A vehicle automatic control method, an automatic driving vehicle and a computer storage medium

Abstract

The application relates to the field of automatic driving, and discloses a vehicle automatic control method, an automatic driving vehicle and a computer storage medium. The vehicle automatic control method comprises the following steps: determining that an obstacle in front of a host vehicle is a preceding vehicle by using a preset deep learning model, and obtaining a predicted driving track of the preceding vehicle; identifying a three-point U-turn intention of the preceding vehicle by using a preset conflict logic model according to an actual driving track of the preceding vehicle and the predicted driving track of the preceding vehicle; judging a target position to be driven by the preceding vehicle after the three-point U-turn of the preceding vehicle, and determining a driving track of the preceding vehicle to the target position; and controlling the host vehicle to drive according to a preset behavior strategy to avoid the preceding vehicle according to the driving track of the preceding vehicle to the target position. Therefore, the automatic driving system can timely respond when the preceding vehicle has the three-point U-turn intention, timely take a correct behavior strategy to avoid the preceding vehicle, improve safety, and make the automatic driving more intelligent.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of automatic driving, in particular to a vehicle automatic control method, an automatic driving vehicle and a computer storage medium. BACKGROUND

[0002] With the rapid development of automobile technology, automobiles are becoming more and more intelligent, and automatic driving of automobiles is also possible. In the process of automatic driving from the starting point to the destination, in addition to avoiding various obstacles on the road to safely go to the destination, various special operation scenarios also need to be dealt with, and various strange driving behaviors of the front vehicle also need to be taken certain behavior strategies to get out of trouble.

[0003] In addition, roads are various in width, and some roads are relatively small. When the vehicle needs to turn around or turn in a narrow road, sometimes only one direction can not pass, and repeated operations in the middle of the road are needed to successfully pass. When the automatic driving system of the automobile encounters the above-mentioned scenario, it cannot understand the purpose of the front vehicle, and the situation of too close distance or unreasonable detour affecting the reversing of the front vehicle may occur. At this time, if the correct avoidance logic is not taken in time, only stopping behind the front vehicle will also cause traffic congestion and other traffic problems. SUMMARY

[0004] The embodiment of the present application aims to provide a vehicle automatic control method, an automatic driving vehicle and a computer storage medium, which can solve the problem that the existing automatic driving system of the automobile cannot understand the purpose of the front vehicle when the front vehicle turns around or turns in the road, causing too close distance or unreasonable detour affecting the reversing of the front vehicle.

[0005] To solve the above technical problems, the first aspect of the present application provides a vehicle automatic control method, comprising:

[0006] determining that the obstacle in front of the host vehicle is a front vehicle by using a preset deep learning model, and obtaining a predicted driving track of the front vehicle;

[0007] identifying a three-point turning intention of the front vehicle by using a preset conflict logic model according to the actual driving track of the front vehicle and the predicted driving track of the front vehicle;

[0008] judging a target position to be driven by the front vehicle after three-point turning according to the three-point turning intention of the front vehicle, and determining a driving track of the front vehicle to the target position;

[0009] controlling the host vehicle to drive according to a preset behavior strategy to avoid the front vehicle according to the driving track of the front vehicle to the target position.

[0010] Optionally, the determining, by the preset deep learning model, that the obstacle in front of the host vehicle is a preceding vehicle and determining a driving trajectory of the preceding vehicle comprises:

[0011] controlling a sensor of the host vehicle to capture a point cloud of the obstacle in front of the host vehicle in real time;

[0012] after obtaining the point cloud of the obstacle in front of the host vehicle, determining, by a preset deep learning model, whether the point cloud of the obstacle in front of the host vehicle is a vehicle contour and a vehicle orientation, and calculating a vehicle speed according to a change in the vehicle orientation and the vehicle position according to the vehicle contour;

[0013] obtaining a predicted driving trajectory of the preceding vehicle according to the vehicle orientation, the vehicle position and the vehicle speed.

[0014] Optionally, the determining, by the preset deep learning model, that the obstacle in front of the host vehicle is a preceding vehicle and determining a driving trajectory of the preceding vehicle further comprises: superimposing the vehicle contour, the vehicle orientation, the vehicle position, the vehicle speed and the driving trajectory of the preceding vehicle in the display screen of the host vehicle.

