WO2020135742A1 - 自动驾驶车辆的横向决策系统及横向决策确定方法 - Google Patents

自动驾驶车辆的横向决策系统及横向决策确定方法 Download PDF

Info

Publication number
WO2020135742A1
WO2020135742A1 PCT/CN2019/129284 CN2019129284W WO2020135742A1 WO 2020135742 A1 WO2020135742 A1 WO 2020135742A1 CN 2019129284 W CN2019129284 W CN 2019129284W WO 2020135742 A1 WO2020135742 A1 WO 2020135742A1
Authority
WO
WIPO (PCT)
Prior art keywords
lane
target
road
obstacle avoidance
autonomous vehicle
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2019/129284
Other languages
English (en)
French (fr)
Inventor
张凯
葛建勇
甄龙豹
和林
常仕伟
王天培
刘宏伟
刘洪亮
卜玉帅
高健
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Great Wall Motor Co Ltd
Original Assignee
Great Wall Motor Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Great Wall Motor Co Ltd filed Critical Great Wall Motor Co Ltd
Publication of WO2020135742A1 publication Critical patent/WO2020135742A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W60/00—Drive control systems specially adapted for autonomous road vehicles
    • B60W60/001—Planning or execution of driving tasks
    • B60W60/0011—Planning or execution of driving tasks involving control alternatives for a single driving scenario, e.g. planning several paths to avoid obstacles
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W30/00—Purposes of road vehicle drive control systems not related to the control of a particular sub-unit, e.g. of systems using conjoint control of vehicle sub-units
    • B60W30/10—Path keeping
    • B60W30/12—Lane keeping
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W30/00—Purposes of road vehicle drive control systems not related to the control of a particular sub-unit, e.g. of systems using conjoint control of vehicle sub-units
    • B60W30/18—Propelling the vehicle
    • B60W30/18009—Propelling the vehicle related to particular drive situations
    • B60W30/18163—Lane change; Overtaking manoeuvres
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W40/00—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models
    • B60W40/02—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models related to ambient conditions
    • B60W40/06—Road conditions
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W2552/00—Input parameters relating to infrastructure
    • B60W2552/05—Type of road, e.g. motorways, local streets, paved or unpaved roads
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W2552/00—Input parameters relating to infrastructure
    • B60W2552/10—Number of lanes
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W2552/00—Input parameters relating to infrastructure
    • B60W2552/53—Road markings, e.g. lane marker or crosswalk
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W2554/00—Input parameters relating to objects
    • B60W2554/20—Static objects

