Automatic driving-oriented structured road traffic environment cognition method

Abstract

The invention is suitable for the technical field of intelligent automobiles, and provides an automatic driving-oriented structured road traffic environment cognition method, which comprises the following steps of: uniformly representing a front driving road section in a road coordinate system; respectively cognizing traffic attributes contained in traffic marking lines, traffic signs and road structures, and indexing the traffic attributes to geometric regions; fusing multi-source attributes, and determining a final type, a speed limit, a boundary line type and a driving-out direction of each region; a geometric region is subdivided into traffic regions on the basis of a rule consistency principle, a structured environment model is constructed, then the connection relation of all the regions is recognized, and influences of traffic lights, traffic participants and obstacles are dynamically fused. According to the method, deep fusion of traffic rule semantics and road geometric information is realized, an environment model which takes a traffic area as a basic unit, is clear in rule and can be dynamically updated is created, and the accuracy and reliability of understanding and decision making of an automatic driving system on a complex traffic environment are remarkably improved.

Description

Technical Field

[0001] This invention belongs to the field of intelligent vehicle technology, and in particular relates to a traffic environment cognition method for structured roads for autonomous driving. Background Technology

[0002] With the development of autonomous vehicles, current driver assistance technologies have been widely used in mass-produced models, while Level 4 autonomous vehicles are still in the demonstration phase.

[0003] Current L4 autonomous driving technology, whether the classic "perception-decision-control" approach or AI-based autonomous driving systems, cannot guarantee that decision-making outcomes will always be legal. The root cause of illegal behavior is that current autonomous driving systems primarily handle traffic regulations by treating them as constraints during motion planning, optimizing them along with other constraints. This turns traffic regulations into "soft constraints," making it impossible to guarantee that every decision outcome will be completely legal. Summary of the Invention

[0004] The purpose of this invention is to provide a traffic environment cognition method for structured roads for autonomous driving, aiming to solve the problems mentioned in the background art.

[0005] The present invention is implemented as follows: a traffic environment cognition method for structured roads for autonomous driving includes the following steps: Step 1: Road segment recognition. Determine the recognition distance based on the current visibility and sensor range, and based on the travel task, represent the road sequence (including intersections, which are recognized as one-way roads with entrances and exits) within the recognition distance in a unified road coordinate system (Frenet coordinate system).

[0006] Step 2: Traffic Marking Recognition. Analyze traffic markings (such as directional, prohibitory, and warning markings) and recognize them as constraints on the position, speed, and direction of travel of motor vehicles (i.e., traffic attributes), and index these attributes into the geometric area in which they apply.

[0007] Step 3: Traffic Sign Recognition. Analyze traffic signs (such as prohibitory, instruction, and warning signs), recognize them as constraints on the position, speed, size, mass, and direction of travel of motor vehicles, and index them into the corresponding geometric areas.

[0008] Step 4: Road Structure Understanding. Understand the road network connections, the number of lanes and their adjacency relationships, and determine the road type (e.g., expressway, highway, urban road).

[0009] Step 5: Traffic Attribute Fusion. The cognitive results from Steps 2-4 are fused to determine the final traffic attributes of each geometric region, including: a) Traffic area type (lane, passage zone, restricted zone); b) Upper and lower speed limits; c) Left and right boundary line type (solid line, dashed line); d) Exit direction.

[0010] Step 6: Traffic Area Recognition. Based on the principle of completely consistent traffic attributes, each geometric region is further divided into multiple "traffic areas". These areas are then pieced together to cover the entire road segment ahead, constructing a traffic environment model based on traffic areas.

[0011] Step 7: Understanding Connections. Determine the connection relationships between the entrances and exits of each traffic area, as well as their left and right adjacent relationships.

[0012] Step 8: Traffic Light Recognition. Identify traffic light colors and change the type of the relevant traffic area according to the color (e.g., when the light is red / yellow, the area type in front of the stop line changes to a passable area).