[0015] Optionally, the identifying, by a preset conflict logic model, a three-point U-turn intention of the preceding vehicle according to the actual driving trajectory of the preceding vehicle and the predicted driving trajectory of the preceding vehicle comprises:

[0016] when the actual driving trajectory of the preceding vehicle is inconsistent with the predicted driving trajectory of the preceding vehicle, identifying, by a preset conflict logic model, a three-point U-turn intention of the preceding vehicle.

[0017] Optionally, the determining a target position to be driven by the preceding vehicle after a three-point U-turn of the preceding vehicle according to the three-point U-turn intention of the preceding vehicle and determining a driving trajectory of the preceding vehicle to the target position comprises:

[0018] determining a target position to be driven by the preceding vehicle after a three-point U-turn of the preceding vehicle according to the three-point U-turn intention of the preceding vehicle;

[0019] performing path planning according to the current vehicle position of the preceding vehicle and the target position of the preceding vehicle to determine a driving trajectory of the preceding vehicle to the target position.

[0020] Optionally, the controlling the host vehicle to drive according to a preset behavior strategy to avoid the preceding vehicle according to the driving trajectory of the preceding vehicle to the target position comprises:

[0021] identifying whether the preceding vehicle is a parked vehicle;

[0022] According to the distance between the host vehicle and the front vehicle and the road width, the host vehicle is controlled to travel according to a preset behavior strategy to avoid the front vehicle.

[0023] Optionally, the identification of whether the front vehicle is a parked vehicle comprises:

[0024] If the front vehicle does not have further action after parking at the current position for a preset time, the front vehicle is determined to be a parked vehicle.

[0025] If the front vehicle has further action after parking at the current position, the front vehicle is determined to be a three-point turning vehicle.

[0026] Optionally, the control of the host vehicle to travel according to a preset behavior strategy according to the distance between the host vehicle and the front vehicle and the road width comprises:

[0027] When the distance between the host vehicle and the front vehicle is close and the road width has space for detouring, it is determined according to the current stage of the front vehicle whether to yield or directly detour.

[0028] When the distance between the host vehicle and the front vehicle is far, if the road width is insufficient for detouring and needs to be parked behind the front vehicle, the host vehicle is parked to wait for the front vehicle to complete three-point turning before continuing to move forward.

[0029] When the road is narrow and there is not enough safe space for detouring, the host vehicle is parked to wait and observe all trajectories of the front vehicle to complete three-point turning, and it is further determined whether the parking position of the host vehicle will affect the space required by the front vehicle to perform three-point turning.

[0030] When the road is wide, according to the driving trajectory of the front vehicle to the target position, if the driving trajectory of the front vehicle will block the host vehicle in the future, the host vehicle is parked to wait and continues to move forward after the front vehicle completes three-point turning.

[0031] Correspondingly, a second aspect embodiment of the present application provides an automatic driving vehicle, comprising a memory, a processor and a computer program stored in the memory and running on the processor, and the computer program is executed by the processor to realize the vehicle automatic control method of the first aspect embodiment of the present application.

[0032] Correspondingly, a third aspect embodiment of the present application provides a computer storage medium, and a vehicle automatic control method program is stored on the computer storage medium, and the vehicle automatic control method program is executed by a processor to realize the vehicle automatic control method of the first aspect embodiment of the present application.