Definitions

  • the invention relates to the technical field of automatic driving, and in particular to a lateral decision system and a lateral decision determination method of an autonomous driving vehicle.
  • Autonomous vehicle refers to an intelligent vehicle that senses the road environment through an onboard sensor system, automatically plans driving routes, and controls the vehicle to reach a predetermined destination. It relies on the Autonomous Driving System (ADS) to achieve its functions. According to the development and design process of ADS, ADS can be divided into five parts: environment awareness system, data fusion system, decision system, control system and execution system.
  • ADS Autonomous Driving System
  • the environment awareness system is used to extract the current driving environment information of vehicles, pedestrians, roads, traffic signs and other vehicles through the on-board sensing system;
  • the data fusion system is used to filter, correlate, track, filter and other data of different sensors in order to Obtain more accurate information on roads, environmental objects, etc.;
  • the decision system is used to logically determine the vehicle behavior of unmanned vehicles based on the driving status, road, and environmental information of different environmental vehicles output by the data fusion system;
  • the control system uses Based on the information output by the data fusion system and decision system, the current horizontal and vertical control changes of the vehicle are calculated and output in real time;
  • the execution system is used to replace the driver's operation of the steering wheel, acceleration and deceleration pedals of the vehicle according to the steering, acceleration and other control variables output by the control system .
  • the decision-making system judges and outputs the lateral and vertical vehicle behaviors of the autonomous vehicle based on the input environmental object targets, roads and other information.
  • the lateral vehicle behaviors include lane keeping, lane changing, abnormal lane changing, etc.
  • the performance is acceleration, deceleration, etc.
  • lane keeping, lane changing and abnormal lane changing are the main behaviors of vehicles during driving.
  • the correct control of these three behaviors by the control system plays a decisive role in driving safety. Therefore, how the decision-making system can correctly judge the lateral behavior of vehicles such as lane keeping, lane-changing, and abnormal lane-changing is an important factor to be considered when designing a vehicle decision-making system.
  • the present invention aims to propose a lateral decision-making system for autonomous vehicles to achieve correct judgment of lateral behavior of the vehicle.
  • a lateral decision-making system for an autonomous vehicle includes an evaluation unit for evaluating target lanes and lane abnormalities required by the autonomous vehicle for lateral decision-making based on road feature information and pre-selected target lines and environmental object targets And a judging unit for judging and outputting the expected lateral behavior of the self-driving vehicle based on the target lane and lane abnormalities evaluated by the evaluation unit in combination with the road feature information, wherein the expected lateral behavior includes the lane Any of keep, change lane and abnormal change lane.
  • the evaluation unit includes: a target lane management module for selecting a target lane of the self-driving vehicle according to the road feature information, wherein the principle of selecting the target lane includes following the principle of road scenes and following the attributes of lanes The principle of not selecting abnormal lanes, and the principle of selecting adjacent lanes and selecting in turn to the right when the lane is abnormal, wherein the road feature information includes the road type, road feature points and the lane attributes, and the lane attributes include A lane feature point attribute and a lane number attribute; and a lane abnormality management module, used to identify an abnormal lane based on the road characteristic information, and provide an obstacle avoidance strategy for the abnormal lane.
  • a target lane management module for selecting a target lane of the self-driving vehicle according to the road feature information, wherein the principle of selecting the target lane includes following the principle of road scenes and following the attributes of lanes The principle of not selecting abnormal lanes, and the principle of selecting adjacent lanes and selecting in turn to the right when the lane is abnormal, where
  • the target lane management module includes: a main lane target lane selection sub-module for selecting a target lane according to the selection principle when the self-driving vehicle is driving in a normal scene of the main lane, wherein the main lane Conventional scenes include acceleration lanes, normal driving lanes, and deceleration lanes, and are used to select target lanes based on changes in the number of lanes of the road ahead relative to the current road when the autonomous vehicle is driving on the special scene of the main lane.
  • the special scene of the main road includes a main road narrowing, a main road widening, a main road bifurcation and/or a tunnel; and a ramp target lane selection sub-module, which is used when the autonomous vehicle is driving on a ramp scene according to the road ahead
  • the target lane is selected with respect to the change of the number of lane attributes of the current road, where the ramp scene includes a regular ramp, a ramp narrows, a ramp widens, a ramp bifurcation, and/or a ramp intersection.
  • the lane abnormality management module includes: a lane abnormality recognition sub-module for analyzing road feature information to filter out static obstacle targets on the road ahead of the self-driving vehicle, and identify whether a lane abnormality is based on the static obstacle targets And an obstacle avoidance sub-module for guiding the autonomous vehicle to avoid obstacles when the lane is abnormal.
  • the obstacle avoidance sub-module for guiding the autonomous vehicle to avoid obstacles when the lane is abnormal includes: determining obstacle avoidance targets according to the static obstacle targets and the dynamic environmental object targets existing in a set area, and Determine the static and dynamic characteristics of the obstacle avoidance target relative to the autonomous vehicle; establish an obstacle avoidance area adapted to road characteristics based on the static and dynamic characteristics of the obstacle avoidance target; based on the static of the obstacle avoidance target Characteristics and dynamic characteristics to determine the accessibility of the obstacle avoidance area; conduct a collision risk assessment on the relevant environmental object targets during the normal lane change of the autonomous vehicle, and determine the feasibility of lane change according to the result of the collision risk assessment And according to the feasibility of the lane change and the accessibility of the obstacle avoidance area, control the autonomous vehicle to change lanes or to bypass the obstacle avoidance target in the current driving lane.
  • the obstacle avoidance sub-module is used to control the autonomous vehicle to change lanes or to bypass the obstacle avoidance target in the current driving lane according to the feasibility of the lane change and the accessibility of the obstacle avoidance area Including: if the lane change is feasible, control the autonomous vehicle to change lanes; otherwise, determine the passability of the obstacle avoidance area; if the obstacle avoidance area is accessible, determine that the autonomous vehicle is in the current driving lane Drive around the obstacle avoidance target.
  • the lateral decision-making system of the self-driving vehicle has the following advantages: it can evaluate the target lane and lane abnormalities, and accordingly make lane keeping, lane change, or abnormal lane change in accordance with the characteristics of the road Lateral decision, so that the vehicle's control system can perform adaptive lateral control based on the lateral decision to ensure the vehicle's driving safety.
  • Another object of the present invention is to propose a method for determining the lateral decision of an autonomous driving vehicle, so as to realize the correct judgment of the lateral behavior of the vehicle.
  • a method for determining a lateral decision of an autonomous vehicle includes: evaluating target lanes and lane abnormalities required by the autonomous vehicle for lateral decision based on road feature information and pre-selected target lines and environmental object targets; and combining The road characteristic information determines and outputs the expected lateral behavior of the autonomous vehicle according to the evaluated target lane and lane abnormality, wherein the expected lateral behavior includes any one of lane keeping, lane changing, and abnormal lane changing.
  • the evaluation of the target lane and the lane abnormality required by the autonomous vehicle for lateral decision-making includes: selecting the target lane of the autonomous vehicle according to the road feature information, wherein the selection principle of the target lane includes follow the principles of road scenes, the principles of lane attributes, the principle of not selecting abnormal lanes, and the principle of selecting adjacent lanes and selecting in turn to the right when lanes are abnormal, where the road feature information includes the road type, road feature points, and the A lane attribute, and the lane attribute includes a lane feature point attribute and a lane number attribute; and an abnormal lane is identified according to the road characteristic information, and an obstacle avoidance strategy for the abnormal lane is provided.
  • the selection principle of the target lane includes follow the principles of road scenes, the principles of lane attributes, the principle of not selecting abnormal lanes, and the principle of selecting adjacent lanes and selecting in turn to the right when lanes are abnormal
  • the road feature information includes the road type, road feature points, and the A lane attribute
  • the lane attribute includes a lane feature
  • the selecting the target lane of the self-driving vehicle according to the road feature information includes: selecting the target lane according to the selection principle when the self-driving vehicle is driving in a normal scene of the main lane, wherein the main lane Conventional lane scenes include acceleration lanes, normal driving lanes, and deceleration lanes; when the autonomous driving vehicle is driving on a special scene of the main lane, the target lane is selected according to the change in the number of lane attributes of the road ahead relative to the current road, where the master Special road scenes include main road narrowing, main road widening, main road bifurcation and/or tunnel; and when the autonomous vehicle is driving on a ramp scene, according to the change in the number of lanes of the road ahead relative to the current road.
  • the identifying an abnormal lane according to the road feature information and providing an obstacle avoidance strategy for the abnormal lane includes: analyzing the road feature information to screen out the static obstacle target of the road ahead of the autonomous vehicle, and based on the The static obstacle target recognizes whether the lane is abnormal; and when the lane is abnormal, guides the autonomous vehicle to avoid obstacles.
  • the guiding the autonomous vehicle to avoid obstacles includes determining an obstacle avoidance target according to the static obstacle target and the dynamic environment object target existing in a set area, and determining that the obstacle avoidance target is relative to all The static and dynamic characteristics of the self-driving vehicle; based on the static and dynamic characteristics of the obstacle avoidance target to establish an obstacle avoidance area adapted to the road characteristics; based on the static and dynamic characteristics of the obstacle avoidance target, determine the avoidance The accessibility of the obstacle area; perform a collision risk assessment on the relevant environmental object targets during the normal lane change of the autonomous vehicle, and determine the lane change feasibility based on the result of the collision risk assessment; and according to the lane change feasibility And the passability of the obstacle avoidance area, controlling the autonomous vehicle to change lanes or to bypass the obstacle avoidance target in the current driving lane.
  • controlling the autonomous vehicle to change lanes or to bypass the obstacle avoidance target in the current driving lane according to the feasibility of the lane change and the accessibility of the obstacle avoidance area includes: If feasible, control the automatic driving vehicle to change lanes, otherwise judge the passability of the obstacle avoidance area, and if the obstacle avoidance area is passable, determine that the automatic driving vehicle bypasses the avoidance in the current driving lane Obstacle goal.
  • the method for determining the lateral decision of the self-driving vehicle has the same advantages as the above-mentioned lateral decision system over the prior art, and will not be repeated here.
  • Another object of the present invention is to propose a machine-readable storage medium to realize the correct judgment of the lateral behavior of the vehicle.
  • a machine-readable storage medium having instructions stored on the machine-readable storage medium is used to cause a machine to execute the above-mentioned method for determining a lateral decision of an autonomous vehicle.
  • FIG. 1 is a schematic diagram of the region division of the vehicle environment in the vehicle body coordinate system according to an embodiment of the present invention
  • FIG. 2 is a schematic structural diagram of a lateral decision system for an autonomous driving vehicle according to an embodiment of the present invention
  • FIG. 3 is an exemplary diagram of target lane selection in a normal driving lane in an embodiment of the present invention
  • FIG. 4(a)-FIG. 4(c) are schematic diagrams of the main road narrowing, the main road widening and the main road bifurcation in the embodiment of the present invention.
  • FIG. 5 is an exemplary diagram of lane abnormality judgment in an embodiment of the present invention.
  • FIG. 6 is an exemplary diagram of lane abnormality recognition of a multi-static obstacle lane in the current lane in an embodiment of the present invention
  • FIG. 7 is a schematic diagram of the vehicle performing obstacle avoidance in the embodiment of the present invention.
  • FIG. 8 is a schematic diagram of a hardware layout of an automatic driving vehicle according to an embodiment of the present invention.
  • FIG. 9 is a schematic flowchart of a method for determining a lateral decision of an autonomous driving vehicle according to an embodiment of the present invention.
  • the “environmental object target” mentioned in the embodiments of the present invention may refer to any object that is moving or stationary in front of, behind, or to the side of the vehicle, for example, vehicles, people, buildings, etc.
  • the “target line” mentioned may Refers to the lane center line, dynamic target line or safety offset line required for lateral decision-making and lateral control of autonomous vehicles (hereinafter referred to as vehicles).
  • the vehicle follows the target line, and the "target lane” corresponds to the "target line”.
  • the decision system will make a decision that the vehicle is driving in the target lane.
  • the “lane anomaly” in the embodiment of the present invention is mainly a target lane anomaly, indicating that the lane is impassable due to static obstacles (such as roadblocks, road cones, and vehicles that cannot move accidents) or red lights at the tunnel entrance. happening.
  • FIG. 1 is a schematic diagram of the area division of the vehicle environment in the vehicle body coordinate system according to the embodiment of the present invention, including the front area of the vehicle, the left front area, etc., the following uses the area division of FIG. 1 to explain the environment object targets, etc. Location.
  • the lateral decision-making system includes: an evaluation unit 100 for evaluating target lanes and lane abnormalities required by the vehicle for lateral decision-making based on road feature information and pre-selected target lines and environmental object targets; and
  • the judging unit 200 is configured to judge and output the expected lateral behavior of the vehicle according to the target lane and the lane abnormality evaluated by the evaluation unit 100 in combination with the road feature information.
  • the road feature information includes a road type, a road feature point and a lane attribute, and the lane attribute includes a lane feature point attribute and a number of lane attributes.
  • the expected lateral behavior includes any one of lane keeping, lane changing and abnormal lane changing.
  • lane keeping means that the vehicle travels along the current lane
  • lane change means that the vehicle moves into the adjacent lane to the left or right.
  • abnormal lanes that is, when the lane keeping and lane changing conditions are not met in front of the lane, the vehicle enters an abnormal lane change (obstacle avoidance state).
  • the three expected horizontal behaviors will be introduced in conjunction with examples below, and will not be repeated here.
  • the evaluation unit 100 includes: a target lane management module 110 for selecting the target lane of the self-driving vehicle according to the road feature information; and a lane abnormality management module 120 for Road feature information identifies abnormal lanes and provides obstacle avoidance strategies for abnormal lanes.
  • the selection principles of the target lane include the principle of following the road scene, the principle of following the attributes of the lane, the principle of not selecting the abnormal lane, and the principle of selecting the adjacent lane and selecting in turn to the right when the lane is abnormal.
  • the principle of following the road scene means that the target lane is to be considered whether the road is the main road or the ramp
  • the principle of following the lane attribute is to consider the change of the lane type when selecting the target lane (judging by the characteristic points of the lane, such as driving in Accelerated lanes) and the number of lanes change.
  • the principle of not selecting abnormal lanes means that the abnormal lanes cannot be used as the target lanes.
  • the principle of selecting adjacent lanes and the order of selection to the right means that when the lanes are abnormal, the neighbors are preferentially selected. For lanes, if multiple lanes are abnormal, select adjacent lanes to the right in turn. It should be noted that the embodiments of the present invention are not limited to these selection principles. In the selection of the target lane, it is necessary to consider more factors in combination with the actual situation. The following will illustrate the four selection principles and some of them here by examples Other selection principles.
  • the target lane management module 110 includes: a main lane target lane selection sub-module 111 for selecting according to the selection principle when the autonomous vehicle is driving in a normal scene of the main lane
  • the target lane is also used to select the target lane according to the change of the number of lane attributes of the road ahead relative to the current road when the autonomous vehicle is driving in the special scene of the main lane; and the ramp target lane selection submodule 112 is used to When the self-driving vehicle is traveling on a ramp scene, the target lane is selected according to the change in the attribute of the number of lanes of the road ahead relative to the current road.
  • the normal scenes of the main lane include acceleration lanes, normal driving lanes, and deceleration lanes. These three lanes belong to the normal lane of the vehicle, and these three lanes can be identified by the lane attributes.
  • the acceleration lane (following the attributes of the lane).
  • the target lane should select the rightmost lane.
  • the target lane is abnormal, it should Select the adjacent lane of the original target lane, and according to the characteristics of the acceleration lane on the right side of the road (follow the road scene), try to select the target lane to the right.
  • the section of the road between the normal driving lane and the starting point of the deceleration lane is less than the warning distance, and the planned deceleration lane on the high speed is called the deceleration lane (following the attributes of the lane).
  • the target lane should be replaced with the far right lane to prepare for entering the deceleration lane and the ramp part in advance.
  • the adjacent lane of the original target lane should be selected, and as far as possible to the right, so that the vehicle can choose the appropriate time to enter the deceleration lane and the ramp part and leave the road section as soon as possible.
  • the target lane selection for normal driving lanes is described in detail below.
  • the normal driving lane here refers to the section from the vehicle exiting the acceleration lane and entering the main highway of the expressway, and away from entering the deceleration lane (following the attributes of the lane), which does not include the special scene of the main lane mentioned above.
  • FIG. 3 is an example diagram of target lane selection in a normal driving lane in an embodiment of the present invention.
  • the position of the vehicle, the original target lane position, and the position of obstacles can be adapted to the following scenarios to change, and will not be shown here one by one.
  • Various changes can be understood by those skilled in the art in conjunction with the text.
  • the principles of target lane selection for normal driving lanes are as follows:
  • Two lanes (for example, only two lanes of C3 and C4): both lanes are normal, and the right lane is the target lane; only one lane is normal (for example, C3 is normal), and the normal lane is the target lane.
  • Three lanes (for example, only three lanes of C2, C3, and C4): all three lanes are normal and the middle lane is the target lane; the middle lane is abnormal and the right lane is the target lane; only one lane is normal and the normal lane is the target lane.
  • the number of lanes is greater than three: the second lane on the left is the target lane. For example, when C1-C4 are normal, select C2 as the target lane.
  • the target lane When the target lane is abnormal, select the target lane according to the principle of gradually to the right. When the abnormality disappears, return to the original target lane. As shown in the current Figure 3, the target lane should be C2. However, there are static obstacles in C2 that cause the C2 lane to be abnormal and cannot pass through. At this time, the target lane is set to the C3 lane. When the vehicle exceeds the obstacle and the C2 lane is normal , The target lane remains the C2 lane. Similarly, if the front of the C2 and C3 lanes is also abnormally unable to pass, then the target lane is placed in the C4 lane, and when there are more lanes, the analogy is in turn.
  • the target lane selects the adjacent lane on the right and selects it to the right in turn, which is helpful for parking the autonomous vehicle in the emergency lane or off the highway more quickly when there is an abnormality in the road ahead.
  • the selection of target lanes in some special scenes can be modified to make it more in line with human driving habits, such as the following selection principles:
  • the target lane is abnormal, and the current lane of the vehicle is normal, the current lane is the target lane.
  • the lane is abnormal, and the nearest normal lane is selected as the target lane.
