycle _Ego Lane of the Lane to which the own vehicle belongs _Ego Longitudinal speed vx of the vehicle can be obtained by the IMU _Ego . The camera and the radar sense states of other vehicles (Tgt), pedestrians (Ped) and signal lamps in the surrounding environment of the own vehicle, and the other vehicle information comprises an abscissa X of the other vehicle in a global coordinate system _Tgt Ordinate Y _Tgt Length L of other vehicle _Tgt Width W _Tgt Longitudinal speed vx of other vehicle _Tgt Course angle information θ _Tgt Lane of the Lane to which the other vehicle belongs _Tgt The method comprises the steps of carrying out a first treatment on the surface of the The pedestrian information includes an abscissa X of the pedestrian in the global coordinate system _Ped Ordinate Y _Ped Velocity vectorThe signal lamp states are TrafficLight, and include three states, namely red (R), green (G) and yellow (Y) lights.

Step three: the decision module provides vehicle decision information. Self-supportingBefore the vehicle enters the intersection, the vehicle makes a Decision on the global path in advance, and the Decision (Decision _Ego ) The three types are left turn (tl), straight turn (gs) and right turn (tr), and the decision module sends the decision information to the CAN bus in a defined bus protocol after making decisions.

The intersection inner area is an intersection area formed by extending four stop lines. Coordinate information of four stop lines is input (R ₁ .s，R ₂ .s，R ₃ .s，R ₄ ·s)，R _i The straight line where s is located is y (R _i S), stop line R ₁ S and R ₂ The intersection point obtained by s extension is P _{IntersectionArea1} ＝y(R ₁ ·s)∩y(R ₂ S); likewise, stop line R ₂ S and R ₃ The intersection point obtained by s extension is P _{IntersectionArea2} ＝y(R ₂ ·s)∩y(R ₃ S); stop line R ₃ S and R ₄ The intersection point obtained by s extension is P _{IntersectionArea3} ＝y(R ₃ ·s)∩y(R ₄ S); stop line R ₄ S and R ₁ The intersection point obtained by s extension is P _{IntersectionArea4} ＝y(R ₄ ·s)∩y(R ₁ S). The region within the intersection can thus be represented by its four vertices:

IntersectionArea＝[P _{IntersectionArea1} ，P _{IntersectionArea2} ，P _{IntersectionArea3} ，P _{IntersectionArea4} ]

traffic light traffic rule triggering conditions: when the vehicle passes through the stop line from the geometric center of the vehicle to enter the intersection, the triggering condition is that the vehicle front end passes through the stop line until the vehicle rear end passes through the stop line. Input vehicle ordinate X in global coordinate system _Ego Y of abscissa _Ego Heading angle theta of bicycle _Ego

And the map information is used for calculating whether the triggering condition of traffic light passing rules is met or not, and the calculation formula is as follows:

representing the midpoint coordinates of the front end of the bicycle;

representing the coordinates of the midpoint of the rear end of the vehicle.

Travel route rule triggering conditions: when the vehicle runs in the intersection, the vehicle should run according to the optimal arc curve, and the triggering condition is that the vehicle runs in the intersection area. Input vehicle ordinate X in global coordinate system _Ego Y of abscissa _Ego Whether the triggering condition of the driving route rule is met or not is calculated, and the calculation formula is as follows:

wherein Crdn _Ego ＝(X _Ego ，Y _Ego ) Representing the geometric center coordinates of the own vehicle.

