Method and system for yielding to pedestrians by autonomous vehicle

Abstract

The application provides a method and system for automatic driving vehicle to give way to pedestrians, comprising the following steps: S1: screening a pedestrian crossing, calculating a stopping line for giving way to pedestrians by the left and right boundaries of the reference line of the vehicle driving and the intersection of the vehicle planning path and the pedestrian crossing; S2: clustering pedestrians and non-motor vehicles into pedestrian flow and non-motor vehicle flow; S3: if the pedestrian crossing is within the preset distance of the vehicle, or there is a pedestrian or non-motor vehicle towards the lane of the vehicle, the vehicle is slowed down. The application adopts a clustering algorithm, takes the cluster combined together as a pedestrian crossing giving way judgment scene, and outputs the clustering results of the pedestrian flow and the non-motor vehicle flow, so that compared with processing each obstacle, processing the flow can significantly improve the operation efficiency.

Description

Technical Field

[0001] This invention relates to the field of autonomous driving, and more specifically, to a method and system for autonomous vehicles to yield to pedestrians. Background Technology

[0002] As key participants in traffic scenarios, the accurate reasoning and prediction of pedestrians' future trajectories are crucial for autonomous driving and road safety. In traffic situations, pedestrian trajectories are influenced not only by their own intentions but also by surrounding pedestrians and vehicles, making pedestrian trajectory prediction an extremely challenging task. In recent years, pedestrian trajectory prediction has received increasing attention in various fields, especially in autonomous driving, where pedestrian identification and the corresponding responses of autonomous vehicles remain significant challenges.

[0003] Patent document CN114379587A discloses a method, device, electronic device, and readable storage medium for avoiding pedestrians in autonomous driving, relating to the fields of artificial intelligence technology such as autonomous driving and intelligent transportation. The method for avoiding pedestrians in autonomous driving includes: in response to detecting a target pedestrian in front of the vehicle, acquiring motion information and / or traffic light information corresponding to the target pedestrian; obtaining the target pedestrian's movement intention based on the motion information and / or traffic light information; and generating a pedestrian avoidance plan based on the target pedestrian's movement intention. However, this invention does not employ a clustering algorithm to group the clusters together as the pedestrian crossing yielding judgment scenario.

[0004] Patent document CN114312836A discloses a method, device, equipment, and storage medium for an autonomous vehicle to yield to pedestrians, relating to the field of computer technology, and particularly to the field of autonomous driving. The specific implementation involves: detecting a pedestrian crossing in the planned driving path, determining the starting position and influence range of the pedestrian crossing, acquiring the movement information of all pedestrians intending to cross the pedestrian crossing within its influence range, and then performing a collision risk assessment based on this information. If the risk assessment value of a collision between any pedestrian and the vehicle exceeds a preset risk threshold, a virtual wall is generated at the starting position of the pedestrian crossing. This invention involves traversing all collision risks between pedestrians and the vehicle and stopping to yield, resulting in low driving efficiency and difficulty in guaranteeing the safety and comfort of autonomous driving.

[0005] Patent document CN114550121A discloses a clustering-based classification and recognition method for lane-changing scenarios in autonomous driving. This classification method includes: 1) collecting data for each sample point to be classified as a lane-changing scenario; 2) preprocessing the data from step 1); 3) treating each preprocessed sample point as a cluster, using an agglomerative hierarchical clustering algorithm to calculate the distance between the sample points of each cluster and the sample points of all other clusters, and combining the two closest clusters; 4) treating each group of clusters as a lane-changing scenario and outputting the clustering results. However, this invention only uses a coarse clustering method, and its applicability is limited in complex scenarios where pedestrians and non-motorized vehicles must yield to pedestrians at intersections. Summary of the Invention

[0006] In view of the deficiencies in the prior art, the purpose of this invention is to provide a method and system for autonomous vehicles to yield to pedestrians.

[0007] A method for an autonomous vehicle to yield to pedestrians according to the present invention includes:

[0008] Step S1: Select pedestrian crossings and calculate stop lines by using the left and right boundaries of the vehicle driving reference line and the intersection relationship between the vehicle's planned path and the pedestrian crossing.

[0009] Step S2: Cluster pedestrians and non-motorized vehicles into pedestrian flow and non-motorized vehicle flow;

[0010] Step S3: If the pedestrian crossing is within the vehicle's preset distance, or if there are pedestrians or non-motorized vehicles heading towards the vehicle lane, the vehicle will slow down.

[0011] Preferably, in step S1:

[0012] The map is used to select the pedestrian crossings that vehicles will pass through within a preset time period in the future.

[0013] If the left and right boundaries of the vehicle's reference line and the vehicle's planned path intersect with the pedestrian crossing, calculate the stop line. If a stop line is provided on the map, stop and yield in front of the stop line. If no stop line is provided on the map, use the position at a preset distance in front of the pedestrian crossing polygon as the stop line. The pedestrian crossing polygon represents the area where pedestrians will pass or be occupied in the future within a preset time period.