[0013] Step 9: Traffic Participant Recognition. Identify the types and states (location, speed, etc.) of traffic participants such as pedestrians and vehicles, and influence the type of traffic area based on their type and location (e.g., when there are pedestrians at a crosswalk, the area becomes a passage zone).

[0014] Step 10: Obstacle Recognition. Identify static or dynamic obstacles and index them to their respective traffic areas.

[0015] This invention provides a method for recognizing the traffic environment of structured roads for autonomous driving. This method decouples traffic regulations from other constraints, systematically and comprehensively analyzes traffic regulations before behavioral decisions or motion planning, and recognizes the structured road as a traffic environment represented by traffic areas. Based on the traffic attribute information of each traffic area, a set of legal traffic area candidates is selected to ensure the legality of subsequent decision results. This method can achieve diverse motion planning by including passage areas in the traffic area candidate set. During decision-making, legality can be temporarily sacrificed for benefits in other value categories, thus achieving an aggressive driver model. If the traffic area candidate set only contains lanes, then a normal driver model or a defensive driver model is implemented. Attached Figure Description

[0016] Figure 1 A diagram illustrating traffic zone types; Figure 2 A technical roadmap for a traffic environment cognition method for structured roads in autonomous driving, provided by an embodiment of the present invention; Figure 3 Intersection recognition map; Figure 4 Example diagram of traffic environment cognition results. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0018] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0019] First, it should be noted that a traffic zone is a two-dimensional enclosed area on a road marked by traffic signs, possessing specific traffic meaning. The traffic attribute information within each traffic zone is completely consistent. Traffic zones are divided into, for example... Figure 1 The three categories shown: (1) A lane that allows motor vehicles to travel and has an entrance and exit, such as a motor vehicle lane; (2) Traffic areas where the road surface is suitable for motor vehicles but traffic regulations do not allow motor vehicles to drive, such as non-motorized vehicles, diversion areas, and central circles; (3) Restricted areas where the road surface is not suitable for motor vehicle driving, such as the median strip in the middle of the highway and the sidewalks on both sides of the road.

[0020] like Figure 2 As shown, an embodiment of the present invention provides a traffic environment cognition method for structured roads for autonomous driving, comprising the following steps: Step 1: Understanding the driving route; To adapt to adverse weather conditions, the perceived distance is determined based on current visibility and the effective sensing range of the vehicle's sensors. Based on the current travel task, multiple interconnected one-way roads within the perceived distance are selected from the current location; that is, the road ahead consists of a sequence of multiple one-way roads.

[0021] Since motor vehicles only enter an intersection from one road and exit to enter another road when passing through an intersection, when recognizing an intersection, based on the current travel task, an intersection can be perceived as a one-way road with an entrance and an exit, such as... Figure 3As shown, if the vehicle's journey is from road 1 to road 4, then intersection 2 is perceived as a straight road; if the vehicle's journey is from road 1 to road 5, then intersection 2 is perceived as a curve. That is, under the premise of this journey task, all structured roads will be perceived as a one-way road with a single entrance and a single exit. Therefore, all structured roads can be represented using a road coordinate system. This road coordinate system belongs to the Frenet coordinate system, and the shape of the road, its geometric regions, traffic areas, traffic participants, obstacles, etc., are all represented based on this road coordinate system.

[0022] The road's attribute data includes: road number, length, coordinate system, shape, and a list of traffic zones, as shown in Table 1.

[0023] Table 1 Road Entity Attribute Data

[0024] Road number: In a traffic environment, each road has a unique number.

[0025] Road length: The length from the beginning to the end of the road along the direction of motor vehicle travel.