Road right rule triggering condition: vehicles cannot prevent the vehicles with high road rights from normally running when running in the intersection. The triggering conditions of the road right rule are the same as the triggering conditions of the driving route, and the triggering conditions are all the areas in the vehicle position at the intersection. Input vehicle ordinate X in global coordinate system _Ego Y of abscissa _Ego Whether the trigger condition of the road weight rule is met or not is calculated, and the calculation formula is as follows:

the pedestrian rule triggering condition is not hindered: when the self-vehicle enters the crosswalk, the normal traffic of pedestrians is not hinderedAnd (3) row. The triggering condition is that the front end of the vehicle enters the crosswalk area, and the center of the vehicle does not enter the crosswalk area. Input vehicle ordinate X in global coordinate system _Ego Y of abscissa _Ego Heading angle theta of bicycle _Ego And map information, calculating whether the triggering condition of the rule of the pedestrian is not prevented from being met, wherein the calculation formula is as follows:

CWArea _i ＝[P _izinl ，P _izinr ，P _izoutl ，P _izoutr ]，i∈{1，2，3，4}

wherein CWArea is provided _i Is a branch crossing R _i Corresponding crosswalk regions, represented by four vertices of the crosswalk region, are P _izinl ＝(R _i .zin.X _i ，R _i .zin.Y _l )，P _izinr ＝(R _i .zin·X _r ，R _i .zin·Y _r )，P _izoutl ＝(R _i .zout.X _l ，R _i .zout.Y _l )，P _izoutr ＝(R _i .zout.X _r ，R _i .zout.Y _r )。

Step five: and (5) monitoring triggering condition judgment. And the monitoring triggering module calculates the triggering condition calculation formula of each rule according to the information read in the step four and outputs an enabling signal result. If the calculation result of the monitoring triggering condition of a certain rule is 1, the rule needs to be monitored at the moment, and a corresponding rule logic judging module starts to work and enters a step six; if the calculation result of the monitoring triggering condition is 0, the condition that the rule is not required to be monitored at the moment is indicated, and the step five is repeated.

Step six: the legislation logic determination module receives the data. When the monitoring triggering condition output enabling signal of a certain rule is 1, the rule logic judgment module of the violation monitoring module starts to read information needed by logic judgment from the CAN bus.

Traffic light traffic rule logic judgment: the vehicle samples time before and after passing the stop line,when the corresponding signal lamps are not green lamps, indicating that the vehicle passes illegally; second, the behavior of the vehicle should be determined to be consistent with the road surface direction marking. Input vehicle ordinate X in global coordinate system _Ego Y of abscissa _Ego Heading angle theta of bicycle _Ego Lane of the Lane to which the own vehicle belongs _Ego And map information, traffic light status TrafficLight, and Decision information definition of own vehicle _Ego The logic for calculating traffic light passing rules judges whether the traffic light passing rules are met or not, and the calculation formula is as follows:

wherein t is the last sampling time; the IllgalTrafficlight is a calculation formula of 'the own vehicle does not pass according to the specified traffic light passing rule', the calculation result is 0 to indicate the compliance, the own vehicle complies with the traffic light rule, and the result is 1 to indicate the violation; illgal Lane Mark is a calculation formula of 'the own vehicle does not travel according to the allowed direction of the lane marking', the calculation result is 0 to indicate the compliance, the own vehicle complies with the lane marking limit, and the calculation result is 1 to indicate the violation, wherein ENLane _Ego For the lane to which the host vehicle belongs after entering the ODD, i.e.t ₀ The initial time of the program running after the host enters the ODD.

And (3) driving route rule logic judgment: when the own vehicle runs in the intersection, the own vehicle should run according to the optimal virtual lane, as shown in fig. 3, firstly, according to the decision of the initial lane and the own vehicle when the own vehicle enters the intersection, the optimal virtual lane and the non-recommended virtual lane are generated, and the generation mode is as follows:

inputting Lane Lane to which own vehicle belongs _Ego Decision information definition of map information and own vehicle _Ego Calculation ofThe optimal virtual lane and the non-recommended virtual lane are calculated as follows:

is provided withThe best virtual lane is:

the virtual lane is not recommended to be the adjacent exit lane of the best virtual lane, so there are at most two:

NREXLane _Ego ＝{R _i’. L _j’-1 ，R _i’. L _j’+1 }

after obtaining the optimal virtual lane and the non-recommended virtual lane, inputting the lane ENLane to which the own vehicle belongs after entering the ODD _Ego Best virtual lane BEXLane _Ego Non-recommended virtual lane NREXLane _Ego And map information, calculating the intersection points of the lanes and the outer boundaries of the corresponding crosswalk in the following calculation modes:

P _enl∩z ＝y(P _enlf ，P _enlr )∩y(P _izoutl ，P _izoutr )

P _enr∩z ＝y(P _enrf ，P _enrr )∩y(P _izoutl ，P _izoutr )

P _bexl∩z ＝y(P _bexlf ，P _bexlr )∩y(P _i’zoutl ，P _i’zoutr )

P _bexr∩z ＝y(P _bexrf ，P _bexrr )∩y(P _i’zoutl ，P _i’zoutr )

P _nrexl∩z ＝y(P _nrexlf ，P _nrexlr )∩y(P _i’zoutl ，P _i’zoutr )

P _nrexr∩z ＝y(P _nrexrf ，P _nrexrr )∩y(P _i’zoutl ，P _i’zoutr )

wherein y (p 1, p 2) is the straight line where the point p1 and the point p2 are located; p (P) _enlf ＝(ENLane _Ego .X _lf ，ENLane _Ego .Y _lf )、P _enlr ＝(ENLane _Ego ·X _lr ，ENLane _Ego .Y _lr )、P _enrf ＝(ENLane _Ego .X _rf ，ENLane _Ego ·Y _rf )、P _enrr ＝(ENLane _Ego .X _rr ，ENLane _Ego .Y _rr ) The left front vertex, the left rear vertex, the right front vertex and the right rear vertex of the entrance lane respectively; p (P) _bexlf ＝(BEXLane _Ego ·X _lf ，BEXLane _Ego .Y _lf )、P _bexlr ＝(BEXLane _Ego .X _lr ，BEXLane _Ego .Y _lr )、P _bexrf ＝(BEXLane _Ego .X _rf ，BEXLane _Ego .Y _rf )、P _bexrr ＝(BEXLane _Ego .X _rr ，BEXLane _Ego .Y _rr ) The left front vertex, the left rear vertex, the right front vertex and the right rear vertex of the optimal virtual lane are respectively; p (P) _nrexlf ＝(NREXLane _Ego .X _lf ，NREXLane _Ego .Y _lf )、P _nrexlr ＝(NREXLane _Ego .X _lr ，NREXLane _Ego .Y _lr )、P _nrexrf ＝(NREXLane _Eoo .X _rf ，NREXLane _Ego .Y _rf )、P _nrexrr ＝(NREXLane _Ego .X _rr ，NREXLane _Ego .Y _rr ) Respectively recommending a left front vertex, a left rear vertex, a right front vertex and a right rear vertex of the virtual lane; p (P) _enl∩z The intersection point of the straight line of the left boundary of the entrance lane and the outer boundary of the corresponding crosswalk; p (P) _enr∩z The intersection point of the straight line of the right boundary of the entrance lane and the outer boundary of the corresponding crosswalk; p (P) _bexl∩z The intersection point of the straight line of the left boundary of the optimal virtual lane and the outer boundary of the corresponding crosswalk; p (P) _bexr∩z The intersection point of the straight line of the right boundary of the optimal virtual lane and the outer boundary of the corresponding crosswalk; p (P) _nrexl∩z Intersection points of a straight line where the left boundary of the non-recommended virtual lane is located and the outer boundary of the corresponding crosswalk; p (P) _nrexr∩z And the intersection point of the straight line where the right boundary of the non-recommended virtual lane is located and the outer boundary of the corresponding crosswalk.