[0014] Preferably, in step S2:

[0015] i. For pedestrian or non-motorized vehicle traffic in motion:

[0016] If the beginning and end of the pedestrian flow are located to the left and right of the planned vehicle path, respectively, it is considered a pedestrian flow that is currently crossing, and vehicles should stop and yield to it;

[0017] ii. If pedestrians or non-motorized vehicles comply with traffic rules, select obstacles within the pedestrian or non-motorized vehicle flow to represent the entire flow, and address the obstacles accordingly:

[0018] Calculate the time from when the vehicle reaches the pedestrian crossing stop line, the time when the obstacle is behind the vehicle during normal passage, the position of the obstacle during the time interval, and the deceleration required for the vehicle to slow down to the stop line:

[0019] a) If a pedestrian or non-motorized vehicle is in the oncoming lane or is more than the preset distance from the vehicle, the vehicle continues to move;

[0020] b) If the deceleration for stopping to yield is lower than the preset value, then stop to yield when the obstacle enters within three lanes of the vehicle;

[0021] c) If the deceleration of the stop-and-yield is greater than the preset value, then stop-and-yield when the obstacle enters within two lanes of the vehicle;

[0022] d) If the deceleration exceeds the vehicle's maximum deceleration limit, the vehicle shall stop and yield after the obstacle enters the vehicle's preset range;

[0023] iii. Situations where pedestrians or non-motorized vehicles do not comply with traffic rules:

[0024] a) Narrow down the screening range for non-motorized vehicle and pedestrian traffic, and do not stop in advance to yield;

[0025] b) If a vehicle is approaching the first pedestrian crossing at an intersection and its deceleration is less than a preset value, it shall stop and yield.

[0026] c) If a vehicle is crossing the second pedestrian crossing at the intersection, the driver should not stop to yield.

[0027] Preferably, in step S2:

[0028] For pedestrians or non-motorized vehicles stationary beside a crosswalk:

[0029] i. When the traffic light for the vehicle's lane is green for either going straight or turning left:

[0030] a) When the first pedestrian crossing a vehicle is about to cross is red, it must not stop to yield to stationary pedestrians or non-motorized vehicles and must apply a speed limit.

[0031] b) When the light turns green for the second pedestrian crossing, set up virtual obstacles and stop before the second pedestrian crossing to yield;

[0032] ii. When the traffic light for the vehicle's lane direction is flashing red, yellow, or green for both going straight and turning left:

[0033] If the first pedestrian crossing a vehicle is about to cross has a green light or is about to turn green, a virtual obstacle is set up at the first pedestrian crossing to allow the vehicle to stop and yield, without considering the handling of the second pedestrian crossing.

[0034] Preferably, the treatment of blind spots at pedestrian crossings is as follows:

[0035] i. Starting from the center of the vehicle's front bumper, connect the two farthest corners of the pedestrian crossing from the vehicle to form a triangle, which represents the vehicle's forward field of vision.

[0036] ii. Traverse all obstacles in the left lane or the right non-motorized vehicle lane. If the obstacle is stationary or its speed is less than the preset value and it is within the vehicle's field of vision, proceed with logic processing. If there is an obstacle higher than the vehicle, there is a blind spot. If the number of obstacles on the right exceeds the preset value, there is a blind spot. After calculating the blind spot, the vehicle's speed is limited so that the vehicle's speed is lower than the preset value before reaching the pedestrian crossing.

[0037] iii. When the distance to the pedestrian crossing is less than a preset value, the blind spot processing is canceled.

[0038] A system for autonomous vehicles to yield to pedestrians according to the present invention includes:

[0039] Module M1: Filters pedestrian crossings and calculates stop lines by using the left and right boundaries of the vehicle's reference line and the intersection relationship between the vehicle's planned path and the pedestrian crossing.

[0040] Module M2: Clusters pedestrians and non-motorized vehicles into pedestrian flow and non-motorized vehicle flow;

[0041] Module M3: If a pedestrian crossing is within a preset distance of the vehicle, or if there are pedestrians or non-motorized vehicles heading towards the vehicle lane, the vehicle will slow down.

[0042] Preferably, in module M1:

[0043] The map is used to select the pedestrian crossings that vehicles will pass through within a preset time period in the future.

[0044] If the left and right boundaries of the vehicle's reference line and the vehicle's planned path intersect with the pedestrian crossing, calculate the stop line. If a stop line is provided on the map, stop and yield in front of the stop line. If no stop line is provided on the map, use the position at a preset distance in front of the pedestrian crossing polygon as the stop line. The pedestrian crossing polygon represents the area where pedestrians will pass or be occupied in the future within a preset time period.

[0045] Preferably, in module M2:

[0046] i. For pedestrian or non-motorized vehicle traffic in motion:

[0047] If the beginning and end of the pedestrian flow are located to the left and right of the planned vehicle path, respectively, it is considered a pedestrian flow that is currently crossing, and vehicles should stop and yield to it;

[0048] ii. If pedestrians or non-motorized vehicles comply with traffic rules, select obstacles within the pedestrian or non-motorized vehicle flow to represent the entire flow, and address the obstacles accordingly:

[0049] Calculate the time from when the vehicle reaches the pedestrian crossing stop line, the time when the obstacle is behind the vehicle during normal passage, the position of the obstacle during the time interval, and the deceleration required for the vehicle to slow down to the stop line:

