A vehicle intelligent obstacle avoidance method

Abstract

This invention provides a vehicle intelligent avoidance method, comprising the following steps: Step S1, acquiring an image captured by a forward-facing camera at the front of the vehicle as the image to be analyzed; Step S2, establishing a planar coordinate system with the midpoint of the vehicle's front as the origin, marking the positions of vehicles in adjacent lanes in the planar coordinate system, and calculating the longitudinal distance between the vehicle and the adjacent vehicles; Step S3, marking the lane line position on the side of the vehicle's lane furthest from the adjacent vehicle in the planar coordinate system based on the image to be analyzed, calculating the standard lateral offset distance of the vehicle based on the lane line position, and obtaining the end coordinate point of the vehicle after avoidance based on the standard distance value and the longitudinal distance value; Step S4, calculating the avoidance trajectory of the vehicle using a planning formula based on the end coordinate point and the origin; The advantage of this invention is that it can reasonably plan a smooth and gentle avoidance trajectory within the lane based on the real-time situation in adjacent lanes and the current lane, thereby ensuring vehicle driving safety.

Description

Technical Field

[0001] This invention relates to the field of autonomous driving, and more specifically to a method and approach for intelligent obstacle avoidance in vehicles. Background Technology

[0002] Autonomous driving refers to a highly centralized control method in which the work performed by the vehicle driver is fully automated. Autonomous driving methods have functions such as automatic vehicle wake-up and start-up and hibernation, automatic parking, automatic driving, automatic door opening and closing, and automatic fault recovery. They also have multiple operating modes such as normal operation, degraded operation, and operation interruption.

[0003] Currently, overcoming the technical challenge of autonomous driving is a key focus. Autonomous driving includes maintaining vehicle centering, lane detection, and maintaining a safe distance from the vehicle in front. However, current autonomous driving technology is still immature, and road conditions are complex. For example, if a vehicle in an adjacent lane is a large trailer or a large vehicle, or if a vehicle in an adjacent lane is close to the vehicle's lane and traveling at a slow speed, and the autonomous driving system still chooses to slow down and avoid it, it will cause the vehicle to travel too slowly and cause congestion in the vehicle's lane. If the autonomous driving system still chooses to stay in the center of the vehicle's lane, it is easy to cause a collision and traffic accident. Therefore, the vehicle needs to choose to avoid the obstacle. The current avoidance methods are usually overtaking or turning to avoid the obstacle. The technical challenge is to achieve avoidance within the vehicle's lane and plan a smooth and gentle avoidance trajectory. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the purpose of this invention is to provide a vehicle intelligent avoidance method. This vehicle intelligent avoidance method can rationally plan a smooth and gentle avoidance trajectory within the lane according to the real-time situation of adjacent lanes and the lane itself, so as to ensure the driving safety of the vehicle.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A vehicle intelligent obstacle avoidance method includes the following steps:

[0007] Step S1: Obtain the image captured by the front-view camera at the front of the vehicle as the image to be analyzed;

[0008] Step S2: Establish a planar coordinate system with the midpoint of the front of the vehicle as the origin; mark the positions of vehicles in adjacent lanes in the planar coordinate system according to the image to be analyzed; and calculate the longitudinal distance between the vehicle in the adjacent lane and the vehicle based on the vehicle positions.

[0009] Step S3: Mark the lane line position of the vehicle's lane away from the adjacent vehicle in the planar coordinate system according to the image to be analyzed; calculate the standard distance value of the vehicle's lateral offset according to the lane line position; and obtain the end coordinate point of the vehicle after avoiding the adjacent vehicle according to the standard distance value and the longitudinal distance value.

[0010] Step S4: Calculate the vehicle's avoidance trajectory using a planning formula based on the end coordinates and the origin.

[0011] Furthermore, step S3 includes step S31, which takes the centerline between the lane line position and the origin as the longitudinal line of the offset point, and indexes the corresponding point on the longitudinal line of the offset point as the end coordinate point according to the longitudinal distance value.

[0012] Furthermore, the trajectory calculation formula is configured as follows:

[0013]

[0014] Where d is the standard distance value and L is the longitudinal distance value.