After obtaining intersection points of an entrance lane, an exit lane and corresponding crosswalk, inputting coordinates and decision information of the intersection points to generate an optimal virtual lane boundary and an un-recommended virtual lane boundary, and taking P if the vehicle decides to turn left or right _enl∩z ＝(x ₁ ，y ₁ )，P _enr∩z ＝(x ₂ ，y ₂ ) Sequentially making the intersection point of the calculated exit lane and the corresponding crosswalk outer boundary be (x) ₃ ，y ₃ ) The circle centers of the arc boundaries of the virtual lane can be respectively obtained by bringing the following steps:

After the circle centers of the arc boundaries of the virtual lanes are obtained, the circle center is set as c= [ x ] _c ，y _c ]The radius of each virtual lane arc boundary isThe starting point and the end point of the arc boundary are respectively the intersection point of the straight line of the entrance lane boundary and the corresponding crosswalk outer boundary, and the intersection point of the straight line of the exit lane boundary and the corresponding crosswalk outer boundary.

The arc curve boundary is expressed as [ starting point, (center, radius), end point ], the optimal virtual lane region can be expressed as:

BVLane＝[[P _enl∩z ，(c _bl ，r _bl )，P _bexl∩z ]，[P _enr∩z ，(c _br ，r _br )，P _bexr∩z ]]

the non-recommended virtual lanes may be expressed as:

NRVLane＝[[P _enl∩z ，(c _nrl ，r _nrl )，P _nrexl∩z ]，[P _enr∩z ，(c _nrr ，r _nrr )，P _nrexr∩z ]]

wherein c _bl ，c _br ，c _nrl ，c _nrr Respectively represent the center of the left boundary of the optimal virtual lane and the right boundary of the optimal virtual laneThe boundary circle center does not recommend the left boundary circle center of the virtual lane and does not recommend the right boundary circle center of the virtual lane; r is (r) _bl ，r _br ，r _nrl ，r _nrr Respectively representing the left boundary radius of the optimal virtual lane, the right boundary radius of the optimal virtual lane, the left boundary radius of the non-recommended virtual lane and the right boundary radius of the non-recommended virtual lane;

if the vehicle makes a decision to go straight, the virtual lane is a quadrilateral region, which can be directly expressed as:

BVLane＝[P _enl∩z ，P _bexl∩z ，P _bexr∩z ，P _enr∩z ]

NRVLane＝[P _enl∩z ，P _nrexl∩z ，P _nrexr∩z ，P _enr∩z ]

after the optimal virtual lane and the non-recommended virtual lane are calculated, inputting the ordinate X of the own vehicle in the global coordinate system _Ego Y of abscissa _Ego And calculating the area of the own vehicle to judge whether the running route of the own vehicle is compliant or not according to the optimal virtual lane BVLLane and the non-recommended virtual lane NRVLane, wherein the calculation formula is as follows:

The calculation result of the calculation formula of FollowBestVirtualLane is 1, which indicates that the calculation formula is compliant with the optimal virtual lane, and the calculation result of the calculation formula is 0, which indicates that the self-vehicle does not obey the virtual lane; the FollownNRVirtualLane is a calculation formula of 'the self-vehicle driving on the non-recommended virtual lane', the calculation result is 1, the self-vehicle driving on the non-recommended virtual lane is indicated, the VirtualLane is a monitoring result of driving route rules, the result is 1, the self-vehicle driving route is in compliance, and the result is 0, the self-vehicle driving route violations are indicated.

Road right rule logic judgment: other vehicles with high road rights may exist when the own vehicle turns in the intersection, and the own vehicle should not prevent the normal running of the vehicles with high road rights. The driver should pay attention to the straight-going vehicles when turning left, pay attention to the left straight-going vehicles when turning right when turning left. The specific monitoring and calculating flow is as follows:

first, inputting the calculated optimal virtual lane and the ordinate X of the own vehicle in the global coordinate system _Ego Y of abscissa _Ego Vehicle entrance lane ENLane _Ego Decision information definition of map information and own vehicle _Ego Generating a virtual stop line Vtopline _k And a highway right monitoring area CheckArea _k . Still set to ENLane _Ego ＝R _i L _j The left branch crossing of the entrance lane of the own vehicle isThe opposite branch crossing is +.>The right branch crossing is +.>N _len ＝R _l .N _en The number of entrance lanes for the left branch road junction; n (N) _oen ＝R _o ·N _en The number of entrance lanes for the opposite branch road junction; n (N) _lex ＝R _l .N _ex The number of exit lanes for the left-hand branch road junction.