[0050] a) If a pedestrian or non-motorized vehicle is in the oncoming lane or is more than the preset distance from the vehicle, the vehicle continues to move;

[0051] b) If the deceleration for stopping to yield is lower than the preset value, then stop to yield when the obstacle enters within three lanes of the vehicle;

[0052] c) If the deceleration of the stop-and-yield is greater than the preset value, then stop-and-yield when the obstacle enters within two lanes of the vehicle;

[0053] d) If the deceleration exceeds the vehicle's maximum deceleration limit, the vehicle shall stop and yield after the obstacle enters the vehicle's preset range;

[0054] iii. Situations where pedestrians or non-motorized vehicles do not comply with traffic rules:

[0055] a) Narrow down the screening range for non-motorized vehicle and pedestrian traffic, and do not stop in advance to yield;

[0056] b) If a vehicle is approaching the first pedestrian crossing at an intersection and its deceleration is less than a preset value, it shall stop and yield.

[0057] c) If a vehicle is crossing the second pedestrian crossing at the intersection, the driver should not stop to yield.

[0058] Preferably, in module M2:

[0059] For pedestrians or non-motorized vehicles stationary beside a crosswalk:

[0060] i. When the traffic light for the vehicle's lane is green for either going straight or turning left:

[0061] a) When the first pedestrian crossing a vehicle is about to cross is red, it must not stop to yield to stationary pedestrians or non-motorized vehicles and must apply a speed limit.

[0062] b) When the light turns green for the second pedestrian crossing, set up virtual obstacles and stop before the second pedestrian crossing to yield;

[0063] ii. When the traffic light for the vehicle's lane direction is flashing red, yellow, or green for both going straight and turning left:

[0064] If the first pedestrian crossing a vehicle is about to cross has a green light or is about to turn green, a virtual obstacle is set up at the first pedestrian crossing to allow the vehicle to stop and yield, without considering the handling of the second pedestrian crossing.

[0065] Preferably, the treatment of blind spots at pedestrian crossings is as follows:

[0066] i. Starting from the center of the vehicle's front bumper, connect the two farthest corners of the pedestrian crossing from the vehicle to form a triangle, which represents the vehicle's forward field of vision.

[0067] ii. Traverse all obstacles in the left lane or the right non-motorized vehicle lane. If the obstacle is stationary or its speed is less than the preset value and it is within the vehicle's field of vision, proceed with logic processing. If there is an obstacle higher than the vehicle, there is a blind spot. If the number of obstacles on the right exceeds the preset value, there is a blind spot. After calculating the blind spot, the vehicle's speed is limited so that the vehicle's speed is lower than the preset value before reaching the pedestrian crossing.

[0068] iii. When the distance to the pedestrian crossing is less than a preset value, the blind spot processing is canceled.

[0069] Compared with the prior art, the present invention has the following beneficial effects:

[0070] 1. This invention uses a clustering algorithm to group the clusters together as the pedestrian crossing yielding judgment scenario, and outputs the clustering results of pedestrian flow and non-motorized vehicle flow. Compared with processing each obstacle, processing the flow can significantly improve the operating efficiency.

[0071] 2. This invention uses the left and right boundaries of the reference line for the vehicle's driving and the position of the vehicle's planned path at a preset distance before intersecting with the abstracted pedestrian crossing polygon as the stopping and yielding line, and reduces speed in advance to ensure safety.

[0072] 3. Based on the existing methods and logic rules for determining the pedestrian yielding mode of autonomous vehicles at zebra crossings, this invention adds relevant logic rules for intersection traffic lights, straight or turning through the second pedestrian crossing, based on the current speed of the autonomous vehicle and the distance to the pedestrian. This makes the movement of the autonomous vehicle more human-like, improves traffic efficiency and enhances safety.

[0073] 4. The present invention adds a blind spot processing technology solution, and the blind spot processing is cancelled in a timely manner according to the blind spot judgment range based on the distance from the sidewalk, thereby improving the driving experience and safety. Attached Figure Description

[0074] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0075] Figure 1 This is a schematic diagram of the process of the present invention. Detailed Implementation

[0076] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the scope of protection of the present invention.

[0077] Example 1:

[0078] This invention provides a method for autonomous vehicles to yield to pedestrians, such as... Figure 1 As shown, it includes:

[0079] Step S1: Select pedestrian crossings and calculate stop lines by using the left and right boundaries of the vehicle driving reference line and the intersection relationship between the vehicle's planned path and the pedestrian crossing.

[0080] Specifically, in step S1:

[0081] The map is used to select the pedestrian crossings that vehicles will pass through within a preset time period in the future.

[0082] If the left and right boundaries of the vehicle's reference line and the vehicle's planned path intersect with the pedestrian crossing, calculate the stop line. If a stop line is provided on the map, stop and yield in front of the stop line. If no stop line is provided on the map, use the position at a preset distance in front of the pedestrian crossing polygon as the stop line. The pedestrian crossing polygon represents the area where pedestrians will pass or be occupied in the future within a preset time period.