[0015] Furthermore, step S3 also includes step S32, which involves obtaining the vehicle speed as a first speed value, obtaining the longitudinal distance between the vehicle and the adjacent vehicle at time t as a first distance value, obtaining the longitudinal distance between the vehicle and the adjacent vehicle at time t2 as a second distance value, calculating the speed of the adjacent vehicle as a second speed value based on the time elapsed from time t to time t2, the first distance value, and the second distance value, and also pre-setting a time distance table, which includes a one-to-one corresponding reference speed range for adjacent vehicles, a reference speed range for the vehicle itself, and a longitudinal distance reference value, and indexing the corresponding longitudinal distance reference value in the time distance table based on the first speed value and the second speed value as the longitudinal distance value.

[0016] Furthermore, step S1 also includes step S11, which determines the vehicle situation in adjacent lanes based on the image to be analyzed. If there is a neighboring vehicle in one adjacent lane, then proceed to step S2. If there are neighboring vehicles in both adjacent lanes, then determine whether the two neighboring vehicles are staggered. If they are staggered, then determine the controllable distance value between the two neighboring vehicles. If the controllable distance value is greater than or equal to the longitudinal distance value, then proceed to step S2.

[0017] Furthermore, step S4 also includes step S41, where when the controllable distance between the two adjacent vehicles is greater than or equal to the longitudinal distance, the speed of the adjacent vehicle far from the vehicle is calculated based on the image to be analyzed as a third speed value. The longitudinal distance reference value corresponding to the first and third speed values ​​indexed in the time-distance table is used as a second longitudinal distance value. The corresponding straight-line end point is indexed on the longitudinal line of the offset point based on the second longitudinal distance value. An offset double line is divided between the lane lines on the side of the straight-line end point far from the adjacent vehicle marked in the plane coordinate system based on the image to be analyzed. The corresponding double-avoidance end point is indexed on the offset double line based on the second longitudinal distance value. The straight-line trajectory is obtained by connecting the end point coordinates and the straight-line end point. The double-avoidance trajectory of the vehicle is calculated by a planning formula based on the straight-line end point and the double-avoidance end point. The overall vehicle trajectory is obtained by connecting the avoidance trajectory, the straight-line trajectory, and the double-avoidance trajectory.

[0018] Furthermore, step S4 also includes step S42, which analyzes whether there are potholes in the vehicle's lane based on the image to be analyzed. If potholes exist, the contour lines of the potholes are planned. Based on the contour lines indexed in the plane coordinate system, several contour points are determined to determine whether each contour point coincides with the avoidance trajectory. If they coincide, the trajectory is updated. If they do not coincide, the avoidance trajectory is executed.

[0019] Furthermore, step S42 also includes step S421, determining the position range of the contour point. If it is located on the avoidance trajectory and / or the re-avoidance trajectory, the longitudinal line and / or the offset double line of the offset point are shifted to the side away from the adjacent vehicle until the trajectory is separated from the contour point, or the longitudinal distance value and / or the secondary longitudinal distance value are adjusted.

[0020] Furthermore, it also includes step S5, acquiring the image to be analyzed during the process of the vehicle avoiding obstacles along the vehicle's driving trajectory, determining the lateral distance between the adjacent vehicle and the avoidance trajectory as the actual distance value, and setting a safe distance value. If the actual distance value is greater than or equal to the safe distance value, the avoidance driving continues; if the actual distance value is less than the safe distance value, a horn command is issued.

[0021] The beneficial effects of this invention are as follows: By visually judging the situation in the two adjacent lanes of the vehicle, it can determine whether the vehicle needs to avoid a collision during autonomous driving. If avoidance is required, a smooth and safe avoidance trajectory can be planned to enable the vehicle to complete the avoidance smoothly. Specifically, by visually judging the longitudinal distance between the vehicle and the adjacent vehicle, the avoidance trajectory is reasonably planned based on the longitudinal distance, the vehicle speed, and the division of the longitudinal line of the offset point. Then, by detecting the position of the vehicle close to the avoidance trajectory, the straight trajectory after the vehicle avoids the first vehicle and the re-avoidance trajectory after the vehicle avoids the second vehicle are planned to enable the vehicle to achieve safe avoidance. Furthermore, based on the longitudinal distance of the adjacent vehicles, it can be determined in advance whether the vehicle can safely complete the avoidance and overtaking. In addition, after the avoidance trajectory is planned, the vehicles in the adjacent lanes are monitored in real time during the avoidance to reduce the collision accidents caused by adjacent vehicles changing lanes or decelerating during the avoidance and overtaking. Attached Figure Description