  • the left and right sides are the same, select the right side as the target lane.
  • the original target lane is C2.
  • C2 and C3 in front are abnormal. Therefore, the vehicle is at C2 and is closer to C1. Therefore, the target lane is placed at C1.
  • target lane selection is not limited to the number of lanes, and any principle that conforms to the above scenario can be adopted.
  • the special scene of the main road mainly includes the main road narrowing, the main road widening, the main road bifurcation (separate subgrade) and/or tunnel.
  • the principle of selecting the target lane under the scenario of narrowing the main lane is: changing the lane attributes (normal lane ⁇ narrowing lane) by 1000m (standard amount) in advance; if the original target lane is a road narrowing lane, the original target lane is set The adjacent normal lane is the target lane.
  • FIG. 4(b) is a schematic diagram of the main road widening in the embodiment of the present invention, where the main road widening refers to the automatic driving vehicle driving on the main road, the number of front lanes increases, which includes left side widening and right side widening Three cases of wide and widening on both sides.
  • the principle of selecting the target lane in the scenario of widening the main lane is: change the lane attributes (normal lane ⁇ widening lane) by 500m (calibration value) in advance; the vehicle travels along the current target lane until it enters the widening area , The number of lanes changes, re-select the target lane.
  • the current number of lanes is 2
  • the target lane is the rightmost lane
  • the vehicle travels along the current road
  • the attribute of the number of lanes where the vehicle is located changes from 2 to 3
  • the target lane is C2.
  • the number of lanes becomes 3 after the left side becomes wider and the right side becomes wider. After the two sides become wider, the number of lanes becomes 4. According to the changed number of lanes, the principles mentioned above are used again Select the target lane.
  • FIG. 4(c) is a schematic diagram of a main road bifurcation in an embodiment of the present invention, where the main road bifurcation is also called a split roadbed, and the road points in two different directions, generally accompanied by changes in the number of lane attributes.
  • the principle of selecting the target lane under the main road bifurcation scenario includes: changing the lane attributes (normal lane ⁇ main road bifurcation) by 500m (calibration value) in advance; taking the 4-lane target direction to the right (the scenario of left) (Similar to this) For example, when there is 1 lane in front of the target side, the lane is the target lane, when there are 2 lanes in front of the target side, the right lane is the target lane, and when there are 3 lanes in front of the target side, the middle lane is Target lane.
  • the selection principle of the target lane of the tunnel is the same as or similar to the normal driving lane corresponding to FIG. 3 described above, and details are not described herein again.
  • the ramp scenario includes a conventional ramp, a ramp narrows, a ramp widens, a ramp bifurcation, and/or a ramp intersection.
  • the main lanes corresponding to the above Figures 4(a)-4(c) are narrowed, the main lane is widened, and the main lane is branched for the target lanes of the ramp narrowing, ramp widening, and ramp bifurcation
  • the target lanes are selected to be the same or similar, the difference is mainly that the main lane becomes a ramp, and those skilled in the art can understand it based on the ramp road conditions, so they will not repeat them here.
  • the target lane selects the rightmost lane.
  • the target lane is selected to be close to the rightmost lane, and the principle of target lane selection follows the principle of keeping to the right as much as possible.
  • ramp merge For the intersection of ramps, or ramp merge, it means that ramps in different directions merge into one.
  • the vehicle is driving on a ramp, and the lane attribute (normal ramp ⁇ interchange ramp) is changed by 500m (calibrated value) in advance.
  • the attribute of the ramp lane number changes.
  • the vehicle runs along the current target lane and merges into the intersection ramp. After the change, re-select the target lane according to the new lane number.
  • the target lane management module 110 selects the target lanes of the main lane and the ramp according to laws and regulations on different speed limits of different lanes on the highway so that the vehicle travels at a faster speed in a predetermined direction
  • Priority driving lanes, and the priority driving lanes are planned to avoid collision hazards caused by large lateral deviations of the vehicle due to inaccurate map positioning, ensuring that the vehicle can drive at a faster speed under the premise of safety , And its target lane selection planning method conforms to people's driving habits.
  • the lane abnormality management module 120 includes: a lane abnormality recognition sub-module 121 for analyzing road feature information to filter out static obstacle targets of the road ahead of the autonomous vehicle, And identify whether the lane is abnormal based on the static obstacle target; and the obstacle avoidance sub-module 122 is used to guide the autonomous vehicle to avoid obstacles when the lane is abnormal.
  • the lane abnormality recognition sub-module 121 is also used to provide the target lane management module 110 with the identified lane abnormality information, so that the target lane management module 110 selects the target lane in combination with the lane abnormality.
  • the lane abnormality recognition sub-module 121 should include three parts: static obstacle target selection, lane abnormality judgment, and lane static multi-static obstacle lane abnormality recognition.
  • the principle of static obstacle target selection includes: extracting road feature information (number of lanes, width of each lane, etc.) of the current driving section of the vehicle, road attachment information, and environmental object target information.
  • road feature information number of lanes, width of each lane, etc.
  • environmental object target information use the nearest environmental object target from the vehicle as a reference to filter static obstacle targets (also called static obstacles) in each lane within a certain range.
  • the static obstacle targets are mainly static object targets such as road cones, roadblocks, and faulty vehicles.
  • the horizontal and vertical distance information of each static obstacle target relative to the own vehicle can be extracted for each lane.
  • FIG. 5 is an exemplary diagram of lane abnormality judgment in the embodiment of the present invention, which takes the current lane where the vehicle is located as an example, and the principle of abnormality judgment of other lanes is similar to this.
  • the driving area of the autonomous vehicle is as shown in ABCE, and the driving width D is the static target 1 and the center of the lane in the range of D2.
  • the vehicle When the vehicle is driving at the tunnel entrance, it is also necessary to identify the traffic lights of each lane of the tunnel entrance.
  • the lane When the lane is a red light, the lane is set to an abnormal lane (from the entrance to the exit is abnormal); until the autonomous vehicle is driven out
  • the system re-identifies whether the road status is a tunnel and re-identifies the traffic lights.
  • FIG. 6 is an exemplary diagram of lane abnormality recognition of a multi-static obstacle lane in the present lane in an embodiment of the present invention.
  • the obstacle avoidance sub-module 122 which is used to guide the autonomous vehicle to avoid obstacles when the lane is abnormal, it may mainly include the following steps: according to the static obstacle target and the dynamic environment objects existing in the set area
  • the target determines the obstacle avoidance target, and determines the static characteristics and dynamic characteristics of the obstacle avoidance target relative to the autonomous vehicle;
  • the obstacle avoidance area adapted to the road characteristics is established based on the static characteristics and dynamic characteristics of the obstacle avoidance target;
  • the static and dynamic characteristics of the obstacle avoidance target to determine the passability of the obstacle avoidance area; to perform collision risk assessment on relevant environmental object targets during normal lane change of the autonomous vehicle, and according to the collision risk assessment
  • the result of is to determine the feasibility of lane change; and according to the feasibility of the lane change and the accessibility of the obstacle avoidance area, control the autonomous vehicle to change lanes or to bypass the obstacle avoidance target in the current driving lane.
  • the functions implemented by the obstacle avoidance sub-module 122 mainly include the following parts.
  • Obstacle avoidance targets include static obstacles and dynamic obstacles. The selection principle is based on the target of the closest object in the area to the autonomous vehicle.
  • Static obstacles are mainly static object targets such as road cones, roadblocks, and faulty vehicles. Obstacle avoidance targets include: 1 static object targets in the front area; 2 static object targets in the left front area; 3 static object targets in the right front area; 4 left side Area static object target; 5 Static object target in the right side area.
  • Dynamic obstacles are mainly moving object targets.
  • Obstacle avoidance targets include: 1Dynamic object targets in the area in front of the front are lower than the speed of the autonomous vehicle; 2Dynamic object targets in the area in the front left are lower than the speed of the autonomous vehicle; 3The area in the front right is low Dynamic object targets for the speed of autonomous vehicles; 4 Dynamic object targets for the left side area; 5 Dynamic object targets for the right side area; 6 Dynamic object targets for the left rear area higher than the speed of the autonomous vehicle; 7 Right rear area high A dynamic object target for the speed of autonomous vehicles.
  • the traditional obstacle avoidance area establishment method usually establishes a fan-shaped area, using 1/2 of the fan-shaped angle as a deflection posture, and avoids obstacles before following.
  • This method is suitable for low-speed autonomous driving vehicles such as city/rural roads without lane lines.
  • the road characteristics should be considered in the establishment of the obstacle avoidance area, so that the obstacle avoidance behavior of the automatic driving conforms to the behavior requirements of the highway for the driver (such as driving in this lane, except overtaking There is no line pressing outside; no dragon driving; no speed, no speed, etc.).
  • FIG. 7 is a schematic diagram of obstacle avoidance of a vehicle in an embodiment of the present invention, wherein the area formed by ABCD is the obstacle avoidance area, the arc length of arcs AC and BD is equal to 200 meters, and the curvature is equal to the curvature of lane line L2, that is, AC and BD Parallel to the road, the size of the area is determined by obstacle avoidance targets G1 and G2.
  • the target G1 is a dynamic object target in the area directly in front of it.
  • the relationship between G1 and the vehicle includes the outer contour points, that is, the lateral closest point G11 and the longitudinal closest point G12.
  • the curve s1 parallel to the road is constructed by G11, and the longitudinal closest point G12 to The intersection point of the vertical line of curve s1 is G13, and G13 is used as the outer contour point of obstacle avoidance for target G1.
  • a safety distance of d2 0.3 m is added to generate a BD curve.
  • the target G2 is a static object target (roadblock) in the front left area.
  • the vehicle width W+safe distance d3 is the most passable judgment condition.
  • the width of the obstacle avoidance area is greater than (W+d3), it is automatically Obstacle avoidance can be carried out by driving the vehicle; otherwise, the autonomous vehicle re-judges other obstacle avoidance areas (such as whether obstacle avoidance areas can be generated on the right).
  • the obstacle avoidance sub-module 122 is used to control the autonomous vehicle to change lanes or bypass the current driving lane based on the feasibility of the lane change and the accessibility of the obstacle avoidance area
  • the obstacle avoidance target driving includes: if the lane change is feasible, control the automatic driving vehicle to change lanes; otherwise, judge the passability of the obstacle avoidance area; if the obstacle avoidance area is passable, determine the automatic driving The vehicle travels around the obstacle avoidance target in the current driving lane.
  • the obstacle avoidance sub-module 122 is used to identify whether a lane change is required, including the following three aspects: generation of a lane change intention (that is, collision risk assessment), lane change direction judgment, and lane change feasibility judgment.
  • the automatic driving vehicle determines whether the vehicle needs to change lanes according to the relative distance between the vehicle and the preceding vehicle and the speed tradeoff, and reduces the frequency of the lane change of the automatic driving vehicle.
  • the autonomous vehicle Assuming that the threshold for the lane change intention expectation factor is set to ⁇ , the speed of the autonomous vehicle V_auto, the target vehicle speed V_trg, the relative distance between the autonomous vehicle and the target vehicle Dis_rely, the autonomous vehicle expects a safe driving distance K*V_auto where K takes priority 0.8 .
  • the self-driving vehicle When a self-driving vehicle runs normally and a static object appears in front of the detection area, the self-driving vehicle should change lanes in advance to avoid collision with the static object in front.
  • the autonomous vehicle Assuming that the threshold of the lane change intention expectation factor is set to ⁇ s, the speed of the autonomous vehicle V_auto, the relative distance Dis_s between the autonomous vehicle and the static obstacle, the autonomous vehicle expects the safe driving distance K*V_auto, where K takes priority 1.
  • the lane-changing intention expectation factor ⁇ s K1*(Dis_s/K*V_auto), where K1 preferentially takes the value 1.
  • K1 preferentially takes the value 1.
  • TTC is the time between the collision of the autonomous vehicle and the front vehicle
  • the left-hand lane change of the self-driving vehicle is prioritized, that is, when the left-front and right-front areas meet the conditions a to f at the same time, the left lane is preferentially selected as the target lane.
  • the self-driving vehicle determines the target lane for lane change based on the above conditions a) to g).
  • Vehicles must strictly comply with road traffic regulations, such as virtual and solid lines, speed limits, lights, horns, traffic lights, and no U-turns.
  • the obstacle avoidance sub-module 122 of the embodiment of the present invention proposes an obstacle avoidance method suitable for vehicles traveling at high speeds and structured roads, which can avoid manual driving that may cause vehicle collisions due to blind spots, and its lane-changing function can improve vehicle driving efficiency and reduce drivers
  • the workload, and the involved automatic lane changing method has a wide range of application, and can be applied to automatic driving systems under curved roads with a large curvature and straight roads, especially under structured roads.
  • the lane abnormality management module 120 of the embodiment of the present invention can recognize the lane situation, and can actively guide the vehicle to avoid obstacles in advance or gradually approach the emergency lane or drive away from the highway to avoid the risk of collision of vehicles.
  • FIG. 8 is a schematic diagram of a hardware arrangement of an autonomous driving vehicle according to an embodiment of the present invention, wherein the decision-making system of the autonomous driving vehicle includes the lateral decision-making system of the foregoing embodiment.
  • control unit 1 the control unit 2, and the control unit 4 constitute an environment awareness system
  • control unit 3 constitutes a lateral decision-making system according to an embodiment of the present invention, which is part of the vehicle's decision-making system.
  • the control unit 1 provides accurate location information for autonomous vehicles, and high-precision GPS+IMU equipment is preferred, with a lateral positioning deviation within 10 cm and a longitudinal positioning deviation within 30 cm.
  • the control unit 2 is used to store and output high-precision lane lines, number of lanes, lane width and other information within 200m from the front and rear of the self-driving vehicle. It preferentially uses hardware devices with storage space greater than 50G and processing memory greater than 1G.
  • the control unit 4 is used for detecting and extracting objects and objects appearing in the range of 360° around the self-driving vehicle, and preferentially selects all-weather sensor detection equipment to avoid misdetection and missed detection of objects and objects caused by rain, snow, fog, and light.
  • the control unit 4 is not limited to the current installation location or the current number.
  • several radar sensors lidar or millimeter wave radar equipment, etc.
  • visual sensors are arranged around the vehicle body. Improve the accuracy and stability of object detection.
  • the control unit 2 obtains the accurate position information of the automatic driving vehicle provided by the control unit 1, and outputs the high-precision map data within 200m in front of and behind the automatic driving vehicle in real time after processing and calculation, including: the latitude and longitude of the discrete points of the lane line (the latitude and longitude are based on the center of the earth) , Discrete point heading angle (take the clockwise direction of 0° in the north direction as evidence), lane line type, lane width, lane number, road boundary and other information, the control unit 3 will receive the lane line offline data through the Ethernet Converted to the plane vehicle coordinate system, providing the road characteristic information required during the vehicle lane change, the control unit 4 simultaneously transmits the object information of the objects in the detection area to the control unit 3 by CAN communication, and the control unit 3 executes the above-mentioned lateral decision The function of the system.
  • the lateral decision-making system of the embodiment of the present invention can evaluate the target lane and lane anomalies, and make lateral decisions on lane keeping, lane-changing, or abnormal lane-changing in accordance with road characteristics in order to facilitate the vehicle control system.
  • An adaptive lateral control can be performed based on this lateral decision to ensure the driving safety of the vehicle.
  • FIG. 9 is a schematic flow chart of a method for determining a lateral decision of an autonomous vehicle according to an embodiment of the present invention.
  • the method for determining a lateral decision is based on the same inventive idea as the lateral decision system described above.
  • the method for determining a lateral decision of an autonomous vehicle may include the following steps S100 and S200:
  • Step S100 based on the road feature information and the pre-selected target line and environmental object target, evaluate the target lane and lane abnormality required by the autonomous vehicle for lateral decision.
  • this step S100 further includes the following sub-steps:
  • Step S110 Select a target lane of the self-driving vehicle according to the road feature information.
  • the selection principle of the target lane includes the principle of following the road scene, the principle of following the attribute of the lane, the principle of not selecting the abnormal lane, and the principle of selecting the adjacent lane and selecting in turn to the right when the lane is abnormal.
  • this step S110 specifically includes: when the self-driving vehicle is driving in a normal scene of the main road, a target lane is selected according to the selection principle, wherein the normal scene of the main road includes an acceleration lane, a normal driving lane and Deceleration lane; when the self-driving vehicle is driving in the special scene of the main road, select the target lane according to the change of the attribute of the number of lanes of the road ahead relative to the current road, wherein the special scene of the main road includes the main road narrowing and the main road Widening, main road bifurcation, and/or tunnel; and when the autonomous vehicle is driving on a ramp scene, the target lane is selected according to the change in the number of lane attributes of the road ahead with respect to the current road, where the ramp scene includes conventional Ramp, ramp narrow, ramp wide, ramp bifurcation and/or ramp intersection.
  • Step S120 identify an abnormal lane according to the road feature information, and provide an obstacle avoidance strategy for the abnormal lane.
  • this step S120 specifically includes: analyzing the road feature information to filter out the static obstacle target of the road ahead of the autonomous vehicle, and identifying whether the lane is abnormal based on the static obstacle target; and when the lane is abnormal, Guiding the autonomous vehicle to avoid obstacles.
  • the guiding the autonomous vehicle to avoid obstacles includes determining an obstacle avoidance target according to the static obstacle target and the dynamic environment object target existing in a set area, and determining that the obstacle avoidance target is relative to all The static and dynamic characteristics of the self-driving vehicle; based on the static and dynamic characteristics of the obstacle avoidance target to establish an obstacle avoidance area adapted to the road characteristics; based on the static and dynamic characteristics of the obstacle avoidance target, determine the avoidance The accessibility of the obstacle area; perform a collision risk assessment on the relevant environmental object targets during the normal lane change of the autonomous vehicle, and determine the lane change feasibility based on the result of the collision risk assessment; and according to the lane change feasibility And the passability of the obstacle avoidance area, controlling the autonomous vehicle to change lanes or to bypass the obstacle avoidance target in the current driving lane.
  • controlling the autonomous vehicle to change lanes or to bypass the obstacle avoidance target in the current driving lane according to the feasibility of the lane change and the accessibility of the obstacle avoidance area includes: If the road is feasible, the automatic driving vehicle is controlled to change lanes, otherwise the passability of the obstacle avoidance area is judged, and if the obstacle avoidance area is passable, it is determined that the automatic driving vehicle bypasses the current driving lane Obstacle avoidance target driving.
  • Step S200 Combine the road feature information, determine and output the expected lateral behavior of the autonomous vehicle according to the evaluated target lane and lane abnormality.
  • the method for determining the lateral decision of the automatic driving vehicle is the same as the specific implementation details and effects of the above-described embodiment of the lateral decision system of the automatic driving vehicle, and details are not repeated herein.
  • phase change memory abbreviation of phase change random access memory, Phase Change Random Access Memory, PRAM, also known as RCM/PCRAM
  • SRAM static random access memory
  • DRAM dynamic Random access memory
  • RAM random access memory
  • ROM read only memory
  • EEPROM electrically erasable programmable read only memory
  • flash memory Flash or other memory Technology
  • CD-ROM compact disc
  • DVD digital versatile disc