If precision _Ego =tl, opposite entry laneLeft and right boundaries of the optimal virtual straight lane and left exit lane +.>The intersection of the straight lines at the right boundary of (2) is set to +.>And->The other opposite entrance lanes R _o .L _k Left and right boundaries of the optimal virtual straight lane and left exit lane +.>The intersection point of the straight lines of the left boundary is set as P _{calr_k} And P _{carr_k} Each of the opposite entrance lanes R _o .L _k The intersection point of the right boundary of the optimal virtual straight lane and the left boundary of the optimal virtual lane of the own vehicle is set as P _{carf_k} P is passed through _{carf_k} To R _o .L _k The intersection point of the right boundary perpendicular line and the right boundary perpendicular line is set as P _{calf_k} Then the road right monitoring area CheckArea _k Can be expressed as:

CheckArea _k ＝[P _{carf_k} ，P _{calr_k} ，P _{carr_k} ，P _{calf_k} ]，k＝1，2...N _oen

the virtual stop line may be expressed as:

Vstopline _k ＝y(c _bl ，P _{calf_k} )，k＝1，2...N _oen

if precision _Ego Tr =Λ trafficlight=r, left entrance laneLeft and right boundaries of the optimal virtual straight lane of (a) and the entrance lane R _i .L _j-1 The intersection of the straight lines at the left boundary of (2) is set to +.>And->(in particular, if j = 1, R is taken _i .L _-1 Straight line where the right boundary of the left lane R is located), the remaining left entrance lanes R _l .L _k Left and right boundaries of the optimal virtual straight lane of (a) and the entrance lane R _i .L _j The intersection point of the straight lines of the left boundary is set as P _{calr_k} And P _{carr_k} Each left entrance lane R _l .L _k The intersection point of the right boundary of the optimal virtual straight lane and the right boundary of the optimal virtual lane of the own vehicle is set as P _{carf_k} P is passed through _{carf_k} To R _l .L _k The intersection point of the left boundary perpendicular line and the left boundary line is set as P _{calf_k} Then the road right monitoring area CheckArea _k Can be expressed as:

CheckArea _k ＝[P _{carf_k} ，P _{calr_k} ，P _{carr_k} ，P _{calf_k} ]，k＝j，j+1…N _len

the virtual stop line may be expressed as:

Vstopline _k ＝y((R _l .L _k .X _rf ，R _l .L _k .Y _rf )，(R _r .L _-k .X _rr ，R _r .L _-k .Y _rr ))，k＝j，j+1...N _len

if precision _Ego Tr =Λ trafficlight=g, left entrance lane R _l .L _k Left and right boundaries of the optimal virtual straight lane of (a) and the entrance lane R _i .L _j The intersection point of the straight lines of the left boundary is set as P _{calr_k} And P _{carr_k} Each left entrance lane R _l .L _k The intersection point of the right boundary of the optimal virtual straight lane and the right boundary of the optimal virtual lane of the own vehicle is set as P _{carf_k} P is passed through _{carf_k} To R _l .L _k The intersection point of the left boundary perpendicular line and the left boundary line is set as P _{calf_k} Then the road right monitoring area CheckArea _k Can be expressed as:

CheckArea _k ＝[P _{carf_k} ，P _{calr_k} ，P _{carr_k} ，P _{calf_k} ]，k＝j，j+1...N _len

the virtual stop line may be expressed as:

calculating virtual stop line Vtopline _k Road right monitoring area CheckArea _k Then, the ordinate X of the own vehicle in the global coordinate system is input _Ego Y of abscissa _Ego Longitudinal speed vx of own vehicle _Ego Decision information definition of map information and own vehicle _Ego Input the abscissa X of other vehicles in the global coordinate system at the same time _Tgt Ordinate Y _Tgt Length L of other vehicle _Tgt Width W _Tgt Course angle information θ _Tgt When the self-vehicle passes through each virtual stop line, whether a highway right vehicle exists in the corresponding highway right monitoring area or not is calculated, and the calculation formula is as follows:

wherein, viola Rightofway is a calculation formula of 'own vehicle violating road right rule', the calculation result is 0 to indicate that the rule is compliant, the own vehicle does not violate the road right rule, and the result is l to indicate that the rule is violating; overlay (ele) ₁ ，ele ₂ ) For atomic proposition, which indicates whether there is an overlap region between two elements, if ele ₁ And ele ₂ With overlap, the output is 1, otherwise the output is 0, ele ₁ And ele ₂ The atomic proposition calculation formula can be a line segment or a plane area as follows:

segment(p ₁ ，p ₂ ) Representing the point p ₂ And p ₂ A segment formed by connection;

representing the midpoint coordinates of the front end of the vehicle; incln (veh) is an atomic proposition, representing the current heading angle of the vehicle veh and the deflection angle of the optimal straight virtual lane direction, and setting the veh entrance lane as R _i .L _j Which is R along the exit lane of the optimal straight virtual lane _i’ .L _j’ The calculation formula is as follows:

the angle range_gs is an angle range for determining the straight running of the vehicle, and is taken as = [0, 15 ]).

The logic judgment of pedestrian rules is not hindered: dividing the crosswalk area into corresponding subareas CWSubA according to each lane of the branch road junction _j When a vehicle enters a certain sub-area, no pedestrians should exist in the corresponding sub-area, and meanwhile, no pedestrians which travel towards the sub-area where the vehicle is located should exist in the adjacent sub-area. The sub-region id of each crosswalk corresponds to the lane id of the branch road, i.e. R _i .CWSubA _j Is R and _i .L _j corresponding crosswalk subregion, R _i .CWSubA _j The calculation formula for each vertex of (a) is as follows:

P _{CWsAlf_ij} ＝y(P _izinl ，P _izinr )∩y((R _i .L _j .X _lf ，R _i .L _j .Y _lf )，(R _i .L _j .X _lr ，R _i .L _j .Y _lr ))

P _{CWSAlr_ij} ＝y(P _izoutl ，P _izoutr )∩y((R _i .L _j .X _lf ，R _i .L _j .Y _lf )，(R _i .L _j .X _lr ，R _i .L _j .Y _lr ))

P _{CWSAlr_ij} ＝y(P _izinl ，P _izinr )∩y((R _i .L _j .X _rf ，R _i .L _j .Y _rf )，(R _i .L _j .X _rr ，R _i .L _j .Y _rr ))

P _{CWSArr_ij} ＝y(P _izoutl ，P _izoutr )∩y((R _i .L _j .X _rf ，R _i .L _j .Y _rf )，(R _i .L _j .x _rr ，R _i .L _j .Y _rr ))

wherein P is _{CWSAlf_ij} 、P _{CWSAlr_ij} 、P _{CWSAlr_ij} 、P _{CWSArr_ij} Respectively crosswalk subareas R _i .CWSubA _j Left front, left rear, right front and right rear vertices of (c), crosswalk subregions may be represented as:

R _i .CWSubA _j ＝[P _{CWSAlf_ij} ，P _{CWSAlr_ij} ，P _{CWSAlr_ij} ，P _{CWSArr_ij} ]