[0022] Figure 1 This is a flowchart of the obstacle avoidance process of the present invention;

[0023] Figure 2 This is the first scene diagram in this invention;

[0024] Figure 3 This is the second scene diagram in this invention;

[0025] Figure 4 This is the third scene diagram in this invention;

[0026] Figure 5 This is the fourth scene diagram in this invention. Detailed Implementation

[0027] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Identical components are denoted by the same reference numerals. It should be noted that the terms "front," "rear," "left," "right," "upper," and "lower" used in the following description refer to directions in the accompanying drawings, and the terms "bottom surface," "top surface," "inner," and "outer" refer to directions toward or away from the geometric center of a specific component, respectively.

[0028] Because current autonomous driving technology is still immature, and road conditions are complex, such as large trailers or other large vehicles in adjacent lanes, or vehicles in adjacent lanes being close to the vehicle's lane at low speeds, if the autonomous driving system still chooses to slow down and avoid the obstacle, it will result in excessively slow speeds and congestion in the vehicle's lane. If the autonomous driving system continues to drive in the center of the lane, it is prone to collisions and traffic accidents. Therefore, vehicles need to choose to avoid obstacles. Current avoidance methods usually involve overtaking or turning to avoid obstacles. Achieving avoidance within the lane and planning a smooth and gentle avoidance trajectory is a technical challenge. Therefore, this invention designs a smart vehicle avoidance method, specifically as follows: Figure 1 As shown, it includes the following steps:

[0029] Step S1: Obtain the image captured by the front-view camera at the front of the vehicle as the image to be analyzed; assuming the road has three lanes and the vehicle is located in the middle lane, the image to be analyzed includes the image within the vehicle's lane and the images within the two adjacent lanes.

[0030] Step S2: Establish a planar coordinate system with the midpoint of the front of the vehicle as the origin. Mark the positions of vehicles in adjacent lanes in the planar coordinate system according to the image to be analyzed. The vehicle positions are mainly the positions of parking spaces. Calculate the longitudinal distance between the vehicle in the adjacent lane and the vehicle based on the vehicle positions.

[0031] Step S3: Mark the lane line position of the vehicle's lane away from the adjacent vehicle in the planar coordinate system according to the image to be analyzed; calculate the standard distance value of the vehicle's lateral offset according to the lane line position; and obtain the end coordinate point of the vehicle after avoiding the adjacent vehicle based on the standard distance value and the longitudinal distance value.

[0032] Step S4: Calculate the vehicle's avoidance trajectory using a programming formula based on the end coordinates and the origin.

[0033] Step S3 includes step S31, which takes the centerline between the lane line position and the origin as the longitudinal line of the offset point, and indexes the corresponding point on the longitudinal line of the offset point as the end coordinate point according to the longitudinal distance value.

[0034] The trajectory calculation formula is configured as follows:

[0035]

[0036] Where d is the standard distance value and L is the longitudinal distance value.

[0037] For example: Figure 2In the first scenario shown, there is car B in the left lane, and this car is car A. There are no vehicles in the lane directly in front of this car or in the lane to its right. Analysis of the image shows that car B is moving slowly and is approaching this car's lane. Since this car's speed is within the speed limit, this car will automatically plan an avoidance trajectory to safely pass car B until it overtakes it. Specifically, a planar coordinate system is first established with the center point of this car's front as the center. The midpoint of car B's rear is marked as G(x1, y1) in the coordinate system. The longitudinal distance L can be calculated from the coordinates of point G, and then the distance between the lane lines to the right of this car is calculated. The distance is calculated by drawing the center line between the two points as the longitudinal line of the offset point. The standard distance d between the vehicle and the longitudinal line of the offset point is calculated. At this time, the trajectory of vehicle A avoiding vehicle B is a curve, and the end point of the curve is on the longitudinal line of the offset point. After the lateral offset, the front of vehicle A is flush with the rear of vehicle B, that is, the end coordinate point H(x2, y2). The avoidance trajectory can be planned through the origin and the H point, that is, the real-time angle of the steering wheel rotation can be planned. The avoidance trajectory can be an arc shape, and the steering wheel angle remains unchanged, so that vehicle A directly travels to the H point. This invention calculates a smooth curve through the trajectory formula.