[0026] Road coordinate system: Since roads are unidirectional and have entrances and exits, a coordinate system is established along the direction of vehicle travel, using the leftmost boundary of the road as a reference line. The road coordinate system is expressed as a series of polylines, which are represented by a sequence of points in a rectangular coordinate system, as shown in the following expression: ; In the formula, For the road coordinate system, This represents a point on the reference line in the road coordinate system. Its specific expression is as follows: ; In the formula, For the first Mileage coordinates of a reference point, For the first The x-coordinate of each reference point in a rectangular coordinate system For the first The ordinate of each reference point in a rectangular coordinate system For the first The heading angle at each reference point.

[0027] Road shape: The road boundary and the road curvature at each reference point are both represented in the road coordinate system, and the specific expressions are as follows: ; In the formula, This represents the left road boundary in the road coordinate system. This represents the right-side road boundary in the road coordinate system. Let be the road curvature at each reference point in the road coordinate system. The first one on the left road boundary The horizontal coordinates at each reference point The first one on the right side of the road boundary The horizontal coordinates at each reference point For the first Road curvature at each reference point.

[0028] Traffic Zone List: Stores the traffic zones belonging to this road. After step 6, the cognitive results will be indexed in this list.

[0029] Step 2: Traffic marking recognition; This section analyzes traffic markings in traffic regulations. Based on the definitions of directional, prohibitory, and warning markings in GB 5768.3, traffic markings are recognized as requirements for the position, speed, and direction of departure of motor vehicles, i.e., traffic attributes. These traffic attributes are then indexed into the geometric area where the traffic marking applies. The geometric area is represented based on its corresponding road coordinate system, using the following formula: ; In the formula, The starting mileage coordinates of the geometric region. The coordinates of the end point of the geometric region.

[0030] Traffic markings describing lane boundaries, such as lane dividers, carriageway edge lines, and tidal flow lane lines, are perceived as traffic attributes that constrain the position of motor vehicles, including: the boundary of the geometric area and whether the vehicle is permitted to enter that geometric area. The perception results of common traffic markings describing lane boundaries are shown in Table 2 below.

[0031] Table 2. Cognitive Results of Traffic Markings Describing Lane Boundaries

[0032] Traffic markings that describe non-lane boundaries, such as pedestrian crossings, stop lines, and directional arrows, are perceived as traffic attributes that constrain the position, speed, and direction of departure of motor vehicles. Specifically: (1) A pedestrian crossing is a geometric area. Only when pedestrians, electric motorcycles and other non-motorized traffic participants are on the pedestrian crossing is the pedestrian crossing recognized as a geometric area in which motor vehicles cannot enter. Otherwise, the pedestrian crossing is recognized as a geometric area in which motor vehicles are allowed to enter. (2) The stop line is the exit edge of a geometric area. When the traffic light is red or yellow, the stop line is perceived as the exit of the geometric area with a speed of 0. (3) The directional arrows constrain the direction of the vehicle's departure and represent the connection relationship of the geometric regions, that is, after exiting the current geometric region, the vehicle can only enter the entrance of a few geometric regions.

[0033] The cognitive results of common traffic markings that do not describe lane boundaries are shown in Table 3 below.

[0034] Table 3. Cognitive Results of Traffic Markings Not Describing Lane Boundaries

[0035] Step 3: Traffic sign recognition; This analysis of traffic signs in traffic regulations is based on the national standard GB 5768.2, which defines prohibitory signs, instruction signs, and warning signs. Traffic signs are interpreted as requirements that constrain the position, speed, size, weight, and direction of travel of motor vehicles—that is, traffic attributes. These traffic attributes are then indexed into the geometric area within which the traffic sign applies. This geometric area is represented based on the road coordinate system to which it belongs.