[0033] The present application provides a vehicle automatic control method, an automatic driving vehicle and a computer storage medium, relative to the prior art, the vehicle automatic control method comprises the following steps: determining that an obstacle in front of a host vehicle is a preceding vehicle by using a preset deep learning model, and obtaining a predicted driving track of the preceding vehicle; identifying a three-point turning intention of the preceding vehicle by using a preset conflict logic model according to an actual driving track of the preceding vehicle and the predicted driving track of the preceding vehicle; determining a target position to be driven by the preceding vehicle after three-point turning of the preceding vehicle according to the three-point turning intention of the preceding vehicle, and determining a driving track of the preceding vehicle to the target position; and controlling the host vehicle to drive according to a preset behavior strategy to avoid the preceding vehicle according to the driving track of the preceding vehicle to the target position. Thus, the automatic driving system can timely respond when the preceding vehicle has a three-point turning intention, timely take a correct behavior strategy to avoid, give the preceding vehicle enough operation space when the preceding vehicle reverses, greatly reduce the contradiction with the preceding vehicle, solve the escape scene, improve the traffic efficiency and safety, and make the automatic driving more intelligent. Thus, the problem that the existing automatic driving system cannot understand the purpose of the preceding vehicle when the preceding vehicle turns or turns in the road, and the preceding vehicle is too close or unreasonable to detour, is solved. BRIEF DESCRIPTION OF DRAWINGS

[0034] One or more embodiments are illustrated by way of example in the figures that form a part of this patent document, these illustrative examples do not limit the embodiments, and elements having the same reference numbers in the figures indicate like elements unless otherwise stated, the figures in the drawings do not constitute a proportional limit.

[0035] Figure 1 is a flowchart of a vehicle automatic control method provided by an embodiment of the present application;

[0036] Figure 2 is a three-point turning schematic diagram of a preceding vehicle driving from reverse to same direction when the preceding vehicle is stuck in reverse in a vehicle automatic control method provided by an embodiment of the present application;

[0037] Figure 3 is a three-point turning schematic diagram of a preceding vehicle driving from reverse to same direction when the preceding vehicle is stuck in reverse in a vehicle automatic control method provided by an embodiment of the present application;

[0038] Figure 4 is a three-point turning schematic diagram of a preceding vehicle driving from same direction to reverse when the preceding vehicle is stuck in same direction in a vehicle automatic control method provided by an embodiment of the present application;

[0039] Figure 5 is a host vehicle avoiding schematic diagram when the host vehicle is close to the preceding vehicle and the road width has space to detour in a vehicle automatic control method provided by an embodiment of the present application;

[0040] Figure 6is a vehicle automatic control method provided by an embodiment of the present application, which is a main vehicle avoidance schematic diagram when there is an obstacle in the middle of the road;

[0041] Figure 7 is a vehicle automatic control method provided by an embodiment of the present application, which is a main vehicle avoidance schematic diagram when there is no obstacle in the middle of the road;

[0042] Figure 8 is a structural schematic diagram of an automatic driving vehicle provided by the present application. DETAILED DESCRIPTION

[0043] In order to facilitate the understanding of the present application, the present application will be described in more detail below in combination with the drawings and specific embodiments. It should be noted that when an element is described as "fixed to" another element, it can be directly on the other element or one or more intervening elements can be present therebetween. When an element is described as "connected to" another element, it can be directly connected to the other element or one or more intervening elements can be present therebetween. The terms "upper", "lower", "inner", "outer", "bottom", etc. used in the specification indicate the orientation or positional relationship shown in the drawings, and are only for the purpose of facilitating the description of the present application and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as limiting the present application. In addition, the terms "first", "second", "third", etc. are only for the purpose of description and cannot be understood as indicating or implying relative importance.

[0044] Unless otherwise defined, all technical and scientific terms used in the specification have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs. The terms used in the specification of the present application are only for the purpose of describing the specific embodiments and are not used to limit the present application. The term "and / or" used in the specification includes any and all combinations of one or more related listed items.

[0045] In addition, the technical features involved in the different embodiments of the present application described below can be combined with each other as long as they do not conflict with each other.

[0046] In one embodiment, as shown in Figure 1 The present application provides a vehicle automatic control method for dealing with a three-point turn of a front vehicle, which comprises:

[0047] S1, determining that the obstacle in front of the host vehicle is a front vehicle by using a preset deep learning model, and obtaining a predicted driving track of the front vehicle, wherein the front vehicle is a vehicle closest to the host vehicle in front of the host vehicle;

[0048] S2, identifying a three-point turning intention of the front vehicle according to the actual driving track of the front vehicle and the predicted driving track of the front vehicle by using a preset conflict logic model;

[0049] S3, judging a target position to be driven by the front vehicle after three-point turning according to the three-point turning intention of the front vehicle, and determining a driving track of the front vehicle to the target position;

[0050] S4, controlling the host vehicle to drive according to a preset behavior strategy to avoid the front vehicle according to the driving track of the front vehicle to the target position.