[0083] Step S2: Cluster pedestrians and non-motorized vehicles into pedestrian flow and non-motorized vehicle flow;

[0084] Specifically, in step S2:

[0085] i. For pedestrian or non-motorized vehicle traffic in motion:

[0086] If the beginning and end of the pedestrian flow are located to the left and right of the planned vehicle path, respectively, it is considered a pedestrian flow that is currently crossing, and vehicles should stop and yield to it;

[0087] ii. If pedestrians or non-motorized vehicles comply with traffic rules, select obstacles within the pedestrian or non-motorized vehicle flow to represent the entire flow, and address the obstacles accordingly:

[0088] Calculate the time from when the vehicle reaches the pedestrian crossing stop line, the time when the obstacle is behind the vehicle during normal passage, the position of the obstacle during the time interval, and the deceleration required for the vehicle to slow down to the stop line:

[0089] a) If a pedestrian or non-motorized vehicle is in the oncoming lane or is more than the preset distance from the vehicle, the vehicle continues to move;

[0090] b) If the deceleration for stopping to yield is lower than the preset value, then stop to yield when the obstacle enters within three lanes of the vehicle;

[0091] c) If the deceleration of the stop-and-yield is greater than the preset value, then stop-and-yield when the obstacle enters within two lanes of the vehicle;

[0092] d) If the deceleration exceeds the vehicle's maximum deceleration limit, the vehicle shall stop and yield after the obstacle enters the vehicle's preset range;

[0093] iii. Situations where pedestrians or non-motorized vehicles do not comply with traffic rules:

[0094] a) Narrow down the screening range for non-motorized vehicle and pedestrian traffic, and do not stop in advance to yield;

[0095] b) If a vehicle is approaching the first pedestrian crossing at an intersection and its deceleration is less than a preset value, it shall stop and yield.

[0096] c) If a vehicle is crossing the second pedestrian crossing at the intersection, the driver should not stop to yield.

[0097] Specifically, in step S2:

[0098] For pedestrians or non-motorized vehicles stationary beside a crosswalk:

[0099] i. When the traffic light for the vehicle's lane is green for either going straight or turning left:

[0100] a) When the first pedestrian crossing a vehicle is about to cross is red, it must not stop to yield to stationary pedestrians or non-motorized vehicles and must apply a speed limit.

[0101] b) When the light turns green for the second pedestrian crossing, set up virtual obstacles and stop before the second pedestrian crossing to yield;

[0102] ii. When the traffic light for the vehicle's lane direction is flashing red, yellow, or green for both going straight and turning left:

[0103] If the first pedestrian crossing a vehicle is about to cross has a green light or is about to turn green, a virtual obstacle is set up at the first pedestrian crossing to allow the vehicle to stop and yield, without considering the handling of the second pedestrian crossing.

[0104] Step S3: If the pedestrian crossing is within the vehicle's preset distance, or if there are pedestrians or non-motorized vehicles heading towards the vehicle lane, the vehicle will slow down.

[0105] Specifically, regarding the handling of blind spots at pedestrian crossings:

[0106] i. Starting from the center of the vehicle's front bumper, connect the two farthest corners of the pedestrian crossing from the vehicle to form a triangle, which represents the vehicle's forward field of vision.

[0107] ii. Traverse all obstacles in the left lane or the right non-motorized vehicle lane. If the obstacle is stationary or its speed is less than the preset value and it is within the vehicle's field of vision, proceed with logic processing. If there is an obstacle higher than the vehicle, there is a blind spot. If the number of obstacles on the right exceeds the preset value, there is a blind spot. After calculating the blind spot, the vehicle's speed is limited so that the vehicle's speed is lower than the preset value before reaching the pedestrian crossing.

[0108] iii. When the distance to the pedestrian crossing is less than a preset value, the blind spot processing is canceled.

[0109] Example 2:

[0110] The present invention also provides a system for autonomous vehicles to yield to pedestrians. The system for autonomous vehicles to yield to pedestrians can be implemented by executing the process steps of the method for autonomous vehicles to yield to pedestrians. That is, those skilled in the art can understand the method for autonomous vehicles to yield to pedestrians as a preferred embodiment of the system for autonomous vehicles to yield to pedestrians.

[0111] A system for autonomous vehicles to yield to pedestrians according to the present invention includes:

[0112] Module M1: Filters pedestrian crossings and calculates stop lines by using the left and right boundaries of the vehicle's reference line and the intersection relationship between the vehicle's planned path and the pedestrian crossing.

[0113] Specifically, in module M1:

[0114] The map is used to select the pedestrian crossings that vehicles will pass through within a preset time period in the future.

[0115] If the left and right boundaries of the vehicle's reference line and the vehicle's planned path intersect with the pedestrian crossing, calculate the stop line. If a stop line is provided on the map, stop and yield in front of the stop line. If no stop line is provided on the map, use the position at a preset distance in front of the pedestrian crossing polygon as the stop line. The pedestrian crossing polygon represents the area where pedestrians will pass or be occupied in the future within a preset time period.