[0036] Prohibitory signs restrict motor vehicles in seven main categories: (1) Prohibiting specific motor vehicles from entering a certain geometric area, such as signs prohibiting passage, prohibiting entry, and prohibiting motor vehicles from entering, the type of geometric area for which the sign is recognized is a passage area; (2) Restrict the speed of motor vehicles when they arrive at the sign, such as stop, yield, and pass signs, and recognize the speed requirements at the exit of the geometric area where the sign is applied; (3) Restricting the speed range of motor vehicles within a section of road, such as speed limit, speed limit lifting, minimum speed limit, and no parking signs, and recognizing the upper and lower speed limits of the geometric area where the sign is effective; (4) Limiting the upper limit of motor vehicle size, such as width limit signs, height limit signs, etc., which are recognized as the requirements for motor vehicle size in the geometric area where the sign is applied; (5) Limits the upper limit of the mass of motor vehicles, such as mass limit signs, axle load limit signs, etc., which are recognized as the mass requirements of motor vehicles in the geometric area where the sign is applied; (6) Restricting motor vehicles from entering certain specific geometric areas after leaving the current geometric area, such as signs prohibiting left turns, U-turns, and going straight, and recognizing the exit direction requirements of the geometric area where the sign is applied; (7) Other specific restrictions, such as the recognition of a no-overtaking sign as changing the left or right boundary line of the geometric area from a dashed line to a solid line, and the recognition of a no-overtaking sign as changing the left or right boundary line of the geometric area from a solid line to a dashed line, etc.

[0037] The recognition results of common prohibition signs are shown in Table 4 below.

[0038] Table 4. Cognitive Results of Prohibition Signs

[0039] The restrictions imposed on motor vehicles by directional signs can be mainly divided into four categories: (1) It is stipulated that after a motor vehicle leaves the current lane, it can only perform specific behaviors, such as going straight, turning left, driving in a roundabout, or following a U-turn sign. This is recognized as the direction of departure in traffic attributes, meaning that it can only enter a few fixed geometric areas. (2) Restrict the speed of motor vehicles when they arrive at signs such as priority for oncoming traffic, crosswalks, and pedestrian signs. The speed is recognized as the exit speed of the current geometric area and only takes effect when it is recognized that there are traffic participants in the specific geometric area ahead. (3) The type of motor vehicle traveling in the designated lane, such as motor vehicle lane, small passenger vehicle lane, bus lane, etc., is recognized as the type of the geometric area. If the motor vehicle meets the requirements of the sign, the type of traffic area is lane; if the motor vehicle does not meet the requirements of the sign, the type of traffic area is passage area. (4) The speed range for driving within the lane is specified, such as the minimum speed limit sign, and is recognized as the upper and lower limits of speed in that geometric area.

[0040] The cognitive results of common directional signs are shown in Table 5.

[0041] Table 5. Cognitive Results of Indicative Signs

[0042] Warning signs primarily constrain motor vehicles by reducing speed at the exit of a geometric area, such as signs for intersections, sharp turns, and consecutive turns. The recognition results of common warning signs are shown in Table 6.

[0043] Table 6. Cognitive Results of Warning Signs

[0044] Step 4: Understanding Road Structure; This method identifies the road network relationships (connections between roads), the number of lanes traveling in the same direction, and the left-right adjacency relationships of each lane. Based on the speed limits stipulated in the "Regulations for the Implementation of the Road Traffic Safety Law of the People's Republic of China" for different road types, this method categorizes road types into three types: expressways, highways, and urban roads. Specifically, the road types belong to the following set: ; in, Road type For highways, For highways, For city roads.

[0045] Step 5: Integration of traffic attributes; By integrating the different sources of traffic attribute information (such as traffic markings, traffic signs, traffic lights, etc.) identified in steps 2 to 4 above, the final value of each traffic attribute in each geometric area is determined.

[0046] Step 5.1: Determine the type of traffic area. Based on the findings of Steps 2 to 4, clarify the traffic area type for each geometric region. There are three types of traffic areas: (1) Define a traffic area where traffic regulations allow the motor vehicle to travel and where there are entrances and exits as a lane, such as a motor vehicle lane; (2) Define traffic areas that meet the conditions for driving this motor vehicle but are not permitted to drive this motor vehicle by regulations as traffic zones, such as non-motor vehicle lanes, diversion zones, and central rings; (3) Define traffic areas that do not meet the conditions for driving this motor vehicle as restricted areas, such as the median strip in the middle of the highway and the sidewalks on both sides of the road.