Landscapes

  • Engineering & Computer Science (AREA)
  • Automation & Control Theory (AREA)
  • Transportation (AREA)
  • Mechanical Engineering (AREA)
  • Human Computer Interaction (AREA)
  • Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Traffic Control Systems (AREA)

Abstract

一种自动驾驶车辆的横向决策系统及横向决策确定方法，该横向决策系统包括：评估单元，用于根据道路特征信息以及预先选择的目标线和环境物体目标，评估自动驾驶车辆进行横向决策所需的目标车道和车道异常情况；以及判断单元，用于结合道路特征信息，根据评估单元所评估的目标车道和车道异常情况，判断并输出自动驾驶车辆的预期横向行为，其中预期横向行为包括车道保持、换道和异常换道中的任意一者。该横向决策系统及方法能够评估出目标车道及车道异常情况，并据此作出符合道路特性的横向决策，以便于车辆的控制系统可基于其决策进行适应性的横向控制。

Description

自动驾驶车辆的横向决策系统及横向决策确定方法 技术领域
本发明涉及自动驾驶技术领域，特别涉及一种自动驾驶车辆的横向决策系统及横向决策确定方法。
背景技术
自动驾驶车辆是指通过车载传感系统感知道路环境，自动规划行车路线并控制车辆到达预定目的地的一种智能车辆，其依靠自动驾驶系统(Autonomous Driving System简称ADS)实现其功能。根据ADS的开发设计过程，可将ADS分为：环境感知系统、数据融合系统、决策系统、控制系统、执行系统五部分。
具体地，环境感知系统用于通过车载传感系统提取车辆、行人、道路、交通标示等车辆当前行驶环境信息；数据融合系统用于将不同传感器数据信息进行筛选、关联、追踪、过滤等处理以便获得更为精确的道路、环境物体目标等信息；决策系统用于根据数据融合系统输出的不同环境车辆行驶状态、道路、环境信息等，通过逻辑判断输出无人驾驶车辆的车辆行为；控制系统用于根据数据融合系统及决策系统输出的信息实时计算输出当前车辆横纵向控制变化量；执行系统用于根据控制系统输出的转向、加速等控制量取代驾驶员对车辆方向盘、加、减速踏板操作过程。
更为具体地，决策系统根据输入的环境物体目标、道路等信息判断并输出自动驾驶车辆的横向、纵向车辆行为，其中横向车辆行为表现为车道保持、换道、异常换道等，纵向车辆行为表现为加速、减速等。其中，车道保持、变道和异常换道是车辆行驶中的主要行为，控制系统对这三种行为的正确控制对行车安全有着举足轻重的作用。因此，决策系统如何实现对车道保持、换道、异常换道等车辆横向行为的正确判断是设计整车决策系统时需要考虑的重要因素。
发明内容
有鉴于此，本发明旨在提出一种自动驾驶车辆的横向决策系统，以实现对车辆横向行为的正确判断。
为达到上述目的，本发明的技术方案是这样实现的：
一种自动驾驶车辆的横向决策系统，包括：评估单元，用于根据道路特征信息以及预先选择的目标线和环境物体目标，评估所述自动驾驶车辆进行横向决策所需的目标车道和车道异常情况；以及判断单元，用于结合所述道路特征信息，根据所述评估单元所评估的目标车 道和车道异常情况，判断并输出所述自动驾驶车辆的预期横向行为，其中所述预期横向行为包括车道保持、换道和异常换道中的任意一者。
进一步的，所述评估单元包括：目标车道管理模块，用于根据所述道路特征信息选择所述自动驾驶车辆的目标车道，其中所述目标车道的选择原则包括遵循道路场景的原则、遵循车道属性的原则、不选择异常车道的原则以及车道异常时选择相邻车道和依次靠右选择的原则，其中所述道路特征信息包括道路类型、道路特征点及所述车道属性，且所述车道属性包括车道特征点属性和车道数属性；以及车道异常管理模块，用于根据所述道路特征信息识别异常车道，并提供针对异常车道的避障策略。
进一步的，所述目标车道管理模块包括：主道目标车道选择子模块，用于在所述自动驾驶车辆行驶于主道常规场景时，根据所述选择原则来选择目标车道，其中所述主道常规场景包括加速车道、正常行驶车道和减速车道，以及用于在所述自动驾驶车辆行驶于主道特殊场景时，根据前方道路相对于当前道路的车道数属性的变化来选择目标车道，其中所述主道特殊场景包括主道变窄、主道变宽、主道分叉和/或隧道；以及匝道目标车道选择子模块，用于在所述自动驾驶车辆行驶于匝道场景时，根据前方道路相对于当前道路的车道数属性的变化来选择目标车道，其中所述匝道场景包括常规匝道、匝道变窄、匝道变宽、匝道分叉和/或匝道交汇。
进一步的，所述车道异常管理模块包括：车道异常识别子模块，用于分析道路特征信息以筛选出所述自动驾驶车辆的前方道路的静态障碍目标，并基于所述静态障碍目标识别是否车道异常；以及避障子模块，用于在车道异常时引导所述自动驾驶车辆进行避障。
进一步的，所述避障子模块用于在车道异常时引导所述自动驾驶车辆进行避障包括：根据所述静态障碍目标以及设定区域存在的动态的所述环境物体目标确定避障目标，并确定所述避障目标相对于所述自动驾驶车辆的静态特性和动态特性；基于所述避障目标的静态特性和动态特性建立适应于道路特性的避障区域；基于所述避障目标的静态特性和动态特性，判断所述避障区域的可通行性；对所述自动驾驶车辆正常换道时的相关环境物体目标进行碰撞风险评估，并根据所述碰撞风险评估的结果确定换道可行性；以及根据所述换道可行性和所述避障区域的可通行性，控制所述自动驾驶车辆进行换道或在当前行驶车道绕开所述避障目标行驶。
进一步的，所述避障子模块用于根据所述换道可行性和所述避障区域的可通行性，控制所述自动驾驶车辆进行换道或在当前行驶车道绕开所述避障目标行驶包括：若换道可行，则控制所述自动驾驶车辆进行换道，否则判断所述避障区域的可通行性，若所述避障区域可通 行，则确定所述自动驾驶车辆在当前行驶车道绕开所述避障目标行驶。
相对于现有技术，本发明所述的自动驾驶车辆的横向决策系统具有以下优势：能够评估出目标车道及车道异常情况，并据此作出符合道路特性的车道保持、换道或异常换道的横向决策，以便于车辆的控制系统可基于该横向决策进行适应性的横向控制，以保证车辆的行驶安全。
本发明的另一目的在于提出一种自动驾驶车辆的横向决策确定方法，以实现对车辆横向行为的正确判断。
为达到上述目的，本发明的技术方案是这样实现的：
一种自动驾驶车辆的横向决策确定方法，包括：根据道路特征信息以及预先选择的目标线和环境物体目标，评估所述自动驾驶车辆进行横向决策所需的目标车道和车道异常情况；以及结合所述道路特征信息，根据所评估的目标车道和车道异常情况，判断并输出所述自动驾驶车辆的预期横向行为，其中所述预期横向行为包括车道保持、换道和异常换道中的任意一者。
进一步的，所述评估所述自动驾驶车辆进行横向决策所需的目标车道和车道异常情况包括：根据所述道路特征信息选择所述自动驾驶车辆的目标车道，其中所述目标车道的选择原则包括遵循道路场景的原则、遵循车道属性的原则、不选择异常车道的原则以及车道异常时选择相邻车道和依次靠右选择的原则，其中所述道路特征信息包括道路类型、道路特征点及所述车道属性，且所述车道属性包括车道特征点属性和车道数属性；以及根据所述道路特征信息识别异常车道，并提供针对异常车道的避障策略。
进一步的，所述根据所述道路特征信息选择所述自动驾驶车辆的目标车道包括：在所述自动驾驶车辆行驶于主道常规场景时，根据所述选择原则来选择目标车道，其中所述主道常规场景包括加速车道、正常行驶车道和减速车道；在所述自动驾驶车辆行驶于主道特殊场景时，根据前方道路相对于当前道路的车道数属性的变化来选择目标车道，其中所述主道特殊场景包括主道变窄、主道变宽、主道分叉和/或隧道；以及在所述自动驾驶车辆行驶于匝道场景时，根据前方道路相对于当前道路的车道数属性的变化来选择目标车道，其中所述匝道场景包括常规匝道、匝道变窄、匝道变宽、匝道分叉和/或匝道交汇。
进一步的，所述根据所述道路特征信息识别异常车道，并提供针对异常车道的避障策略包括：分析道路特征信息以筛选出所述自动驾驶车辆的前方道路的静态障碍目标，并基于所述静态障碍目标识别是否车道异常；以及在车道异常时，引导所述自动驾驶车辆进行避障。
进一步的，所述引导所述自动驾驶车辆进行避障包括：根据所述静态障碍目标以及设定区域存在的动态的所述环境物体目标确定避 障目标，并确定所述避障目标相对于所述自动驾驶车辆的静态特性和动态特性；基于所述避障目标的静态特性和动态特性建立适应于道路特性的避障区域；基于所述避障目标的静态特性和动态特性，判断所述避障区域的可通行性；对所述自动驾驶车辆正常换道时的相关环境物体目标进行碰撞风险评估，并根据所述碰撞风险评估的结果确定换道可行性；以及根据所述换道可行性和所述避障区域的可通行性，控制所述自动驾驶车辆进行换道或在当前行驶车道绕开所述避障目标行驶。
进一步的，所述根据所述换道可行性和所述避障区域的可通行性，控制所述自动驾驶车辆进行换道或在当前行驶车道绕开所述避障目标行驶包括：若换道可行，则控制所述自动驾驶车辆进行换道，否则判断所述避障区域的可通行性，若所述避障区域可通行，则确定所述自动驾驶车辆在当前行驶车道绕开所述避障目标行驶。
所述自动驾驶车辆的横向决策确定方法与上述横向决策系统相对于现有技术所具有的优势相同，在此不再赘述。
本发明的另一目的在于提出一种机器可读存储介质，以实现对车辆横向行为的正确判断。
为达到上述目的，本发明的技术方案是这样实现的：
一种机器可读存储介质，该机器可读存储介质上存储有指令，该指令用于使得机器执行上述的自动驾驶车辆的横向决策确定方法。
本发明的其它特征和优点将在随后的具体实施方式部分予以详细说明。
附图说明
构成本发明的一部分的附图用来提供对本发明的进一步理解，本发明的示意性实施方式及其说明用于解释本发明，并不构成对本发明的不当限定。在附图中：
图1是本发明实施例在车体坐标系下对本车环境进行区域划分的示意图；
图2是本发明实施例的一种自动驾驶车辆的横向决策系统的结构示意图；
图3是本发明实施例中正常行驶车道的目标车道选择的示例图；
图4(a)-图4(c)分别是本发明实施例中的主道变窄、主道变宽和主道分叉的示意图；
图5是本发明实施例中车道异常判断的示例图；
图6是本发明实施例中本车道多静态障碍物车道异常识别的示例图；
图7是本发明实施例中车辆进行避障的示意图；
图8是本发明实施例的自动驾驶车辆的硬件布置示意图；以及
图9是本发明实施例的自动驾驶车辆的横向决策确定方法的流程示意图。
具体实施方式
需要说明的是，在不冲突的情况下，本发明中的实施方式及实施方式中的特征可以相互组合。
本发明实施例中提到的“环境物体目标”可以指在车辆前方、后方或侧方的移动的或静止的任意物体，例如，车辆、人、建筑物等，提到的“目标线”可以指自动驾驶车辆(以下简称车辆)横向决策、横向控制所需的车道中心线、动态目标线或安全偏移线等，车辆跟随目标线行驶，“目标车道”与“目标线”相对应，横向决策系统将作出车辆在目标车道行驶的决策。另外，本发明实施例中的“车道异常”主要是目标车道异常，表明是由于静态障碍物(如：路障、路锥、无法移动事故车辆等)或隧道入口红灯等引起的车道无法通行的情况。
另外，图1是本发明实施例在车体坐标系下对本车环境进行区域划分的示意图，包括车辆的前方区域、左前方区域等，下文均以图1的区域划分来说明环境物体目标等所处的位置。
下面将参考附图并结合实施方式来详细说明本发明。
图2是本发明实施例的一种自动驾驶车辆的横向决策系统的结构示意图。如图2所述，所述横向决策系统包括：评估单元100，用于根据道路特征信息以及预先选择的目标线和环境物体目标，评估车辆进行横向决策所需的目标车道和车道异常情况；以及判断单元200，用于结合所述道路特征信息，根据所述评估单元100所评估的目标车道和车道异常情况，判断并输出车辆的预期横向行为。
其中，所述道路特征信息包括道路类型、道路特征点及车道属性，而车道属性又包括车道特征点属性和车道数属性，这些信息将在下文中具体应用，在此不再赘述。另外，预先选择的目标线和环境物体目标是车辆的决策系统在进行纵向决策和横向决策之前，通过其他功能模块所得到的，本领域技术人员可参考相关技术理解，本发明实施例并不关注于此。