[0116] Module M2: Clusters pedestrians and non-motorized vehicles into pedestrian flow and non-motorized vehicle flow;

[0117] Specifically, in module M2:

[0118] i. For pedestrian or non-motorized vehicle traffic in motion:

[0119] If the beginning and end of the pedestrian flow are located to the left and right of the planned vehicle path, respectively, it is considered a pedestrian flow that is currently crossing, and vehicles should stop and yield to it;

[0120] ii. If pedestrians or non-motorized vehicles comply with traffic rules, select obstacles within the pedestrian or non-motorized vehicle flow to represent the entire flow, and address the obstacles accordingly:

[0121] Calculate the time from when the vehicle reaches the pedestrian crossing stop line, the time when the obstacle is behind the vehicle during normal passage, the position of the obstacle during the time interval, and the deceleration required for the vehicle to slow down to the stop line:

[0122] a) If a pedestrian or non-motorized vehicle is in the oncoming lane or is more than the preset distance from the vehicle, the vehicle continues to move;

[0123] b) If the deceleration for stopping to yield is lower than the preset value, then stop to yield when the obstacle enters within three lanes of the vehicle;

[0124] c) If the deceleration of the stop-and-yield is greater than the preset value, then stop-and-yield when the obstacle enters within two lanes of the vehicle;

[0125] d) If the deceleration exceeds the vehicle's maximum deceleration limit, the vehicle shall stop and yield after the obstacle enters the vehicle's preset range;

[0126] iii. Situations where pedestrians or non-motorized vehicles do not comply with traffic rules:

[0127] a) Narrow down the screening range for non-motorized vehicle and pedestrian traffic, and do not stop in advance to yield;

[0128] b) If a vehicle is approaching the first pedestrian crossing at an intersection and its deceleration is less than a preset value, it shall stop and yield.

[0129] c) If a vehicle is crossing the second pedestrian crossing at the intersection, the driver should not stop to yield.

[0130] Specifically, in module M2:

[0131] For pedestrians or non-motorized vehicles stationary beside a crosswalk:

[0132] i. When the traffic light for the vehicle's lane is green for either going straight or turning left:

[0133] a) When the first pedestrian crossing a vehicle is about to cross is red, it must not stop to yield to stationary pedestrians or non-motorized vehicles and must apply a speed limit.

[0134] b) When the light turns green for the second pedestrian crossing, set up virtual obstacles and stop before the second pedestrian crossing to yield;

[0135] ii. When the traffic light for the vehicle's lane direction is flashing red, yellow, or green for both going straight and turning left:

[0136] If the first pedestrian crossing a vehicle is about to cross has a green light or is about to turn green, a virtual obstacle is set up at the first pedestrian crossing to allow the vehicle to stop and yield, without considering the handling of the second pedestrian crossing.

[0137] Module M3: If a pedestrian crossing is within a preset distance of the vehicle, or if there are pedestrians or non-motorized vehicles heading towards the vehicle lane, the vehicle will slow down.

[0138] Specifically, regarding the handling of blind spots at pedestrian crossings:

[0139] i. Starting from the center of the vehicle's front bumper, connect the two farthest corners of the pedestrian crossing from the vehicle to form a triangle, which represents the vehicle's forward field of vision.

[0140] ii. Traverse all obstacles in the left lane or the right non-motorized vehicle lane. If the obstacle is stationary or its speed is less than the preset value and it is within the vehicle's field of vision, proceed with logic processing. If there is an obstacle higher than the vehicle, there is a blind spot. If the number of obstacles on the right exceeds the preset value, there is a blind spot. After calculating the blind spot, the vehicle's speed is limited so that the vehicle's speed is lower than the preset value before reaching the pedestrian crossing.

[0141] iii. When the distance to the pedestrian crossing is less than a preset value, the blind spot processing is canceled.

[0142] Example 3:

[0143] Example 3 is a preferred example of Example 1, and is used to illustrate the present invention in more detail.

[0144] This invention relates to the field of autonomous driving technology, and more particularly to the fields of artificial intelligence technology such as autonomous driving and intelligent transportation. Specifically, it provides a method for autonomous vehicles to yield to pedestrians. This solves the problem of yielding to pedestrians (non-motorized vehicles) when crossing crosswalks during autonomous driving.

[0145] To enable autonomous vehicles to obey traffic rules and cross pedestrian crossings efficiently and safely, the following main steps are taken:

[0146] 1. Use high-precision maps to identify future pedestrian crossings;

[0147] 2. The vehicle's reference line and the planned path intersect with the abstracted pedestrian crossing polygon. If there is an intersection, the processing logic is entered to calculate the stop line. If there is a stop line provided by the map, the vehicle yields before the stop line; otherwise, the position 1m in front of the polygon is taken as the stop line.

[0148] 3. If the pedestrian crossing is within the planned distance, reduce the speed limit to 30 kph in advance (regardless of whether there are pedestrians, to ensure safety and prevent pedestrians or non-motorized vehicles from suddenly entering blind spots such as flower beds and green belts);

[0149] 4. If there are pedestrians and non-motorized vehicles in and around the pedestrian crossing, first use a clustering algorithm to cluster pedestrians and non-motorized vehicles into human flow and non-motorized vehicle flow. The speed and direction of the flow are basically consistent. By clustering obstacles, the flow can be processed by processing each obstacle, which can significantly improve the running time.

[0150] 5. For pedestrians or non-motorized vehicles traveling towards the lane, regardless of whether yielding is required in subsequent calculations, slow down again in advance to ensure safety and prevent untimely braking in case of abnormal situations.