[0047] Therefore, the traffic area type belongs to the following set: ; in, As a traffic area type, For lanes, For passage, This is a restricted area.

[0048] The steps to determine the traffic zone type for each geometric region are as follows: 1) Set the traffic area type of the geometric area between two adjacent lane lines to lane by default; 2) Based on the prohibition signs in the traffic signs, such as no entry, no driving, and no motor vehicle entry signs, change the traffic area type of the geometric area affected by the sign to a passable area; 3) Based on the instructions in the traffic signs, such as signs for motor vehicle lanes, small passenger vehicle lanes, and bus lanes, if the motor vehicle does not meet the requirements of the sign for motor vehicle type, the traffic area type of the geometric area where the sign applies will be changed to a traffic zone.

[0049] The results of the common traffic area types are shown in Table 7.

[0050] Table 7. Cognitive Results of Traffic Zone Types

[0051] Step 5.2: Determine the upper and lower speed limits; According to the "Regulations for the Implementation of the Road Traffic Safety Law of the People's Republic of China": 1) Prioritize the speed indicated by traffic signs and markings, such as speed limit signs and speed limit markers, to determine the upper and lower speed limits within the geometric area; 2) When there are no speed limit signs or markings in the geometric area, the speed limit is 30 km / h when performing specific actions in the geometric area, such as turning, going down a steep slope, or making a U-turn. 3) If there are no traffic signs or markings in the geometric area, and no specific actions are performed, then the upper and lower speed limits are determined based on the cognitive results of step 4 (road type and number of lanes in the same direction). For example, the upper speed limit on a city road with only one lane for motor vehicles in the same direction is 50 km / h.

[0052] Common cognitive results for determining the upper and lower limits of speed are shown in Table 8.

[0053] Table 8. Cognitive Results of Upper and Lower Speed ​​Limits

[0054] Step 5.3: Determine the type of the left and right boundary lines; According to the national standard GB 5768.3, traffic markings can be divided into two types: dashed lines and solid lines. Therefore, the left and right boundary line types belong to the following set: ; in, For boundary line type, The line is dashed. The line is solid.

[0055] The types of left and right boundary lines are jointly determined based on the cognitive results of steps 2 and 3. For example, if the left boundary line of the current geometric area is a dashed line, but a no-overtaking sign exists in this geometric area, then the type of the left boundary line changes from dashed to solid. Traffic regulations regarding the left and right positions of motor vehicles are entirely reflected in the types of left and right boundary lines. When the boundary line is dashed, motor vehicles can cross it to enter adjacent geometric areas; when the boundary line is solid, motor vehicles cannot cross it to enter adjacent geometric areas. The cognitive results for common left and right boundary line types are shown in Table 9.

[0056] Table 9. Cognitive Outcomes of Left and Right Boundary Line Types

[0057] Step 5.4: Determine the exit direction; Based on the cognitive results from steps 2 to 4, the exit directions for each geometric region are determined. The exit direction refers to the connection between a vehicle exiting the current geometric region and entering the entrance of another geometric region, including lane keeping, going straight at an intersection, turning left at an intersection, turning right at an intersection, and making a U-turn. Therefore, the exit directions belong to the following set: ; in, To drive out of the direction, For lane keeping, To turn left at the intersection To turn right at the intersection For going straight at the intersection, To turn around.

[0058] 1) Based on the road network relationship, determine the exit direction of each geometric area, and then eliminate those that do not comply with traffic regulations; 2) Based on the understanding of traffic markings, eliminate non-compliant exit directions. For example, if the directional arrow indicates a left turn, eliminate other exit directions and retain only left turns at intersections.

[0059] 3) Eliminate non-compliant exit directions based on the understanding of traffic signs. For example, if there is a sign prohibiting left turns, then eliminate left turns at intersections.

[0060] The common results of recognizing the direction of departure are shown in Table 10.