其中，所述预期横向行为包括车道保持、换道和异常换道中的任意一者。可理解的是，车道保持即车辆沿当前车道行驶；换道即车辆向左或右进入相邻车道行驶，在换道过程中需考虑周围环境车辆对自动驾驶车辆换道过程的可能影响；异常换道即本车道前方不满足车道保持、换道条件时，车辆进入异常换道(避障状态)。下文还将结合示例对这三种预期横向行为进行介绍，在此则不再赘述。
在优选的实施例中，所述评估单元100包括：目标车道管理模块 110，用于根据所述道路特征信息选择所述自动驾驶车辆的目标车道；以及车道异常管理模块120，用于根据所述道路特征信息识别异常车道，并提供针对异常车道的避障策略。
其中，对于目标车道管理模块110，目标车道的选择原则包括遵循道路场景的原则、遵循车道属性的原则、不选择异常车道的原则以及车道异常时选择相邻车道和依次靠右选择的原则。举例而言，遵循道路场景的原则是指选择目标车道是要考虑道路是主道还是匝道，遵循车道属性的原则是指选择目标车道时要考虑车道类型变化(通过车道特征点判断，例如驶入加速车道)和车道数变化，不选择异常车道的原则是指不能将异常车道作为目标车道，车道异常时选择相邻车道和依次靠右选择的原则是指在车道异常时，优先选择其相邻车道，若多条车道异常，则依次向右选择相邻车道。需说明的是，本发明实施例并不只局限于这几种选择原则，在目标车道的选择中，需要结合实际情况，考虑更多的因素，下文将通过示例说明这里的四种选择原则以及一些其他的选择原则。
在更为优选的实施例中，所述目标车道管理模块110包括：主道目标车道选择子模块111，用于在所述自动驾驶车辆行驶于主道常规场景时，根据所述选择原则来选择目标车道，还用于在所述自动驾驶车辆行驶于主道特殊场景时，根据前方道路相对于当前道路的车道数属性的变化来选择目标车道；以及匝道目标车道选择子模块112，用于在所述自动驾驶车辆行驶于匝道场景时，根据前方道路相对于当前道路的车道数属性的变化来选择目标车道。
下面具体介绍主道目标车道选择子模块111所关注的主道常规场景和主道特殊场景下的目标车道选择以及匝道目标车道选择子模块112所关注的匝道场景下的目标车道选择。
一、主道常规场景
其中，主道常规场景包括加速车道、正常行驶车道和减速车道，这三种车道属于车辆的正常行车道，可通过车道属性识别这三种车道。
进一步地，在驶出匝道汇入主道之前的部分称为加速车道(遵循车道属性)，当自动驾驶车辆驶入加速车道时，目标车道应选择最右侧车道，当目标车道异常时，应选择原目标车道相邻车道，并根据加速车道在道路右侧(遵循道路场景)的特点，尽量靠右选择目标车道。
进一步地，将正常行驶车道距离减速车道起点距离小于预警距离的这段道路及高速上规划的减速车道部分都称为减速车道(遵循车道属性)，当自动驾驶车辆驶入减速车道时，根据减速车道在道路最右侧(遵循道路场景)的特点，应将目标车道置换为最右侧车道，为进入减速车道及匝道部分提前做好准备。当目标车道异常时，应选择原目标车道相邻车道，并尽量靠右，便于车辆选取适当的时机尽快进入 减速车道及匝道部分驶离该路段。
下面详细介绍针对正常行驶车道的目标车道选择。在此的正常行驶车道的是指从车辆驶出加速车道汇入高速公路主道，到驶离进入减速车道的路段(遵循车道属性)，其中不包括上述的主道特殊场景。
图3是本发明实施例中正常行驶车道的目标车道选择的示例图，该图中本车位置、原目标车道位置及障碍物位置可适应于下述场景进行变化，在此不再一一示出各类变化，本领域技术人员可结合文字进行理解。结合图3，正常行驶车道目标车道选择的原则主要有以下几点：
1、两车道(例如只有C3、C4两条车道)：两条车道均正常，右侧车道为目标车道；只有一条车道正常(例如C3正常)，正常车道为目标车道。
2、三车道(例如只有C2、C3、C4三条车道)：三条车道均正常，中间车道为目标车道；中间车道不正常，右侧车道为目标车道；只有一条车道正常，正常车道为目标车道。
3、车道数大于三：左侧第二车道为目标车道，例如C1-C4都正常时，选择C2为目标车道。
4、目标车道异常时，本着逐步向右的原则选择目标车道，当异常消失时，回到原目标车道。如当前的图3所示，目标车道本应为C2，然而C2存在静态障碍物导致C2车道异常，无法通过，此时将目标车道置为C3车道，当车辆超过障碍物，且C2车道正常时，目标车道仍然变为C2车道。同理若C2车道、C3车道前方同样异常无法通过，则将目标车道置于C4车道，当车道数更多时，依次类推。这是因为当车道数较多时，目标车道选择靠近左侧，因为高速路况中，左侧车速更快(符合遵循道路场景的原则)，车辆能够在以一个较快的车速行驶，而当原目标车道出现异常时，目标车道选择右侧相邻车道，并依次向右选择，这有利于在前方道路出现异常时候更加快速的将自动驾驶车辆停在应急车道上或者驶离高速公路。
进一步地，除上述4种选择原则以外，还可对部分特殊场景目标车道选择进行修正，使其更符合人的驾驶习惯，例如以下选择原则：
1)目标车道异常，而当前车辆所在车道正常，则当前车道为目标车道。
2)本车道异常，选取最近的正常车道为目标车道。当左右侧相同时，选取右侧为目标车道。结合图3，原目标车道为C2，此时前方C2和C3异常，因此时车辆在C2，距离C1较近，故将目标车道置于C1。
需说明的是，上述目标车道选择不局限于车道数量，凡符合上述场景均可以采用相应原则。
二、主道特殊场景
其中，主道特殊场景主要包括主道变窄、主道变宽、主道分叉(分离式路基)和/或隧道。
1、主道变窄
图4(a)是本发明实施例中的主道变窄的示意图，其中主道变窄指的是自动驾驶车辆在主道上面行驶，前方车道数量减少，其包括左侧变窄、右侧变窄及两侧变窄三种情况。本发明实施例中，主道变窄情景下的目标车道选择原则为：提前1000m(标定量)变更车道属性(正常车道→变窄车道)；若原目标车道为道路变窄车道，置原目标车道相邻正常车道为目标车道。
其中，在此的主道变窄以及下面的主道变宽、主道分叉和隧道所对应的车道异常时的选择原则与主道常规场景近似，在此不再赘述。
2、主道变宽
图4(b)是本发明实施例中主道变宽的示意图，其中主道变宽指的是自动驾驶车辆在主道上面行驶，前方车道数量增多，其包括左侧变宽、右侧变宽及两侧变宽三种情况。
本发明实施例中，主道变宽情景下的目标车道选择原则为：提前500m(标定值)变更车道属性(正常车道→变宽车道)；车辆沿着当前目标车道行驶，直至进入变宽区域，车道数量发生变化，重新选择目标车道。参考图4(b),当前车道数为2，目标车道为最右侧车道，车辆沿当前道路行驶，本车所在车道数属性由2变为3，目标车道为C2。以原车道数为2为例，左侧变宽、右侧变宽后车道数变为3，双侧变宽后，车道数变为4，根据变更的车道数重新按上文提及的原则选择目标车道。
3、主道分叉
图4(c)是本发明实施例中主道分叉的示意图，其中主道分叉又称为分离式路基，道路指向两个不同的方向，一般伴随着车道数属性的变化。
本发明实施例中，主道分叉情景下的目标车道选择原则包括：提前500m(标定值)变更车道属性(正常车道→主道分叉)；以4车道目标方向为右(为左的情景与此类似)为例，当目标侧前方1条车道时，该车道为目标车道，当目标侧前方2条车道时，右侧车道为目标车道，当目标侧前方3条车道时，中间车道为目标车道。
参考图4(c)，分叉右侧前方有两条车道，依照上面原则，在过了分叉点车辆应走右侧车道，在车道数跳变前，将目标车道置最右侧C4车道，车辆车道保持进入分叉目标方向侧，即行驶于C2车道。
4、隧道
本发明实施例中，隧道目标车道的选择原则与上述的图3所对应 的正常行驶车道相同或相似，在此不再赘述。
三、匝道场景
其中，所述匝道场景包括常规匝道、匝道变窄、匝道变宽、匝道分叉和/或匝道交汇。需说明的是，针对匝道变窄、匝道变宽、匝道分叉的目标车道选择与上述图4(a)-图4(c)对应的主道变窄、主道变宽、主道分叉的目标车道选择相同或相似，区别主要在于主道变为匝道，本领域技术人员是可以结合匝道路况进行理解，故在此不再赘述。
对于常规匝道，当车辆驶入匝道时，目标车道选择最右侧车道。当最右侧车道异常时，目标车道选择临近最右侧车道，目标车道选择原则遵从尽量靠右原则。
对于匝道交汇，或称匝道合并，是指不同方向的匝道合并成为一个。此情景下，车辆行驶在匝道上，提前500m(标定值)变更车道属性(普通匝道→交汇匝道)，匝道车道数属性发生变化，车辆沿着当前目标车道行驶汇入交汇匝道，当车道数属性变化后，按照新的车道数，重新选择目标车道。
本发明实施例中，目标车道管理模块110根据法律法规对高速公路上不同车道的不同速度限制，将主道及匝道的目标车道选择为使车辆按照预先设定的方向以一个较快的速度行进的优先行驶车道，并且由于规划了优先行驶车道，避免由于地图定位不准，造成车辆行驶出现较大横向偏差而导致的碰撞危险，保证了车辆在安全的前提下可以以一个较快的速度行驶，并且其目标车道选择规划方式符合人的驾驶习惯。
继续参考图2，在优选的实施例中，所述车道异常管理模块120包括：车道异常识别子模块121，用于分析道路特征信息以筛选出所述自动驾驶车辆的前方道路的静态障碍目标，并基于所述静态障碍目标识别是否车道异常；以及避障子模块122，用于在车道异常时引导所述自动驾驶车辆进行避障。
其中，车道异常识别子模块121还用于向目标车道管理模块110提供识别到的车道异常的信息，以使目标车道管理模块110结合车道异常情况来选择目标车道。
进一步地，结合上文，当车辆行驶在非隧道入口路段时，车道异常识别子模块121应包括静态障碍目标选取、车道异常判断以及本车道多静态障碍物车道异常识别三个部分。
1、静态障碍目标选取
本发明实施例中，静态障碍目标选取原则包括：提取车辆当前行驶路段的道路特征信息(车道数、各车道宽度等)、道路附属物信息、环境物体目标信息。在各车道内，以距离车辆最近的环境物体目标为 参考，筛选一定范围内各车道静态障碍目标(也称为静态障碍物)，静态障碍目标主要是路锥、路障、故障车辆等静态物体目标，也包含目标速度小于某一阈值时的动态目标。并且，可按车道提取各静态障碍目标相对于本车的横纵向距离信息等。
2、车道异常判断原则
图5是本发明实施例中车道异常判断的示例图，其以车辆所在的本车道为例，而其他车道的异常判断原则与此类似。如图5所示，由于静态目标1与静态目标2的作用，在D2范围内，自动驾驶车辆可行驶区域如ABCE所示，可行驶宽度D，其为D2范围内静态目标1与本车道中心线横向距离最近的点距离值l 1(正值左)，与静态目标2与本车道中心线横向距离最近的点距离值l 2(负值右)绝对值的和(l 1为本车道中心线横向距离-E点横向距离；l 2为本车道中心线横向距离-C点横向距离)。如果没有横向最近点，即前方没有静态障碍目标，则l 1和l 2取固定值(选确定的标定量，可称为TBD)。当车道的可行驶宽度D小于TBD时，认为车道异常，车辆无法通行，若可行驶宽度D大于TBD时，认为道路正常，车辆可以正常通过。
当车辆行驶于隧道口时，还需要识别隧道入口各车道红绿灯情况，当车道为红灯时，则把该车道置为异常车道(从入口到出口均为异常状态)；直到自动驾驶车辆驶出隧道，系统重新识别道路状态是否为隧道，并重新识别红绿灯。
3、本车道多静态障碍物车道异常识别
图6是本发明实施例中本车道多静态障碍物车道异常识别的示例图。参考图6，车道存在静态目标1、2其中两者之间距离D3，可行驶宽度D＝(abs(l 1)+abs(l 2))大于设定阈值TBD(优先取值2.8m)，车辆可安全通过静态目标1；当D3大于设定阈值TBD1(TBD1的取值与当前车速线性相关K*V，最小取值25m，其中K为比例系数，V为车速)时，可行驶宽度D4＝(abs(l 1’)+abs(l 2’))大于设定阈值TBD(优先取值2.8m)，车辆可安全通过静态目标2。因此，虽车道异常，但车辆仍可安全通过。