[0151] Pedestrian and non-motorized vehicle traffic will be divided into two categories for processing: stationary and moving traffic.

[0152] For pedestrian or non-motorized vehicle traffic:

[0153] 1. If the beginning and ending points of the pedestrian flow are located to the left and right of the planned path of the vehicle, it means that the pedestrian flow is crossing and you must stop and give way no matter what;

[0154] 2. If pedestrians or non-motorized vehicles comply with traffic rules, select key obstacles in the pedestrian or non-motorized vehicle flow to represent the entire flow, and only handle the key obstacles (different scope of handling according to traffic rules, such as pedestrians crossing the crosswalk on the left when a vehicle is turning left is considered to be complying with traffic rules, while pedestrians crossing against a red light when a vehicle is going straight is considered to be violating traffic rules):

[0155] a. Calculate the time from when the vehicle arrives at the pedestrian crossing stop line to when the obstacle is behind the vehicle during normal passage using the physical velocity-time dynamics formula. Determine the position of the key obstacle during this time period and calculate the deceleration required for the vehicle to slow down to the stop line. (During this process, minimum passage speeds will be set for pedestrians and non-motorized vehicles separately. When the obstacle's speed is low, the speed will be increased to prevent the obstacle from accelerating from a standstill with a very low initial speed, or from failing to yield to vulnerable groups (such as the elderly or children) who are moving slowly.)

[0156] i. If the pedestrian is still in the oncoming lane or far away from your vehicle (in cases where the road is narrow and you cannot rely solely on lane information), then you do not need to yield to the pedestrian.

[0157] ii. Different yielding ranges are handled based on the deceleration during yielding:

[0158] 1) If the deceleration for stopping to yield is comfortable, then stop to yield when the critical obstacle will move into the vicinity of your vehicle within three lanes.

[0159] 2) If the deceleration of stopping to yield is large, then yield only when the key obstacle moves into a range of two lanes from your vehicle;

[0160] 3) If the deceleration exceeds the vehicle's maximum deceleration limit, yielding will only be done if a critical obstacle would encroach on the vehicle's safe zone.

[0161] 3. If you do not obey traffic rules:

[0162] a. Narrow the screening scope for non-motorized vehicle and pedestrian traffic, no longer consider pedestrians and non-motorized vehicles that may enter the crosswalk from outside the crosswalk area, and no longer give way in advance;

[0163] b. If your vehicle is about to cross the first pedestrian crossing at the intersection, you may slow down and yield, provided that you have a small deceleration.

[0164] c. If your vehicle is about to cross the second pedestrian crossing at an intersection (mainly for straight-ahead traffic), do not yield to pedestrians in advance for two reasons: 1. Your vehicle is traveling normally at high speed, while the pedestrian is violating traffic rules; 2. Sudden braking in the middle of an intersection could cause a rear-end collision. Therefore, for pedestrians who do not comply with traffic rules, brake only if a collision is likely, and do not brake in advance to yield.

[0165] For pedestrians or non-motorized vehicles stationary beside a crosswalk:

[0166] The main scope of this operation is to allow vehicles to yield to each other when making right turns by reading traffic light information.

[0167] 1. When the traffic light for your lane is green for going straight / turning left:

[0168] a. If the first pedestrian crossing that a vehicle is about to cross is red, there is no need to brake in advance to yield to stationary pedestrians / non-motorized vehicles; only speed limits should be applied.

[0169] b. When the second pedestrian crossing that a vehicle is about to pass through has a green light, and stationary pedestrians / non-motorized vehicles intend to cross, virtual obstacles need to be set up in advance and vehicles need to stop and yield before the second pedestrian crossing.

[0170] 2. When the traffic light in your lane is flashing red, yellow, or green for both straight and left turns:

[0171] At this time, the first pedestrian crossing that the vehicle is about to cross is green or about to turn green, and the stationary pedestrian / non-motorized vehicle flow will intend to cross. At this time, virtual obstacles need to be set up at the first pedestrian crossing to allow vehicles to stop and yield, and there is no need to consider the handling of the second pedestrian crossing.

[0172] Addressing blind spots at pedestrian crossings:

[0173] Blind spots refer to the areas that may contain stationary, moving, or soon-to-be-moving pedestrians or non-motorized vehicles that cannot be detected by the vehicle's cameras or lidar. If blind spot mitigation is not considered, a pedestrian may suddenly appear, and the vehicle may not be able to brake in time, resulting in danger.

[0174] 1. First, determine the scope of blind spot handling (when the vehicle needs to turn right, similar to a person driving):

[0175] a. A long row of motor vehicles is parked on the left, blocking pedestrians and non-motorized vehicles that may pass on the left;

[0176] b. There is a group of stationary non-motorized vehicles parked on the right. There may be pedestrians and non-motorized vehicles about to cross the pedestrian crossing at the edge in front of the non-motorized vehicles.

[0177] 2. As mentioned above, our blind spot assessment mainly addresses pedestrians / non-motorized vehicles whose view is obstructed when turning right. We only need to address the first pedestrian crossing; after passing the first, there is no blind spot at the second crossing. Therefore, we calculate the existence of a blind spot before proceeding to the first pedestrian crossing:

[0178] a. Starting from the center of the front bumper of the vehicle, and ending at the two farthest corners of the pedestrian crossing from the vehicle, a triangle is formed. This area is considered to be the vehicle's forward field of vision.