[0051] In the embodiment, by providing a vehicle automatic control method, the front vehicle is determined to be a front vehicle by using a preset deep learning model, and a predicted driving track of the front vehicle is obtained; a three-point turning intention of the front vehicle is identified according to the actual driving track of the front vehicle and the predicted driving track of the front vehicle by using a preset conflict logic model; a target position to be driven by the front vehicle after three-point turning is judged according to the three-point turning intention of the front vehicle, and a driving track of the front vehicle to the target position is determined; the host vehicle is controlled to drive according to a preset behavior strategy to avoid the front vehicle according to the driving track of the front vehicle to the target position. Thus, the automatic driving system of the host vehicle can timely respond when the front vehicle has a three-point turning intention, and timely take a correct behavior strategy to avoid, give the front vehicle enough operation space when reversing, greatly reduce the contradiction with the front vehicle, solve the escape scene, improve the traffic efficiency and safety, and make the automatic driving more intelligent. Thus, the problem that the existing automatic driving system of the vehicle cannot understand the purpose of the front vehicle when the front vehicle turns or turns in the road, resulting in too close distance or unreasonable detour to affect the reversing of the front vehicle, can be solved.

[0052] In one embodiment, in step S1, the front vehicle is determined to be a front vehicle by using a preset deep learning model, and a predicted driving track of the front vehicle is obtained. Specifically, it includes:

[0053] S11, controlling the sensor of the host vehicle to capture the point cloud of the front obstacle of the host vehicle in real time.

[0054] For example, the sensor of the host vehicle includes a radar, which reflects the point cloud of the front obstacle of the host vehicle by laser and transmits to the automatic driving system. Thus, the point cloud of the front obstacle of the host vehicle is captured in real time.

[0055] S12, after obtaining the point cloud of the obstacle in front of the host vehicle, determining whether the point cloud of the obstacle in front of the host vehicle is a vehicle contour and a vehicle orientation by using a preset deep learning model, and calculating a vehicle position according to the vehicle contour, and calculating a vehicle speed according to the change of the vehicle position and the vehicle orientation.

[0056] Specifically, the sensor of the host vehicle includes a radar. The radar reflects the point cloud of the obstacle in front of the host vehicle by laser, and transmits to the automatic driving system.

[0057] After the automatic driving system obtains the point cloud of the obstacle in front of the host vehicle, the automatic driving system clusters the point cloud, compares the clustered image with a preset image in an image database, confirms that the clustered image is a vehicle image, draws a frame for the clustered image, labels a vehicle contour and an orientation for the frame to form labeled data, and then performs deep learning training on the labeled data by using a preset deep learning model. In this way, after a large amount of labeled data is sent to the preset deep learning model for deep learning training, the automatic driving system can automatically determine whether the point cloud of the obstacle in front of the host vehicle is a vehicle contour and a vehicle orientation, calculate a vehicle position according to the vehicle contour, and calculate a vehicle speed according to the change of the vehicle position and the vehicle orientation. Thus, the vehicle information of the front vehicle can be captured in real time.

[0058] Similarly, after the automatic driving system obtains the point cloud of the obstacle in front of the host vehicle, the automatic driving system clusters the point cloud, compares the clustered image with an image in an image database, and determines that the clustered image includes a vehicle category (for example, a car, a truck, or a bus) image. After the vehicle category image is labeled, labeled data is formed, and then the labeled data is subjected to deep learning training by using a preset deep learning model. In this way, after a large amount of labeled data is sent to the preset deep learning model for deep learning training, the automatic driving system can automatically identify the category of the front vehicle, that is, the vehicle category information of the front vehicle can be captured in real time.

[0059] S13, obtaining a predicted driving trajectory of the front vehicle according to the vehicle orientation, the vehicle position, and the vehicle speed.