进一步地，对于避障子模块122，其用于在车道异常时引导所述自动驾驶车辆进行避障，主要可包括以下步骤：根据所述静态障碍目标以及设定区域存在的动态的所述环境物体目标确定避障目标，并确定所述避障目标相对于所述自动驾驶车辆的静态特性和动态特性；基于所述避障目标的静态特性和动态特性建立适应于道路特性的避障区域；基于所述避障目标的静态特性和动态特性，判断所述避障区域的可通行性；对所述自动驾驶车辆正常换道时的相关环境物体目标进行碰撞风险评估，并根据所述碰撞风险评估的结果确定换道可行性；以及根据所述换道可行性和所述避障区域的可通行性，控制所述自动 驾驶车辆进行换道或在当前行驶车道绕开所述避障目标行驶。
举例而言，避障子模块122所实现的功能主要包括以下几个部分。
1、避障目标选取
避障目标包括静态障碍物和动态障碍物。选取原则以区域内距离自动驾驶车辆最近物体目标为参考。静态障碍物主要是路锥、路障、故障车辆等静态物体目标，避障目标包括：①正前方区域静态物体目标；②左前方区域静态物体目标；③右前方区域静态物体目标；④左侧方区域静态物体目标；⑤右侧方区域的静态物体目标。动态障碍物主要是运动的物体目标，避障目标包括：①正前方区域低于自动驾驶车辆速度的动态物体目标；②左前方区域低于自动驾驶车辆速度的动态物体目标；③右前方区域低于自动驾驶车辆速度的动态物体目标；④左侧方区域的动态物体目标；⑤右侧方区域的动态物体目标；⑥左后方区域高于自动驾驶车辆速度的动态物体目标；⑦右后方区域高于自动驾驶车辆速度的动态物体目标。
2、避障区域建立
传统避障区域建立方法通常建立扇形区域，以扇形角度的1/2作为偏转姿态，躲避障碍物成功后再进行跟随，这种方法适用于无车道线的城市/乡村道路等低速自动驾驶车辆。本发明实施例中，避障区域的建立除了考虑避障目标外，还应考虑道路特性，使自动驾驶的避障行为符合高速公路对于驾驶员的行为要求(如在本车道内行驶，除超车外不压线；不画龙行驶；速度不忽快忽慢等)。
图7是本发明实施例中车辆进行避障的示意图，其中ABCD构成的区域这里的避障区域，弧AC和弧BD的弧长等于200米，曲率等于车道线L2的曲率，即AC与BD平行于道路，该区域的尺寸大小通过避障目标G1和G2确定。
目标G1为正前方区域的动态物体目标，G1与本车之间的关系包括外轮廓点，即横向最近点G11和纵向最近点G12，经G11构建平行于道路的曲线s1，纵向最近点G12到曲线s1垂线的交点为G13，将G13作为目标G1进行避障的外轮廓点，考虑避障安全加入d2＝0.3米的安全距离生成BD曲线。
目标G2为左前方区域的静态物体目标(修路路障)，G2最终选取的外轮廓点为G21，考虑避障安全加入d1＝0.1米的安全距离生成AC曲线。
3、避障区域可通行性判断
根据第2部分生成避障区域ABCD后，车辆是否能够通过该区域则需要进行判断，车辆宽度W+安全距离d3最为可通行性判断的条件，当避障区域宽度大于(W+d3)时，自动驾驶车辆可以进行避障；否则自动驾驶车辆重新判断其它避障区域(如右侧是否可以生成 避障区域)。
在优选的实施例中，所述避障子模块122用于根据所述换道可行性和所述避障区域的可通行性，控制所述自动驾驶车辆进行换道或在当前行驶车道绕开所述避障目标行驶包括：若换道可行，则控制所述自动驾驶车辆进行换道，否则判断所述避障区域的可通行性，若所述避障区域可通行，则确定所述自动驾驶车辆在当前行驶车道绕开所述避障目标行驶。具体地，结合实际，所述避障子模块122用于识别是否需要换道包括以下的换道意图产生(即碰撞风险评估)、换道方向判断和换道可行性判断三个方面，
1、换道意图产生
车辆正常行驶时，探测区域前方出现低于车辆最高限速非静态物体目标时，自动驾驶车辆根据本车与前车相对距离、速度权衡确定车辆是否需要换道，降低自动驾驶车辆换道频率。
假设换道意图期望因子阈值设定为η，自动驾驶车辆速度V_auto,目标车辆速度V_trg,自动驾驶车辆与目标车辆之间相对距离Dis_rely,自动驾驶车辆期望安全行车距离K*V_auto其中K优先取0.8。
换道意图期望因子β＝K1*(V_auto/V_trg)+K2*(Dis_rely/K*V_auto)，其中K1+K2＝1当换道意图期望因子β小于η时自动驾驶车辆意图满足。
自动驾驶车辆正常行驶时，探测区域前方出现静态物体目标时，自动驾驶车辆应提前换道避免与前方静态物体发生碰撞。
假设换道意图期望因子阈值设定为ηs，自动驾驶车辆速度V_auto,自动驾驶车辆与静态障碍物之间相对距离Dis_s，自动驾驶车辆期望安全行车距离K*V_auto，其中K优先取1。
换道意图期望因子βs＝K1*(Dis_s/K*V_auto)，其中K1优先取值1，当换道意图期望因子βs小于ηs时自动驾驶车辆意图满足。
2、换道方向判断
自动驾驶车辆换道方向判断需满足以下条件：
a)前方区域(左前方、正前方、右前方)存在非静态物体目标。
b)左前或右前方区域物体目标速度与本车道前车速度之差大于速度阈值ΔV，ΔV优先取值5km/h。
c)车辆与左前或右前车之间距离大于车辆期望安全行车距离K3*V_auto，其中K3优先取0.6。
d)车辆正侧方无车。
e)根据后方区域(左后方、右后方)环境车辆与自动驾驶车辆之间关系进行碰撞风险评估，根据自动驾驶车辆与环境物体目标TTC值(TTC即自动驾驶车辆与前方车辆碰撞时间，TTC＝相对速度/相对距离；相对速度＝本车速度-前车速度)确定自动驾驶车辆换道可行性， 优先推荐选取TTC值大于2。
f)后方区域(左后方、右后方)环境车辆与自动驾驶车辆之间相对距离大于自动驾驶车辆期望安全行车距离K4*V_auto，其中K4优先取0.3。
g)换到条件判断中自动驾驶车辆左换道优先，即当左前、右前区域同时满足a～f条件时，优先选择左侧车道为目标车道。
自动驾驶车辆根据上述条件a)～g)确定换道的目标车道。
3、换道可行性判断
车辆须谨遵道路交通法规，如：虚实线、限速、灯光、喇叭、红绿灯、禁止掉头等。
本发明实施例的避障子模块122提出一种适用于车辆高速行驶、结构化道路的避障方法，可避免人工驾驶可能因盲区引起车辆碰撞，且其换道功能能够提高车辆行驶效率降低驾驶员工作量，且涉及的自动换道方法适用范围广，可以适用于较大曲率弯曲道路及笔直道路下自动驾驶系统尤其是结构化道路下的自动驾驶系统。
在此，本发明实施例的车道异常管理模块120，能识别车道情况，并可以提前主动引导车辆避障或者逐渐向应急车道靠拢或者驶离高速公路，避免车辆发生碰撞危险。
需说明的是，车辆的决策系统与环境感知系统及它们各自的功能模块可对应理解为车辆上的控制单元，下面将基于这一理解来说明本发明实施例的自动驾驶车辆的硬件布置。图8是本发明实施例的自动驾驶车辆的硬件布置示意图，其中所述自动驾驶车辆的决策系统中包括有上述实施例的横向决策系统。
如图8所示，控制单元1、控制单元2、控制单元4构成环境感知系统，控制单元3构成本发明实施例的横向决策系统，其是车辆的决策系统的部分。控制单元1为自动驾驶车辆提供准确位置信息，优先选用高精度GPS+IMU设备，横向定位偏差10cm以内，纵向定位偏差30cm以内。控制单元2用于存储、输出自动驾驶车辆前后方200m范围内高精度车道线、车道数、车道宽度等信息，优先使用存储空间大于50G，处理内存大于1G硬件设备。控制单元4用于探测提取自动驾驶车辆周围360°范围内出现的物体目标，优先选用全天候传感器探测设备避免因雨、雪、雾、光照等引起物体目标误检、漏检等。其中，控制单元4不仅仅局限于当前安装位置也不局限于当前数量，为提高物体探测准确性在车身周围布置若干雷达传感器(激光雷达或毫米波雷达设备等)、视觉传感器，通过设备冗余提高物体目标检测准确、稳定性。
其中，控制单元2获取控制单元1提供自动驾驶车辆准确位置信息，处理运算后实时输出自动驾驶车辆前后方200m范围内高精度地 图数据，包括：车道线离散点经纬度(经纬度以地心为原点)、离散点航向角(以正北方向为0°顺时针为证)、车道线类型、车道宽度、车道数量、道路边界等信息，控制单元3将通过以太网方式接收到车道线离线数据通过坐标转换至平面车辆坐标系下，提供车辆换道过程中所需的道路特征信息，控制单元4同时将探测区域内物体目标信息以CAN通讯方式传输至控制单元3，控制单元3执行上述的横向决策系统的功能。
由此可知，本发明实施例的横向决策系统是易于通过硬件来实现的。
综上所述，本发明实施例的横向决策系统能够评估出目标车道及车道异常情况，并据此作出符合道路特性的车道保持、换道或异常换道的横向决策，以便于车辆的控制系统可基于该横向决策进行适应性的横向控制，以保证车辆的行驶安全。
图9是本发明实施例的一种自动驾驶车辆的横向决策确定方法的流程示意图，该横向决策确定方法与上述的横向决策系统是基于同样的发明思路的。如图9所示，所述自动驾驶车辆的横向决策确定方法可以包括以下步骤S100及步骤S200：
步骤S100，根据道路特征信息以及预先选择的目标线和环境物体目标，评估所述自动驾驶车辆进行横向决策所需的目标车道和车道异常情况。
优选地，该步骤S100又包括以下子步骤：
步骤S110，根据所述道路特征信息选择所述自动驾驶车辆的目标车道。
其中，所述目标车道的选择原则包括遵循道路场景的原则、遵循车道属性的原则、不选择异常车道的原则以及车道异常时选择相邻车道和依次靠右选择的原则。
更为优选地，该步骤S110具体包括：在所述自动驾驶车辆行驶于主道常规场景时，根据所述选择原则来选择目标车道，其中所述主道常规场景包括加速车道、正常行驶车道和减速车道；在所述自动驾驶车辆行驶于主道特殊场景时，根据前方道路相对于当前道路的车道数属性的变化来选择目标车道，其中所述主道特殊场景包括主道变窄、主道变宽、主道分叉和/或隧道；以及在所述自动驾驶车辆行驶于匝道场景时，根据前方道路相对于当前道路的车道数属性的变化来选择目标车道，其中所述匝道场景包括常规匝道、匝道变窄、匝道变宽、匝道分叉和/或匝道交汇。
步骤S120，根据所述道路特征信息识别异常车道，并提供针对异常车道的避障策略。
更为优选地，该步骤S120又具体包括：分析道路特征信息以筛 选出所述自动驾驶车辆的前方道路的静态障碍目标，并基于所述静态障碍目标识别是否车道异常；以及在车道异常时，引导所述自动驾驶车辆进行避障。
进一步地，所述引导所述自动驾驶车辆进行避障包括：根据所述静态障碍目标以及设定区域存在的动态的所述环境物体目标确定避障目标，并确定所述避障目标相对于所述自动驾驶车辆的静态特性和动态特性；基于所述避障目标的静态特性和动态特性建立适应于道路特性的避障区域；基于所述避障目标的静态特性和动态特性，判断所述避障区域的可通行性；对所述自动驾驶车辆正常换道时的相关环境物体目标进行碰撞风险评估，并根据所述碰撞风险评估的结果确定换道可行性；以及根据所述换道可行性和所述避障区域的可通行性，控制所述自动驾驶车辆进行换道或在当前行驶车道绕开所述避障目标行驶。
更进一步地，所述根据所述换道可行性和所述避障区域的可通行性，控制所述自动驾驶车辆进行换道或在当前行驶车道绕开所述避障目标行驶包括：若换道可行，则控制所述自动驾驶车辆进行换道，否则判断所述避障区域的可通行性，若所述避障区域可通行，则确定所述自动驾驶车辆在当前行驶车道绕开所述避障目标行驶。
步骤S200，结合所述道路特征信息，根据所评估的目标车道和车道异常情况，判断并输出所述自动驾驶车辆的预期横向行为。
需说明的是，本发明实施例的自动驾驶车辆的横向决策确定方法与上述自动驾驶车辆的横向决策系统的实施例的具体实施细节及效果相同，在此则不再赘述。
本发明另一实施例还提供一种机器可读存储介质，该机器可读存储介质上存储有指令，该指令用于使得机器执行上述的横向决策确定方法。所述机器可读存储介质包括但不限于相变内存(相变随机存取存储器的简称，Phase Change Random Access Memory，PRAM，亦称为RCM/PCRAM)、静态随机存取存储器(SRAM)、动态随机存取存储器(DRAM)、其他类型的随机存取存储器(RAM)、只读存储器(ROM)、电可擦除可编程只读存储器(EEPROM)、快闪记忆体(Flash Memory)或其他内存技术、只读光盘只读存储器(CD-ROM)、数字多功能光盘(DVD)或其他光学存储、磁盒式磁带，磁带磁磁盘存储或其他磁性存储设备等各种可以存储程序代码的介质。
以上所述仅为本发明的较佳实施方式而已，并不用以限制本发明，凡在本发明的精神和原则之内，所作的任何修改、等同替换、改进等，均应包含在本发明的保护范围之内