[0179] b. Iterate through all motor vehicles in the left lane or non-motor vehicles in the right lane. If they are stationary or moving at a very low speed and are within the polygon of the vehicle's field of vision, proceed with the following logic: If there are trucks, lorries, or other vehicles taller than the vehicle, or multiple cars, the field of vision is considered to be obstructed and a blind spot exists; if there are also multiple non-motor vehicles on the right, the field of vision is also considered to be obstructed and a blind spot exists.

[0180] 3. If the blind spot is calculated, the speed of the vehicle is limited so that the vehicle reaches a low speed before reaching the pedestrian crossing (if the vehicle brakes directly, it will give people a bad experience, and there may be no pedestrians or pedestrians who have no intention of crossing but the vehicle stops).

[0181] 4. When the distance to the pedestrian crossing is relatively close, the blind spot treatment is cancelled. At this time, the field of vision can see pedestrians and non-motorized vehicles that may pass on the left and right sides, and then the above logic for handling pedestrian flow can be resumed.

[0182] Those skilled in the art will understand that, in addition to implementing the system, apparatus, and their modules provided by this invention in purely computer-readable program code, the same program can be implemented in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system, apparatus, and their modules provided by this invention can be considered a hardware component, and the modules included therein for implementing various programs can also be considered structures within the hardware component; alternatively, modules for implementing various functions can be considered both software programs implementing the method and structures within the hardware component.

[0183] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A method for yielding to a pedestrian by an autonomous vehicle, the method comprising: include: Step S1: Select pedestrian crossings and calculate stop lines by using the left and right boundaries of the vehicle driving reference line and the intersection relationship between the vehicle's planned path and the pedestrian crossing. Step S2: Cluster pedestrians and non-motorized vehicles into pedestrian flow and non-motorized vehicle flow; Step S3: If the pedestrian crossing is within the vehicle's preset distance, or if there are pedestrians or non-motorized vehicles heading towards the vehicle lane, the vehicle will slow down. In step S2: i. For pedestrian or non-motorized vehicle traffic in motion: If the beginning and end of the pedestrian flow are located to the left and right of the planned vehicle path, respectively, it is considered a pedestrian flow that is currently crossing, and vehicles should stop and yield to it; ii. If pedestrians or non-motorized vehicles comply with traffic rules, select obstacles within the pedestrian or non-motorized vehicle flow to represent the entire flow, and address the obstacles accordingly: Calculate the time from when the vehicle reaches the pedestrian crossing stop line, the time when the obstacle is behind the vehicle during normal passage, the position of the obstacle during the time interval, and the deceleration required for the vehicle to slow down to the stop line: a) If a pedestrian or non-motorized vehicle is in the oncoming lane or is more than the preset distance from the vehicle, the vehicle continues to move; b) If the deceleration for stopping to yield is lower than the preset value, then stop to yield when the obstacle enters within three lanes of the vehicle; c) If the deceleration of the stop-and-yield is greater than the preset value, then stop-and-yield when the obstacle enters within two lanes of the vehicle; d) If the deceleration exceeds the vehicle's maximum deceleration limit, the vehicle shall stop and yield after the obstacle enters the vehicle's preset range; iii. Situations where pedestrians or non-motorized vehicles do not comply with traffic rules: a) Narrow down the screening range for non-motorized vehicle and pedestrian traffic, and do not stop in advance to yield; b) If a vehicle is approaching the first pedestrian crossing at an intersection and its deceleration is less than a preset value, it shall stop and yield. c) If a vehicle is crossing the second pedestrian crossing at the intersection, the driver should not stop to yield.

2. The method of claim 1, wherein, In step S1: The map is used to select the pedestrian crossings that vehicles will pass through within a preset time period in the future; If the left and right boundaries of the vehicle's reference line and the vehicle's planned path intersect with the pedestrian crossing, calculate the stop line. If a stop line is provided on the map, stop and yield in front of the stop line. If no stop line is provided on the map, use the position at a preset distance in front of the pedestrian crossing polygon as the stop line. The pedestrian crossing polygon represents the area where pedestrians will pass or be occupied in the future within a preset time period.

3. The method of claim 1, wherein, In step S2: For pedestrians or non-motorized vehicles stationary beside a crosswalk: i. When the traffic light for the vehicle's lane is green for either going straight or turning left: a) When the first pedestrian crossing a vehicle is about to cross is red, it must not stop to yield to stationary pedestrians or non-motorized vehicles and must apply a speed limit. b) When the light turns green for the second pedestrian crossing, set up virtual obstacles and stop before the second pedestrian crossing to yield; ii. When the traffic light for the vehicle's lane direction is flashing red, yellow, or green for both going straight and turning left: If the first pedestrian crossing a vehicle is about to cross has a green light or is about to turn green, a virtual obstacle is set up at the first pedestrian crossing to allow the vehicle to stop and yield, without considering the handling of the second pedestrian crossing.