R _i .CWSubA _j adjacent sub-region R of (2) _i .CWSubAN _j Is R _i .CWSubA _j-1 Or R is _i .CWSubA _j+1 In particular, when j=1, R _i .CWSubA _j-1 Is R _i .CWSubA _-1 When j= -1, R _i .CWSubA _j+1 Is R _i .CWSubA ₁ The method comprises the steps of carrying out a first treatment on the surface of the The crosswalk subregion where the own vehicle is located can be expressed as:

the adjacent subarea of the crosswalk subarea where the own vehicle is located is denoted as CWSubAN _Ego 。

Input vehicle ordinate X in global coordinate system _Ego Y of abscissa _Ego Heading angle theta of bicycle _Ego Map information and abscissa X of pedestrian in global coordinate system _Ped Ordinate Y _Ped Velocity vectorCalculating whether pedestrians exist in a corresponding subarea when a vehicle enters a subarea, and simultaneously, adjacent subareas Whether there is a pedestrian in the domain traveling toward the sub-area where the own vehicle is located. The calculation formula is as follows:

wherein Crdn _ped ＝(X _Ped ，Y _Ped ) Representing coordinates of pedestrians in a global coordinate system;for atomic proposition, a vector with origin p is denoted +.>Is set to be area= [ p ] ₁ ，p ₂ ，p ₃ ，p ₄ ]The calculation formula is as follows:

the ImpedePederstran is a calculation formula of 'pedestrian obstacle of own vehicle', the calculation result of 0 indicates that the rule is compliant, the rule of pedestrian obstacle is not violated by own vehicle, and the calculation result of 1 indicates that the rule is violated.

Step seven: the rule logic determination module performs rule violation monitoring. And D, calculating a logic judgment calculation formula of each rule according to the information read in the step six, and outputting a calculation result to the CAN bus in real time.

According to the monitoring system provided by the embodiment of the invention, logic judgment is carried out by combining the intersection scene information at the current moment acquired by the sensing system and the high-precision map information, and the violation condition of the vehicle in the whole process of passing through the intersection is monitored in real time, so that the real-time violation monitoring of the vehicle end of the automatic driving vehicle at the intersection is realized. Through the auxiliary monitoring of the high-precision map, the input requirement of a sensing system can be reduced, the accuracy of monitoring judgment is improved, the rule is divided into a trigger domain and a logic judgment part, the calculation force requirement of the system can be reduced, and the rule division modularization has the characteristic of good maintainability. Therefore, the system has the advantages of real-time monitoring, high accuracy, low sensing requirement, low calculation force requirement, good maintainability, strong floor property and the like. Has the following advantages:

1. the technical scheme of the invention provides a real-time violation monitoring system for the vehicle end of the automatic driving vehicle in the crossroad scene, and the automatic driving vehicle can correctly understand and obey traffic regulations by digitizing the regulations of the crossroad, and meanwhile, the system can correctly monitor the behavior compliance of the automatic driving vehicle.

2. According to the method and the system, the automatic driving automobile can be guided to make a compliance decision according to the compliance monitoring result, so that the automatic driving automobile can improve legal compliance, more accord with human driving in behavior performance, improve the safety of human on the automatic driving automobile, and reduce the occurrence of dangerous behaviors caused by excessive reaction due to illegal operation of the automatic driving automobile.

3. If a traffic accident happens accidentally in the driving process, the real-time compliance monitoring result of the invention is also convenient for the responsibility tracing work after the accident.

Claims

1. A real-time violation monitoring method for an automatic driving automobile, which is characterized by being applied to an intersection, and comprising the following steps:

If the static component element information of the crossroad, the self-vehicle state information and the environment information meet the preset triggering conditions of any regulations, acquiring data corresponding to the regulations and judging whether the regulations are illegal or not;

if the static component element information of the crossroad, the self-vehicle state information and the environment information meet the preset triggering conditions of any regulations, the data corresponding to the regulations are obtained to judge whether the regulations are violated, and the method comprises the following steps:

calculating the intersection points of the lanes to which the self-vehicle belongs, the optimal virtual lanes and the outer boundaries of the non-recommended virtual lanes and the crosswalk;