[0061] Table 10 Cognitive Results of the Direction of Exit

[0062] Step 6: Traffic Area Recognition; Based on the merged traffic attribute information from step 5, and adhering to the principle of complete consistency in traffic rules within each traffic area, the geometric regions of the road ahead are further divided into multiple traffic areas. Since the effects (traffic attributes) and scope (i.e., geometric regions) of traffic regulations are understood, the intersection of the geometric regions of each traffic attribute forms a traffic area. Multiple traffic areas will then "patch together" to cover the entire road ahead, resulting in a traffic environment model based on traffic area representation. Figure 4 As shown in the image, the upper part is a top view of a structured road, with vehicles traveling from left to right. The lower part of the image shows the corresponding traffic environment model. The traffic areas enclosed by solid gray lines are lanes, such as traffic areas 1-2, 1-3, 2-4, and 3-2; the closed areas enclosed by dashed black lines are passage zones, such as traffic areas 1-5, 2-5, and 3-5; and the closed areas enclosed by dashed gray lines are no-entry zones, such as traffic areas 1-1, 1-5, and 2-1.

[0063] Step 7: Understanding Connections; Based on the road network relationships identified in Step 4 and the traffic areas identified in Step 6, determine the entrance and exit connections of each traffic area; based on the left and right adjacency relationships of each lane identified in Step 4 and the traffic areas identified in Step 6, determine the left and right adjacency relationships of each traffic area. For example... Figure 4 As shown, traffic area 2-2 is adjacent to traffic area 2-1 on the left and traffic area 2-3 on the left. The entrance connects to traffic area 1-2 and the exit connects to traffic area 3-2.

[0064] Step 8: Traffic light recognition; According to GB 14886, motor vehicle traffic lights are divided into three types: red, yellow, and green, therefore they belong to the following set: ; in, For the color of motor vehicle signal lights, It is red. It is yellow. It is green.

[0065] The system can recognize the current traffic light color for motor vehicles and change the type of the relevant traffic area according to the current traffic light color, as shown in Table 11.

[0066] Table 11 Cognitive Results of Traffic Lights

[0067] Step 9: Traffic participant perception; We recognize the types of traffic participants, including pedestrians, motor vehicles, and trucks. Traffic participant types are categorized into two types: motor vehicles and non-motor vehicles. Therefore, traffic participant types belong to the following set: ; in, For traffic participant types, For motor vehicles, It is a non-motorized vehicle.

[0068] The attribute data of traffic participants include: ID, type, bounding box size, position, attitude, speed, angular velocity, acceleration, and angular acceleration. All of these attributes are represented in the road coordinate system. Specific examples are shown in Table 12.

[0069] Table 12 Attribute data of traffic participants

[0070] Traffic participants are indexed according to their type within the traffic area of ​​their location. The type of the relevant traffic area is changed according to the type of traffic participants within the traffic area. For example, if there are non-motorized vehicle types among the traffic participants in the traffic area of ​​a pedestrian crossing, the traffic area type is changed from lane to passage zone; if there are no non-motorized vehicle types among the traffic participants in the traffic area of ​​a pedestrian crossing, the traffic area type remains lane. Common examples are shown in Table 13.

[0071] Table 13 Results of the role of traffic participants

[0072] Step 10: Obstacle Recognition; The system identifies obstacles that impede the vehicle's operation, i.e., objects that do not have the ability to move independently, and indexes each obstacle into the traffic area where it is located.

[0073] The obstacle's attribute data includes: number, bounding box size, position, attitude, velocity, angular velocity, acceleration, and angular acceleration. All of these attributes are represented in the road coordinate system, as shown in Table 14.