4. The method for an autonomous vehicle to yield to pedestrians according to claim 1, characterized in that: Addressing blind spots at pedestrian crossings: i. Starting from the center of the vehicle's front bumper, connect the two farthest corners of the pedestrian crossing from the vehicle to form a triangle, which represents the vehicle's forward field of vision. ii. Traverse all obstacles in the left lane or the right non-motorized vehicle lane. If the obstacle is stationary or its speed is less than a preset value and it is within the vehicle's field of vision, proceed to logic processing. If there are obstacles higher than the vehicle, there will be blind spots in the field of vision; The number of obstacles on the right exceeds the preset value, and there is a blind spot in the field of vision. After calculating the blind spot, the vehicle speed is limited, and the vehicle speed is lower than the preset value before reaching the pedestrian crossing. iii. When the distance to the pedestrian crossing is less than a preset value, the blind spot processing is canceled.

5. A system for yielding to a pedestrian by an autonomous vehicle, the system comprising: include: Module M1: Filters pedestrian crossings and calculates stop lines by using the left and right boundaries of the vehicle's reference line and the intersection relationship between the vehicle's planned path and the pedestrian crossing. Module M2: Clusters pedestrians and non-motorized vehicles into pedestrian flow and non-motorized vehicle flow; Module M3: If a pedestrian crossing is within a preset distance of the vehicle, or if there are pedestrians or non-motorized vehicles heading towards the vehicle lane, the vehicle will slow down. In module M2: i. For pedestrian or non-motorized vehicle traffic in motion: If the beginning and end of the pedestrian flow are located to the left and right of the planned vehicle path, respectively, it is considered a pedestrian flow that is currently crossing, and vehicles should stop and yield to it; ii. If pedestrians or non-motorized vehicles comply with traffic rules, select obstacles within the pedestrian or non-motorized vehicle flow to represent the entire flow, and address the obstacles accordingly: Calculate the time from when the vehicle reaches the pedestrian crossing stop line, the time when the obstacle is behind the vehicle during normal passage, the position of the obstacle during the time interval, and the deceleration required for the vehicle to slow down to the stop line: a) If a pedestrian or non-motorized vehicle is in the oncoming lane or is more than the preset distance from the vehicle, the vehicle continues to move; b) If the deceleration for stopping to yield is lower than the preset value, then stop to yield when the obstacle enters within three lanes of the vehicle; c) If the deceleration of the stop-and-yield is greater than the preset value, then stop-and-yield when the obstacle enters within two lanes of the vehicle; d) If the deceleration exceeds the vehicle's maximum deceleration limit, the vehicle shall stop and yield after the obstacle enters the vehicle's preset range; iii. Situations where pedestrians or non-motorized vehicles do not comply with traffic rules: a) Narrow down the screening range for non-motorized vehicle and pedestrian traffic, and do not stop in advance to yield; b) If a vehicle is approaching the first pedestrian crossing at an intersection and its deceleration is less than a preset value, it shall stop and yield. c) If a vehicle is crossing the second pedestrian crossing at the intersection, the driver should not stop to yield.

6. The system for autonomous vehicles to yield to pedestrians according to claim 5, characterized in that, In module M1: The map is used to select the pedestrian crossings that vehicles will pass through within a preset time period in the future; If the left and right boundaries of the vehicle's reference line and the vehicle's planned path intersect with the pedestrian crossing, calculate the stop line. If a stop line is provided on the map, stop and yield in front of the stop line. If no stop line is provided on the map, use the position at a preset distance in front of the pedestrian crossing polygon as the stop line. The pedestrian crossing polygon represents the area where pedestrians will pass or be occupied in the future within a preset time period.

7. The system for yielding to a pedestrian for an autonomous vehicle of claim 5, wherein, In module M2: For pedestrians or non-motorized vehicles stationary beside a crosswalk: i. When the traffic light for the vehicle's lane is green for either going straight or turning left: a) When the first pedestrian crossing a vehicle is about to cross is red, it must not stop to yield to stationary pedestrians or non-motorized vehicles and must apply a speed limit. b) When the light turns green for the second pedestrian crossing, set up virtual obstacles and stop before the second pedestrian crossing to yield; ii. When the traffic light for the vehicle's lane direction is flashing red, yellow, or green for both going straight and turning left: If the first pedestrian crossing a vehicle is about to cross has a green light or is about to turn green, a virtual obstacle is set up at the first pedestrian crossing to allow the vehicle to stop and yield, without considering the handling of the second pedestrian crossing.

8. The system for autonomous vehicles to yield to pedestrians according to claim 5, characterized in that: Addressing blind spots at pedestrian crossings: i. Starting from the center of the vehicle's front bumper, connect the two farthest corners of the pedestrian crossing from the vehicle to form a triangle, which represents the vehicle's forward field of vision. ii. Traverse all obstacles in the left lane or the right non-motorized vehicle lane. If the obstacle is stationary or its speed is less than a preset value and it is within the vehicle's field of vision, proceed to logic processing. If there are obstacles higher than the vehicle, there will be blind spots in the field of vision; The number of obstacles on the right exceeds the preset value, and there is a blind spot in the field of vision. After calculating the blind spot, the vehicle speed is limited, and the vehicle speed is lower than the preset value before reaching the pedestrian crossing. iii. When the distance to the pedestrian crossing is less than a preset value, the blind spot processing is canceled.

Images