Claims (13)

  1. 一种自动驾驶车辆的横向决策系统，其特征在于，所述自动驾驶车辆的横向决策系统包括：
    评估单元，用于根据道路特征信息以及预先选择的目标线和环境物体目标，评估所述自动驾驶车辆进行横向决策所需的目标车道和车道异常情况；以及
    判断单元，用于结合所述道路特征信息，根据所述评估单元所评估的目标车道和车道异常情况，判断并输出所述自动驾驶车辆的预期横向行为，其中所述预期横向行为包括车道保持、换道和异常换道中的任意一者。
  2. 根据权利要求1所述的自动驾驶车辆的横向决策系统，其特征在于，所述评估单元包括：
    目标车道管理模块，用于根据所述道路特征信息选择所述自动驾驶车辆的目标车道，其中所述目标车道的选择原则包括遵循道路场景的原则、遵循车道属性的原则、不选择异常车道的原则以及车道异常时选择相邻车道和依次靠右选择的原则，其中所述道路特征信息包括道路类型、道路特征点及所述车道属性，且所述车道属性包括车道特征点属性和车道数属性；以及
    车道异常管理模块，用于根据所述道路特征信息识别异常车道，并提供针对异常车道的避障策略。
  3. 根据权利要求2所述的自动驾驶车辆的横向决策系统，其特征在于，所述目标车道管理模块包括：
    主道目标车道选择子模块，用于在所述自动驾驶车辆行驶于主道常规场景时，根据所述选择原则来选择目标车道，其中所述主道常规场景包括加速车道、正常行驶车道和减速车道；以及用于在所述自动驾驶车辆行驶于主道特殊场景时，根据前方道路相对于当前道路的车道数属性的变化来选择目标车道，其中所述主道特殊场景包括主道变窄、主道变宽、主道分叉和/或隧道；以及
    匝道目标车道选择子模块，用于在所述自动驾驶车辆行驶于匝道场景时，根据前方道路相对于当前道路的车道数属性的变化来选择目标车道，其中所述匝道场景包括常规匝道、匝道变窄、匝道变宽、匝道分叉和/或匝道交汇。
  4. 根据权利要求2所述的自动驾驶车辆的横向决策系统，其特征在于，所述车道异常管理模块包括：
    车道异常识别子模块，用于分析道路特征信息以筛选出所述自动驾驶车辆的前方道路的静态障碍目标，并基于所述静态障碍目标识别是否车道异常；以及
    避障子模块，用于在车道异常时引导所述自动驾驶车辆进行避障。
  5. 根据权利要求4所述的自动驾驶车辆的横向决策系统，其特征在于，所述避障子模块用于在车道异常时引导所述自动驾驶车辆进行避障包括：
    根据所述静态障碍目标以及设定区域存在的动态的所述环境物体目标确定避障目标，并确定所述避障目标相对于所述自动驾驶车辆的静态特性和动态特性；
    基于所述避障目标的静态特性和动态特性建立适应于道路特性的避障区域；
    基于所述避障目标的静态特性和动态特性，判断所述避障区域的可通行性；
    对所述自动驾驶车辆正常换道时的相关环境物体目标进行碰撞风险评估，并根据所述碰撞风险评估的结果确定换道可行性；以及
    根据所述换道可行性和所述避障区域的可通行性，控制所述自动驾驶车辆进行换道或在当前行驶车道绕开所述避障目标行驶。
  6. 根据权利要求5所述的自动驾驶车辆的横向决策系统，其特征在于，所述避障子模块用于根据所述换道可行性和所述避障区域的可通行性，控制所述自动驾驶车辆进行换道或在当前行驶车道绕开所述避障目标行驶包括：
    若换道可行，则控制所述自动驾驶车辆进行换道，否则判断所述避障区域的可通行性，若所述避障区域可通行，则确定所述自动驾驶车辆在当前行驶车道绕开所述避障目标行驶。
  7. 一种自动驾驶车辆的横向决策确定方法，其特征在于，所述自动驾驶车辆的横向决策确定方法包括：
    根据道路特征信息以及预先选择的目标线和环境物体目标，评估所述自动驾驶车辆进行横向决策所需的目标车道和车道异常情况；以及
    结合所述道路特征信息，根据所评估的目标车道和车道异常情况，判断并输出所述自动驾驶车辆的预期横向行为，其中所述预期横向行为包括车道保持、换道和异常换道中的任意一者。
  8. 根据权利要求7所述的自动驾驶车辆的横向决策确定方法， 其特征在于，所述评估所述自动驾驶车辆进行横向决策所需的目标车道和车道异常情况包括：
    根据所述道路特征信息选择所述自动驾驶车辆的目标车道，其中所述目标车道的选择原则包括遵循道路场景的原则、遵循车道属性的原则、不选择异常车道的原则以及车道异常时选择相邻车道和依次靠右选择的原则，其中所述道路特征信息包括道路类型、道路特征点及所述车道属性，且所述车道属性包括车道特征点属性和车道数属性；以及
    根据所述道路特征信息识别异常车道，并提供针对异常车道的避障策略。
  9. 根据权利要求8所述的自动驾驶车辆的横向决策确定方法，其特征在于，所述根据所述道路特征信息选择所述自动驾驶车辆的目标车道包括：
    在所述自动驾驶车辆行驶于主道常规场景时，根据所述选择原则来选择目标车道，其中所述主道常规场景包括加速车道、正常行驶车道和减速车道；
    在所述自动驾驶车辆行驶于主道特殊场景时，根据前方道路相对于当前道路的车道数属性的变化来选择目标车道，其中所述主道特殊场景包括主道变窄、主道变宽、主道分叉和/或隧道；以及
    在所述自动驾驶车辆行驶于匝道场景时，根据前方道路相对于当前道路的车道数属性的变化来选择目标车道，其中所述匝道场景包括常规匝道、匝道变窄、匝道变宽、匝道分叉和/或匝道交汇。
  10. 根据权利要求8所述的自动驾驶车辆的横向决策确定方法，其特征在于，所述根据所述道路特征信息识别异常车道，并提供针对异常车道的避障策略包括：
    分析道路特征信息以筛选出所述自动驾驶车辆的前方道路的静态障碍目标，并基于所述静态障碍目标识别是否车道异常；以及
    在车道异常时，引导所述自动驾驶车辆进行避障。
  11. 根据权利要求10所述的自动驾驶车辆的横向决策确定方法，其特征在于，所述引导所述自动驾驶车辆进行避障包括：
    根据所述静态障碍目标以及设定区域存在的动态的所述环境物体目标确定避障目标，并确定所述避障目标相对于所述自动驾驶车辆的静态特性和动态特性；
    基于所述避障目标的静态特性和动态特性建立适应于道路特性的避障区域；
    基于所述避障目标的静态特性和动态特性，判断所述避障区域的可通行性；
    对所述自动驾驶车辆正常换道时的相关环境物体目标进行碰撞风险评估，并根据所述碰撞风险评估的结果确定换道可行性；以及
    根据所述换道可行性和所述避障区域的可通行性，控制所述自动驾驶车辆进行换道或在当前行驶车道绕开所述避障目标行驶。
  12. 根据权利要求11所述的自动驾驶车辆的横向决策确定方法，其特征在于，所述根据所述换道可行性和所述避障区域的可通行性，控制所述自动驾驶车辆进行换道或在当前行驶车道绕开所述避障目标行驶包括：
    若换道可行，则控制所述自动驾驶车辆进行换道，否则判断所述避障区域的可通行性，若所述避障区域可通行，则确定所述自动驾驶车辆在当前行驶车道绕开所述避障目标行驶。
  13. 一种机器可读存储介质，该机器可读存储介质上存储有指令，该指令用于使得机器执行权利要求7至12中任意一项所述的自动驾驶车辆的横向决策确定方法。
PCT/CN2019/129284 2018-12-29 2019-12-27 自动驾驶车辆的横向决策系统及横向决策确定方法 Ceased WO2020135742A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201811642027.5 2018-12-29
CN201811642027.5A CN110667578B (zh) 2018-12-29 2018-12-29 自动驾驶车辆的横向决策系统及横向决策确定方法

Publications (1)

Publication Number Publication Date
WO2020135742A1 true WO2020135742A1 (zh) 2020-07-02

Family

ID=69065715

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/129284 Ceased WO2020135742A1 (zh) 2018-12-29 2019-12-27 自动驾驶车辆的横向决策系统及横向决策确定方法

Country Status (2)

Country Link
CN (1) CN110667578B (zh)
WO (1) WO2020135742A1 (zh)

Cited By (26)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112650224A (zh) * 2020-12-11 2021-04-13 国汽(北京)智能网联汽车研究院有限公司 自动驾驶仿真的方法、装置、设备及存储介质
CN112710317A (zh) * 2020-12-14 2021-04-27 北京四维图新科技股份有限公司 自动驾驶地图的生成方法、自动驾驶方法及相关产品
CN112735187A (zh) * 2021-01-29 2021-04-30 重庆长安汽车股份有限公司 一种自动识别应急车道的系统和方法
CN113022570A (zh) * 2021-03-19 2021-06-25 武汉理工大学 一种车辆换道行为识别方法及设备
CN113085853A (zh) * 2021-04-26 2021-07-09 中汽研（天津）汽车工程研究院有限公司 一种车道内主动躲闪大型车辆的辅助驾驶系统
CN113104038A (zh) * 2021-03-31 2021-07-13 江铃汽车股份有限公司 车辆换道控制方法、装置、电子设备及可读存储介质
CN113823118A (zh) * 2021-02-19 2021-12-21 石家庄铁道大学 一种联合紧急程度和博弈论的智能网联车辆换道方法
CN113848913A (zh) * 2021-09-28 2021-12-28 北京三快在线科技有限公司 一种无人驾驶设备的控制方法及控制装置
CN114360246A (zh) * 2021-12-28 2022-04-15 北京汇通天下物联科技有限公司 一种高速公路出口匝道的预警方法、装置及存储介质
CN114407929A (zh) * 2022-01-29 2022-04-29 上海木蚁机器人科技有限公司 无人驾驶绕障处理方法、装置、电子设备及存储介质
CN114435403A (zh) * 2022-02-22 2022-05-06 重庆长安汽车股份有限公司 一种基于环境信息的导航定位校核系统及方法
CN114506324A (zh) * 2020-10-23 2022-05-17 上海汽车集团股份有限公司 一种车道决策方法及相关装置
CN114506347A (zh) * 2022-03-24 2022-05-17 重庆长安汽车股份有限公司 一种自动驾驶通过施工路段的系统和方法
CN114802244A (zh) * 2021-01-11 2022-07-29 广东科学技术职业学院 控制无人驾驶车辆的方法
CN114822078A (zh) * 2021-01-28 2022-07-29 武汉智行者科技有限公司 一种目标参考线切换控制方法、装置及存储介质
CN114999152A (zh) * 2022-05-25 2022-09-02 清华大学 面向混合车流的匝道合流边缘云管控方法
CN115158308A (zh) * 2021-04-02 2022-10-11 清华大学 一种智能车辆主动避障控制方法及装置、存储介质、终端
CN115169908A (zh) * 2022-07-14 2022-10-11 吉林大学 网联环境下基于知觉风险场的车辆行驶状态安全评估方法
CN115195748A (zh) * 2022-08-15 2022-10-18 中汽研（天津）汽车工程研究院有限公司 基于数据驱动的智能汽车拟人化决策规划系统及方法
CN115384502A (zh) * 2022-08-30 2022-11-25 吉林大学 一种基于驾驶风格的自动驾驶车辆车道选择方法
CN115402354A (zh) * 2022-09-26 2022-11-29 苏州挚途科技有限公司 一种针对匝道汇入的车辆控制方法、装置及设备
CN116142193A (zh) * 2023-02-28 2023-05-23 驭势科技（北京）有限公司 车辆换道决策方法、装置、电子设备和存储介质
CN116229715A (zh) * 2023-02-13 2023-06-06 武汉理工大学 一种道路交织区连续流生成方法与系统
CN117079478A (zh) * 2023-08-16 2023-11-17 北京中交华安科技有限公司 一种基于边缘计算的合流区交通冲突辨识装置
CN118430306A (zh) * 2024-04-22 2024-08-02 福建理工大学 一种自动驾驶车辆智能选择跟驰车辆的方法及终端
CN118605542A (zh) * 2024-08-07 2024-09-06 成都赛力斯科技有限公司 城区驾驶场景中的车辆绕障方法、电子设备及存储介质