[0038] Step S3 also includes step S32, which involves obtaining the vehicle's speed as a first speed value, obtaining the longitudinal distance between the vehicle and adjacent vehicles at time t as a first distance value, obtaining the longitudinal distance between the vehicle and adjacent vehicles at time t2 as a second distance value, calculating the speed of the adjacent vehicle as the second speed value based on the time elapsed from time t to time t2, the first distance value, and the second distance value, and also pre-setting a time-distance table, which includes a one-to-one corresponding reference speed range for adjacent vehicles, a reference speed range for the vehicle itself, and a longitudinal distance reference value. The corresponding longitudinal distance is indexed in the time-distance table based on the first speed value and the second speed value. The longitudinal distance is determined by the reference value. Since the speeds of both vehicle A and vehicle B affect the range of the longitudinal distance value, if vehicle B is too close to vehicle A's lane, the longitudinal distance value will decrease as vehicle A moves forward a certain distance. This results in a larger arc of vehicle A's avoidance trajectory curve, a larger steering wheel angle, and a smaller safety factor. Therefore, a distance-time table is set up to index reasonable longitudinal distance values ​​based on the speeds of vehicles A and B. As long as the actual longitudinal distance between vehicle A and vehicle B is greater than or equal to the indexed longitudinal distance value, the system can reasonably plan a smooth avoidance trajectory.

[0039] Step S1 also includes step S11, which determines the vehicle situation in adjacent lanes based on the image to be analyzed. If there is a neighboring vehicle in one adjacent lane, then proceed to step S2. If there are neighboring vehicles in both adjacent lanes, then determine whether the two neighboring vehicles are staggered. If they are staggered, then determine the controllable distance value between the two neighboring vehicles. If the controllable distance value is greater than or equal to the longitudinal distance value, then proceed to step S2.

[0040] For example: Figure 3 The second scene and Figure 4 In the third scenario, both the second and third scenarios involve a car B in the left lane and a car C in the right lane. In the second scenario, the controllable distance L1 between the rear of car C and the front of car B is greater than or equal to the indexed longitudinal distance L, so the trajectory planning for the second avoidance can be safely achieved. However, in the third scenario, the rear of car C is aligned with the front of car B. If car A completes the first avoidance and moves to the right lane, it is likely to collide with car C, so the trajectory planning for the second avoidance cannot be achieved.

[0041] Step S4 also includes step S41, where when the controllable distance between two adjacent vehicles is greater than or equal to the longitudinal distance, the speed of the adjacent vehicle that is far from the vehicle is calculated based on the image to be analyzed as the third speed value. The longitudinal distance reference value corresponding to the first speed value and the third speed value is indexed in the time distance table as the second longitudinal distance value. The straight end point is indexed on the longitudinal line of the offset point based on the second longitudinal distance value. The lane line position on the side of the straight end point that is far from the adjacent vehicle is marked in the plane coordinate system based on the image to be analyzed to obtain the offset double line. The double avoidance end point is indexed on the offset double line based on the second longitudinal distance value. The straight trajectory is obtained by connecting the end coordinate point and the straight end point. The double avoidance trajectory of the vehicle is calculated by the planning formula based on the straight end point and the double avoidance end point. The avoidance trajectory, the straight trajectory and the double avoidance trajectory are connected to obtain the overall vehicle driving trajectory.