[0074] Table 14 Attribute Data of Obstacles

[0075] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A traffic environment cognition method for structured roads for autonomous driving, characterized in that, Includes the following steps: Step 1: Road segment cognition. Determine the cognition distance based on the current environment, and represent the road sequence within the cognition distance, including intersections, as multiple one-way roads in the road coordinate system based on the travel task. Step 2: Traffic marking recognition. Analyze traffic markings and recognize them as traffic attributes related to the position, speed, or direction of travel of motor vehicles, and index the traffic attributes to the geometric region in which they apply. Step 3: Traffic sign recognition. Analyze traffic signs and recognize them as traffic attributes related to the position, speed, size, mass, or direction of travel of motor vehicles, and index the traffic attributes to the geometric area in which they apply. Step 4: Road structure recognition, obtaining the road network connection relationship, number of lanes, lane adjacency relationship and road type of the road segment ahead; Step 5: Traffic attribute fusion. The cognitive results of steps 2-4 are fused to determine the final traffic attributes of each geometric region. The final traffic attributes include at least the traffic region type, upper speed limit, lower speed limit, left boundary line type, right boundary line type, and departure direction. Step 6: Traffic area recognition. Based on the principle of completely consistent traffic attributes, each geometric area is divided into multiple traffic areas. The multiple traffic areas are spliced ​​together to form a traffic environment model of the road segment ahead. Step 7: Connectivity Recognition. Based on the road network connectivity and lane adjacency relationships, determine the entrance connectivity, exit connectivity, left-side adjacency, and right-side adjacency relationships for each traffic area. Step 8: Traffic light recognition, identifying traffic light colors, and dynamically adjusting the traffic area type of the relevant traffic area based on the colors; Step 9: Traffic participant recognition, identifying the type and status of traffic participants, indexing them to their respective traffic zones, and dynamically adjusting the traffic zone type based on the type of the traffic participant; Step 10: Obstacle Recognition, identify obstacles and index them to the traffic area where they are located.

2. The traffic environment cognition method for structured roads for autonomous driving according to claim 1, characterized in that, In step 1, the intersection is recognized as a one-way road with a single entrance and a single exit, which are determined by the travel task.

3. The traffic environment cognition method for structured roads for autonomous driving according to claim 1, characterized in that, In step 5, the steps for determining the traffic area type of the geometric region are as follows: Set the traffic area type of the geometric area between two adjacent lane lines to lane. Based on the prohibition signs within the traffic signs, the type of the corresponding geometric area is changed to a traffic zone; If the vehicle does not meet the requirements of the sign according to the instructions in the traffic sign, the traffic area type of the geometric area where the sign applies will be changed to a passage zone. Areas that do not meet the conditions for motor vehicle driving are defined as restricted zones.

4. The traffic environment cognition method for structured roads for autonomous driving according to claim 1, characterized in that, In step 5, the determination of the upper and lower speed limits follows the following priority order: Prioritize the speed values ​​indicated by traffic signs or markings; If there are no clear speed signs or markings, the speed limit will be set to a preset value within the geometric area where turning, U-turn, or steep downhill actions are performed; If there are neither speed signs or markings nor specific behaviors, then the legally mandated default speed limit and speed minimum are determined based on the road type and the number of lanes in the same direction.

5. The traffic environment cognition method for structured roads for autonomous driving according to claim 1, characterized in that, In step 5, the specific steps for determining the exit direction are as follows: All possible exit directions are obtained based on the road network relationship; Based on the directional arrow traffic markings, retain the outbound direction that is consistent with the arrow direction and eliminate other directions; Based on the prohibited traffic signs, the prohibited directions are eliminated from the set of all possible directions.

6. The traffic environment cognition method for structured roads for autonomous driving according to claim 1, characterized in that, In step 8, when the traffic light is red or yellow, the type of the traffic area in front of the associated stop line is changed from lane to passage zone; when the traffic light is green, its type is restored to lane.

7. The traffic environment cognition method for structured roads for autonomous driving according to claim 1, characterized in that, In step 9, when a non-motorized vehicle type traffic participant is identified as being located within the pedestrian crossing area, the type of the pedestrian crossing area is changed from lane to passage zone.

Images