Families Citing this family (28)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP6861669B2 (ja) * 2018-06-15 2021-04-21 本田技研工業株式会社 車両制御装置、車両制御方法、およびプログラム
CN115087573B (zh) * 2020-03-31 2023-05-05 华为技术有限公司 一种关键目标选取方法、装置及系统
CN111489569B (zh) * 2020-05-28 2021-08-31 兰州理工大学 一种立交匝道车道宽度动态调整系统及其调整方法
CN113741412B (zh) * 2020-05-29 2023-09-01 杭州海康威视数字技术股份有限公司 自动驾驶设备的控制方法、装置及存储介质
CN111746542B (zh) * 2020-06-04 2023-04-14 重庆长安汽车股份有限公司 车辆智能换道提醒的方法、系统、车辆及存储介质
CN111862604B (zh) * 2020-07-20 2022-03-04 北京京东乾石科技有限公司 无人车的控制方法及装置、计算机存储介质、电子设备
CN112068559B (zh) * 2020-08-28 2022-10-11 重庆长安汽车股份有限公司 无人驾驶车辆的偏移控制方法、系统、车辆及存储介质
CN111994076A (zh) * 2020-09-02 2020-11-27 中国第一汽车股份有限公司 自动驾驶车辆的控制方法及装置
CN112435466B (zh) * 2020-10-23 2022-03-22 江苏大学 混合交通流环境下cacc车辆退变为传统车的接管时间预测方法及系统
CN114446041B (zh) * 2020-10-30 2023-03-03 华为终端有限公司 一种车辆变道管理方法及变道管理装置
CN114475602B (zh) * 2020-11-12 2023-05-09 宇通客车股份有限公司 一种车辆、车辆调头方法及装置
CN114550474B (zh) * 2020-11-24 2023-03-03 华为技术有限公司 一种横向规划约束确定方法及装置
WO2022126349A1 (zh) * 2020-12-15 2022-06-23 华为技术有限公司 控制方法和控制装置
CN113096390B (zh) * 2021-03-25 2022-04-08 南京航空航天大学 一种车辆驶离主道的控制方法
CN115214708B (zh) * 2021-04-19 2026-04-03 华为技术有限公司 一种车辆意图预测方法及其相关装置
CN113320545A (zh) * 2021-07-01 2021-08-31 江苏理工学院 一种基于线控智能车辆的预路口行为决策方法
CN113548049B (zh) * 2021-07-27 2022-05-31 武汉理工大学 一种基于有限状态机的智能车辆驾驶行为决策方法及系统
CN113619602B (zh) * 2021-08-20 2023-03-10 华为技术有限公司 引导车辆行驶的方法及相关系统、存储介质
CN113945221B (zh) * 2021-09-26 2024-02-13 华中科技大学 一种考虑近迫感效应的自动驾驶车道宽度确定方法
CN114275039B (zh) * 2021-12-27 2022-11-04 联创汽车电子有限公司 智能驾驶车辆横向控制方法和模块
CN116353589A (zh) * 2021-12-27 2023-06-30 比亚迪股份有限公司 基于无人驾驶系统的车辆控制方法及装置和车辆
CN115320553B (zh) * 2022-08-30 2024-07-23 桂林电子科技大学 一种面向弯道情况下aeb系统的前车定位判断方法
CN115447616B (zh) * 2022-10-26 2024-05-17 重庆长安汽车股份有限公司 车辆驾驶客观指标的生成方法及装置
CN115909784B (zh) * 2022-12-07 2023-10-27 长安大学 多车道的智能网联车辆合流控制方法和控制装置
CN116300583B (zh) * 2023-02-01 2026-02-10 国汽智控(北京)科技有限公司 车辆控制方法、装置及车辆
CN116564097B (zh) * 2023-07-11 2023-10-03 蘑菇车联信息科技有限公司 车辆的路口通行决策方法、装置及系统、电子设备
CN118270039A (zh) * 2023-11-21 2024-07-02 比亚迪股份有限公司 车辆及其控制方法
CN118323141B (zh) * 2024-06-12 2024-10-11 知行汽车科技(苏州)股份有限公司 一种控制车辆行驶的方法、装置、设备及介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105980811A (zh) * 2014-02-25 2016-09-28 爱信艾达株式会社 路径搜索系统、路径搜索方法、计算机程序、以及成本表的数据结构
CN107408343A (zh) * 2015-03-31 2017-11-28 爱信艾达株式会社 自动驾驶辅助系统、自动驾驶辅助方法以及计算机程序
CN107430807A (zh) * 2015-03-31 2017-12-01 爱信艾达株式会社 自动驾驶辅助系统、自动驾驶辅助方法以及计算机程序
CN107533805A (zh) * 2015-04-27 2018-01-02 爱信艾达株式会社 自动驾驶辅助系统、自动驾驶辅助方法以及计算机程序

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3400419B1 (en) * 2016-01-05 2025-08-27 Mobileye Vision Technologies Ltd. Trained navigational system with imposed constraints
EP3483022B1 (en) * 2016-07-05 2020-11-04 Nissan Motor Co., Ltd. Travel control method and travel control device
JP6583185B2 (ja) * 2016-08-10 2019-10-02 トヨタ自動車株式会社 自動運転システム及び自動運転車両
CN106740835B (zh) * 2016-11-21 2019-10-22 北汽福田汽车股份有限公司 车辆自适应巡航控制方法、装置和车辆
CN106681319A (zh) * 2016-12-09 2017-05-17 重庆长安汽车股份有限公司 自动换道系统及方法

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105980811A (zh) * 2014-02-25 2016-09-28 爱信艾达株式会社 路径搜索系统、路径搜索方法、计算机程序、以及成本表的数据结构
CN107408343A (zh) * 2015-03-31 2017-11-28 爱信艾达株式会社 自动驾驶辅助系统、自动驾驶辅助方法以及计算机程序
CN107430807A (zh) * 2015-03-31 2017-12-01 爱信艾达株式会社 自动驾驶辅助系统、自动驾驶辅助方法以及计算机程序
CN107533805A (zh) * 2015-04-27 2018-01-02 爱信艾达株式会社 自动驾驶辅助系统、自动驾驶辅助方法以及计算机程序

Cited By (38)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114506324A (zh) * 2020-10-23 2022-05-17 上海汽车集团股份有限公司 一种车道决策方法及相关装置
CN114506324B (zh) * 2020-10-23 2024-03-15 上海汽车集团股份有限公司 一种车道决策方法及相关装置
CN112650224A (zh) * 2020-12-11 2021-04-13 国汽(北京)智能网联汽车研究院有限公司 自动驾驶仿真的方法、装置、设备及存储介质
CN112710317A (zh) * 2020-12-14 2021-04-27 北京四维图新科技股份有限公司 自动驾驶地图的生成方法、自动驾驶方法及相关产品
CN114802244A (zh) * 2021-01-11 2022-07-29 广东科学技术职业学院 控制无人驾驶车辆的方法
CN114822078A (zh) * 2021-01-28 2022-07-29 武汉智行者科技有限公司 一种目标参考线切换控制方法、装置及存储介质
CN112735187A (zh) * 2021-01-29 2021-04-30 重庆长安汽车股份有限公司 一种自动识别应急车道的系统和方法
CN113823118A (zh) * 2021-02-19 2021-12-21 石家庄铁道大学 一种联合紧急程度和博弈论的智能网联车辆换道方法
CN113823118B (zh) * 2021-02-19 2022-07-08 石家庄铁道大学 一种联合紧急程度和博弈论的智能网联车辆换道方法
CN113022570B (zh) * 2021-03-19 2022-06-10 武汉理工大学 一种车辆换道行为识别方法及设备
CN113022570A (zh) * 2021-03-19 2021-06-25 武汉理工大学 一种车辆换道行为识别方法及设备
CN113104038B (zh) * 2021-03-31 2022-12-20 江铃汽车股份有限公司 车辆换道控制方法、装置、电子设备及可读存储介质
CN113104038A (zh) * 2021-03-31 2021-07-13 江铃汽车股份有限公司 车辆换道控制方法、装置、电子设备及可读存储介质
CN115158308A (zh) * 2021-04-02 2022-10-11 清华大学 一种智能车辆主动避障控制方法及装置、存储介质、终端
CN113085853B (zh) * 2021-04-26 2022-05-17 中汽研（天津）汽车工程研究院有限公司 一种车道内主动躲闪大型车辆的辅助驾驶系统
CN113085853A (zh) * 2021-04-26 2021-07-09 中汽研（天津）汽车工程研究院有限公司 一种车道内主动躲闪大型车辆的辅助驾驶系统
CN113848913B (zh) * 2021-09-28 2023-01-06 北京三快在线科技有限公司 一种无人驾驶设备的控制方法及控制装置
CN113848913A (zh) * 2021-09-28 2021-12-28 北京三快在线科技有限公司 一种无人驾驶设备的控制方法及控制装置
CN114360246A (zh) * 2021-12-28 2022-04-15 北京汇通天下物联科技有限公司 一种高速公路出口匝道的预警方法、装置及存储介质
CN114360246B (zh) * 2021-12-28 2023-03-17 北京汇通天下物联科技有限公司 一种高速公路出口匝道的预警方法、装置及存储介质
CN114407929A (zh) * 2022-01-29 2022-04-29 上海木蚁机器人科技有限公司 无人驾驶绕障处理方法、装置、电子设备及存储介质
CN114407929B (zh) * 2022-01-29 2023-12-12 上海木蚁机器人科技有限公司 无人驾驶绕障处理方法、装置、电子设备及存储介质
CN114435403A (zh) * 2022-02-22 2022-05-06 重庆长安汽车股份有限公司 一种基于环境信息的导航定位校核系统及方法
CN114435403B (zh) * 2022-02-22 2023-11-03 重庆长安汽车股份有限公司 一种基于环境信息的导航定位校核系统及方法
CN114506347A (zh) * 2022-03-24 2022-05-17 重庆长安汽车股份有限公司 一种自动驾驶通过施工路段的系统和方法
CN114999152A (zh) * 2022-05-25 2022-09-02 清华大学 面向混合车流的匝道合流边缘云管控方法
CN114999152B (zh) * 2022-05-25 2024-04-30 清华大学 面向混合车流的匝道合流边缘云管控方法
CN115169908A (zh) * 2022-07-14 2022-10-11 吉林大学 网联环境下基于知觉风险场的车辆行驶状态安全评估方法
CN115195748A (zh) * 2022-08-15 2022-10-18 中汽研（天津）汽车工程研究院有限公司 基于数据驱动的智能汽车拟人化决策规划系统及方法
CN115384502A (zh) * 2022-08-30 2022-11-25 吉林大学 一种基于驾驶风格的自动驾驶车辆车道选择方法
CN115402354A (zh) * 2022-09-26 2022-11-29 苏州挚途科技有限公司 一种针对匝道汇入的车辆控制方法、装置及设备
CN116229715A (zh) * 2023-02-13 2023-06-06 武汉理工大学 一种道路交织区连续流生成方法与系统
CN116229715B (zh) * 2023-02-13 2024-01-12 武汉理工大学 一种道路交织区连续流生成方法与系统
CN116142193A (zh) * 2023-02-28 2023-05-23 驭势科技（北京）有限公司 车辆换道决策方法、装置、电子设备和存储介质
CN117079478A (zh) * 2023-08-16 2023-11-17 北京中交华安科技有限公司 一种基于边缘计算的合流区交通冲突辨识装置
CN117079478B (zh) * 2023-08-16 2024-04-05 北京中交华安科技有限公司 一种基于边缘计算的合流区交通冲突辨识装置
CN118430306A (zh) * 2024-04-22 2024-08-02 福建理工大学 一种自动驾驶车辆智能选择跟驰车辆的方法及终端
CN118605542A (zh) * 2024-08-07 2024-09-06 成都赛力斯科技有限公司 城区驾驶场景中的车辆绕障方法、电子设备及存储介质

Also Published As

Publication number Publication date
CN110667578A (zh) 2020-01-10
CN110667578B (zh) 2021-09-17

Similar Documents

Publication Publication Date Title
WO2020135742A1 (zh) 自动驾驶车辆的横向决策系统及横向决策确定方法
CN111383474B (zh) 自动驾驶车辆的决策系统及方法
US10703362B2 (en) Autonomous driving autonomous system, automated driving assistance method, and computer program
JP6443550B2 (ja) シーン評価装置、走行支援装置、シーン評価方法
JP6451847B2 (ja) 運転計画装置、走行支援装置、運転計画方法
JP6443552B2 (ja) シーン評価装置、走行支援装置、シーン評価方法
JPWO2017013750A1 (ja) 運転計画装置、走行支援装置、運転計画方法
WO2017126250A1 (ja) 運転支援方法及び装置
CN106355918A (zh) 无人驾驶车辆的控制方法、装置和系统
WO2019030916A1 (ja) 車線情報管理方法、走行制御方法及び車線情報管理装置
CN115083190A (zh) 多车道交通路口的自动驾驶系统及方法
US12384413B2 (en) Vehicle determining a driving route based on pass priority and a method for operating the vehicle
KR102596624B1 (ko) 자율 주행 차량들의 방향 전환들에 대한 시그널링
US12116007B2 (en) Trajectory limiting for autonomous vehicles
CN111391830A (zh) 自动驾驶车辆的纵向决策系统及纵向决策确定方法
JP6443551B2 (ja) シーン評価装置、走行支援装置、シーン評価方法
CN119536015A (zh) 一种智能网联汽车决策规划模块仿真测试系统及其方法
CN114217601A (zh) 自驾车的混合决策方法及其系统
KR20230001869A (ko) 차량 제어 장치, 그를 포함한 시스템 및 그 방법
US11460848B1 (en) Biased trajectory progress metric
CN115285141B (zh) 智能驾驶过路口的控制方法、系统以及可读存储介质
JP2021033667A (ja) 地図データ生成方法
CN117184123A (zh) 基于顺应先验对具有不确定测量的道路行为体进行状态识别
WO2020012209A1 (ja) 経路算出方法、運転制御方法及び経路算出装置
CN119261899B (zh) 车辆换道方法及装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19901839

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19901839

Country of ref document: EP

Kind code of ref document: A1