[0042] For example: Figure 2 The first scene shown and Figure 3 In the second scenario shown, the trajectory of car A when avoiding overtaking is in three segments. The first scenario includes an avoidance trajectory, a straight trajectory, and a re-avoidance trajectory. The re-avoidance trajectory is when car A reaches position A1 (when the center point of car A1 is aligned with the front of car B). At this point, car A plans to return to the center of its lane and drive normally. According to the second scenario, the initial point of the avoidance trajectory is the origin, and the ending point is point H. The initial point of the straight trajectory is point H, and the ending point is point K. The starting point of the re-avoidance trajectory is point K. An offset double line is drawn on the left half of the lane. The re-avoidance trajectory's ending point M is marked according to the offset double line. There are two ways to divide the offset double line: one is to take the center line between the left lane line of car A's lane and the straight trajectory, and the other is to take the center line between the center line of car A's lane and the left lane line of car A's lane. The choice is mainly based on car B's speed and the situation. Generally, the second method is considered as the offset double line.

[0043] Because driving into potholes during avoidance can cause vehicle swaying and passenger discomfort, and if the vehicle avoidance speed is too high, turning into the pothole can easily cause loss of vehicle control, step S4 also includes step S42, which analyzes whether there are potholes in the vehicle's lane based on the image to be analyzed. If potholes exist, the contour line of the pothole is planned, and several contour points are indexed according to the contour line in the plane coordinate system. It is determined whether each contour point coincides with the avoidance trajectory. If they coincide, the trajectory is updated. If they do not coincide, the avoidance trajectory is executed. Step S42 also includes step S421, which determines the position range of the contour point. If it is located on the avoidance trajectory and / or the secondary avoidance trajectory, the longitudinal line and / or the secondary offset line of the offset point are shifted to the side away from the adjacent vehicle until the trajectory is separated from the contour point, or the longitudinal distance value and / or the secondary longitudinal distance value are adjusted.

[0044] For example: Figure 5 The fourth scenario shown can be analyzed by comparing images to determine if there are potholes on the road surface. If so, the outline of the potholes can be marked based on image processing. Figure 2-5 Although the avoidance trajectory is only in the form of a line, the vehicle has width. Therefore, based on the width of the vehicle, the trajectory area can be drawn by extending the trajectory line. It can be determined whether the pothole area is within the trajectory area and whether the pothole is located on the wheel line within the trajectory area. If it is, an adjustment is required. Furthermore, if there is overlap, it is determined whether the pothole is on the avoidance trajectory or the repeat avoidance trajectory. Or, if there are two potholes on the avoidance trajectory and the repeat avoidance trajectory respectively, then the longitudinal line or the offset line of the offset point needs to be shifted a certain distance. For example, if the pothole q is located on the avoidance trajectory, the longitudinal line of the offset point P is adjusted to P1. At this time, point H is changed to h. After vehicle A avoids to point h, it travels forward in a straight line for a certain distance and then turns back to the straight trajectory of HK, safely avoiding the pothole.

[0045] It also includes step S5, which involves acquiring the image to be analyzed during the process of the vehicle avoiding obstacles along the vehicle's driving trajectory, determining the lateral distance between the adjacent vehicle and the avoidance trajectory as the actual distance value, and setting a safe distance value. If the actual distance value is greater than or equal to the safe distance value, the avoidance driving continues; if the actual distance value is less than the safe distance value, a horn command is issued.