If yes, judging that the self-vehicle driving route is not illegal; if not, judging that the self-vehicle driving route is illegal;

if the static component element information of the crossroad, the vehicle state information and the environment information meet the preset triggering conditions of any regulations, the data corresponding to the regulations are obtained to judge whether the regulations are violated, and the method comprises the following steps: when judging that the geometric center of the vehicle is positioned in an area surrounded by straight lines where four parking lines are positioned according to the static component element information and the vehicle state information of the crossroad, meeting the road right rule triggering condition; generating a virtual stop line and a highway right monitoring area according to the optimal virtual lane, the vehicle state information and the static component element information of the crossroad and the signal lamp state; when the own vehicle passes through the virtual stop line, judging whether the other vehicle exists in the highway right monitoring area; if yes, judging that the own vehicle violates the road right rule; if the road right rule is not violated, judging that the own vehicle does not violate the road right rule;

the generating a virtual stop line and a highway right monitoring area according to the optimal virtual lane and self-vehicle state information and the static component element information of the crossroad comprises the following steps: determining a lane of the highway right according to the signal lamp state and the self-vehicle decision, and obtaining a highway right monitoring area according to the lane of the highway right; and generating a virtual stop line according to the lane to which the own vehicle belongs and the optimal virtual lane.

2. The method according to claim 1, characterized in that: if the static component element information of the crossroad, the self-vehicle state information and the environment information meet the preset triggering conditions of any regulations, the data corresponding to the regulations are obtained to judge whether the regulations are violated, and the method comprises the following steps:

3. The method as recited in claim 1, further comprising:

and judging whether the self-vehicle driving route is within the optimal virtual lane boundary according to the self-vehicle coordinates, and if not, recording data for self-vehicle learning for improving the self-vehicle driving route selection capability.

4. The method according to claim 1, characterized in that: if the static component element information of the crossroad, the self-vehicle state information and the environment information meet the preset triggering conditions of any regulations, the data corresponding to the regulations are obtained to judge whether the regulations are violated, and the method comprises the following steps:

if so, judging that the own vehicle hinders the pedestrians.

5. The method according to claim 4, wherein: if no pedestrians exist in the pedestrian crossing subarea, judging whether pedestrians which move towards the subarea where the self-vehicle is located exist in adjacent subareas of the pedestrian crossing subarea;

6. A real-time violation monitoring device at a vehicle end of an automatic driving vehicle, characterized in that the device is applied to an intersection, and comprises:

The violation judging unit is used for acquiring data corresponding to any regulations and judging whether the rule is violated if the static component element information of the crossroad, the vehicle state information and the environment information meet the preset triggering conditions of any regulations;

the violation judging unit is specifically used for meeting the rule triggering condition of the driving route when judging that the geometric center of the vehicle is positioned in the area surrounded by the straight lines where the four parking lines are positioned according to the static component element information of the crossroad and the vehicle state information; generating an optimal virtual lane and an un-recommended virtual lane according to a lane to which a vehicle belongs and a vehicle decision, wherein the un-recommended virtual lane is an adjacent exit lane of the optimal virtual lane; calculating the intersection points of the lanes to which the self-vehicle belongs, the optimal virtual lanes and the outer boundaries of the non-recommended virtual lanes and the crosswalk; generating an optimal virtual lane boundary and an un-recommended virtual lane boundary according to the intersection point and the self-vehicle decision; judging whether the vehicle coordinate is in the optimal virtual lane boundary and the non-recommended virtual lane boundary or not according to the vehicle coordinate; if yes, judging that the self-vehicle driving route is not illegal; if not, judging that the self-vehicle driving route is illegal;

7. An automatic driving automobile end real-time violation monitoring system, which is characterized by comprising:

a processor;

a memory for storing the processor-executable instructions;

the processor is configured to perform the method of any of claims 1-5.