[0046] The above are merely preferred embodiments of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principle of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A vehicle intelligent obstacle avoidance method, characterized in that: Includes the following steps: Step S1: Acquire the image captured by the front-view camera at the front of the vehicle as the image to be analyzed; Step S2: Establish a planar coordinate system with the midpoint of the front of the vehicle as the origin; mark the positions of vehicles in adjacent lanes in the planar coordinate system according to the image to be analyzed; and calculate the longitudinal distance between the vehicle in the adjacent lane and the vehicle based on the vehicle positions. Step S3: Mark the lane line position of the vehicle's lane away from the adjacent vehicle in the plane coordinate system according to the image to be analyzed; calculate the standard distance value of the vehicle's lateral offset according to the lane line position; and obtain the end coordinate point of the vehicle after avoiding the adjacent vehicle according to the standard distance value and the longitudinal distance value. Step S3 includes step S31, which takes the centerline between the lane line position and the origin as the longitudinal line of the offset point, and indexes the corresponding point on the longitudinal line of the offset point as the end coordinate point according to the longitudinal distance value. Step S3 further includes step S32, which involves obtaining the vehicle speed as a first speed value, obtaining the longitudinal distance between the vehicle and the adjacent vehicle at time t as a first distance value, obtaining the longitudinal distance between the vehicle and the adjacent vehicle at time t2 as a second distance value, calculating the speed of the adjacent vehicle as a second speed value based on the time elapsed from time t to time t2, the first distance value, and the second distance value, and also pre-setting a time distance table, which includes a one-to-one corresponding reference speed range for adjacent vehicles, a reference speed range for the vehicle itself, and a longitudinal distance reference value, and indexing the corresponding longitudinal distance reference value in the time distance table based on the first speed value and the second speed value as the longitudinal distance value. Step S4: Calculate the avoidance trajectory of the vehicle using a planning formula based on the end coordinates and the origin. Step S4 further includes step S41, where when the controllable distance between two adjacent vehicles is greater than or equal to the longitudinal distance, the speed of the adjacent vehicle far from the vehicle is calculated based on the image to be analyzed as the third speed value. The longitudinal distance reference value corresponding to the first and third speed values ​​indexed in the time-distance table is used as the second longitudinal distance value. The corresponding straight-line end point is indexed on the longitudinal line of the offset point based on the second longitudinal distance value. The offset double line is divided between the lane lines on the side of the straight-line end point far from the adjacent vehicle marked in the plane coordinate system based on the image to be analyzed. The corresponding double-avoidance end point is indexed on the offset double line based on the second longitudinal distance value. The straight-line trajectory is obtained by connecting the end point coordinates and the straight-line end point. The double-avoidance trajectory of the vehicle is calculated by a planning formula based on the straight-line end point and the double-avoidance end point. The overall vehicle driving trajectory is obtained by connecting the avoidance trajectory, the straight-line trajectory, and the double-avoidance trajectory.

2. The intelligent vehicle avoidance method according to claim 1, characterized in that: The avoidance trajectory formula is configured as follows: Where d is the standard distance value and L is the longitudinal distance value.

3. The intelligent vehicle avoidance method according to claim 1, characterized in that: Step S1 further includes step S11, which determines the vehicle situation in adjacent lanes based on the image to be analyzed. If there is a neighboring vehicle in one adjacent lane, then proceed to step S2. If there are neighboring vehicles in both adjacent lanes, then determine whether the two neighboring vehicles are staggered. If they are staggered, then determine the controllable distance value between the two neighboring vehicles. If the controllable distance value is greater than or equal to the longitudinal distance value, then proceed to step S2.

4. The intelligent vehicle avoidance method according to claim 1, characterized in that: Step S4 further includes step S42, which analyzes whether there are potholes in the vehicle's lane based on the image to be analyzed. If potholes exist, the contour lines of the potholes are planned. Based on the contour lines indexed in the plane coordinate system, several contour points are determined to determine whether each contour point coincides with the avoidance trajectory. If they coincide, the trajectory is updated. If they do not coincide, the avoidance trajectory is executed.

5. The intelligent vehicle avoidance method according to claim 4, characterized in that: Step S42 further includes step S421, determining the position range of the contour point. If it is located on the avoidance trajectory and / or the re-avoidance trajectory, the longitudinal line and / or the offset double line of the offset point are shifted to the side away from the adjacent vehicle until the trajectory is separated from the contour point, or the longitudinal distance value and / or the secondary longitudinal distance value are adjusted.

6. The intelligent vehicle avoidance method according to claim 5, characterized in that: It also includes step S5, which involves acquiring the image to be analyzed during the process of the vehicle avoiding obstacles along the vehicle's driving trajectory, determining the lateral distance between the adjacent vehicle and the avoidance trajectory as the actual distance value, and setting a safe distance value. If the actual distance value is greater than or equal to the safe distance value, the avoidance driving continues; if the actual distance value is less than the safe distance value, a horn command is issued.

Images