[0060] After determining the vehicle orientation, the vehicle position, the vehicle speed, and the vehicle category, the above data is labeled to form labeled data, and then the labeled data is subjected to deep learning training using a preset deep learning model. In this way, after a large amount of labeled data is sent to the preset deep learning model for deep learning training, the future trajectory of the front vehicle can be predicted according to the real-time state of the above data of the front vehicle, so that the predicted driving trajectory of the front vehicle can be obtained.

[0061] As an optional embodiment, in step S1, the step S1 can also include: S14, superimposing the vehicle contour, the vehicle category, the vehicle orientation, the vehicle position, the vehicle speed, and the driving trajectory of the front vehicle in the display screen of the host vehicle.

[0062] Specifically, the internal of the automatic driving system generally includes a map for automatic driving, and the automatic driving system superimposes the vehicle category, the vehicle orientation, the vehicle position, the vehicle speed, and the driving trajectory of the front vehicle in the map, so that various vehicle information of the front vehicle is clearly displayed in the display screen of the host vehicle, facilitating the driver to obtain various vehicle information of the front vehicle, and improving the safety of vehicle driving.

[0063] In this embodiment, by determining that the obstacle in front of the host vehicle is the front vehicle using the preset deep learning model, and obtaining the predicted driving trajectory of the front vehicle, the driving trajectory and the influence range of the front vehicle can be known in advance, traffic accidents with the front vehicle can be avoided, and safety can be improved.

[0064] In one embodiment, in step S2, the three-point U-turn intention of the front vehicle is identified using a preset conflict logic model according to the actual driving trajectory of the front vehicle and the predicted driving trajectory of the front vehicle. Specifically, when the actual driving trajectory of the front vehicle is inconsistent with the predicted driving trajectory of the front vehicle, the three-point U-turn intention of the front vehicle is identified using a preset conflict logic model.

[0065] Specifically, generally, when the driving direction of the vehicle is blocked, if the road in the driving direction cannot be rotated in place or driven horizontally, the vehicle can only be adjusted in a space by driving forward and backward to change the pose of the vehicle, and then driven to the target position. This driving method is called three-point U-turn driving method.

[0066] Specifically, the vehicle in the three-point turn has the following three stages: in the first stage, the front vehicle drives forward and stops at a suitable position; in the second stage, the front vehicle reverses a turn; and in the third stage, the front vehicle continues to drive forward after reversing the turn. When the road is narrow, the above three stages may have a patrol situation, that is, when the third stage is executed, if it is predicted that the third stage will also be stuck and cannot get out, the first stage will be repeated, that is, the front vehicle will be rubbed back and forth to get out of the trouble.

[0067] For example, as shown in Figures 2 to 4 , the front vehicle drives forward, stops at the B point position after driving from the A point position, then reverses a turn to reach the C point position, and finally continues to drive forward.

[0068] For example, as shown in Figure 2 , the front vehicle is stuck in the same direction, and the three-point turn diagram for changing from reverse direction to the same direction is shown. In the first scenario, there are a first lane and a second lane, the first lane and the second lane are reverse lanes, and the front vehicle needs to turn around, that is, to drive from the first lane to the second lane. At this time, the host vehicle drives in the second lane, and the front vehicle drives in the first lane. The front vehicle drives forward in the first lane, turns at the A point position of the first lane, and stops at the B point position of the second lane, then reverses a turn to reach the C point position, and finally continues to drive forward in the second lane, that is, the front vehicle changes from reverse direction to the same direction as the host vehicle.

[0069] For example, as shown in Figure 3 , the front vehicle is stuck in the reverse direction, and the three-point turn diagram for changing from reverse direction to the same direction is shown. In the second scenario, there are a first lane and a second lane, the first lane and the second lane are reverse lanes, and the front vehicle needs to turn around, that is, to drive from the first lane to the second lane. At this time, the host vehicle drives in the second lane, and the front vehicle drives in the first lane. The front vehicle drives forward in the first lane, stops at the B point position of the first lane after driving from the A point position of the first lane, then reverses a turn to reach the C point position of the second lane, and finally continues to drive forward in the second lane, that is, the front vehicle changes from reverse direction to the same direction as the host vehicle.

[0070] For example, as shown in Figure 4As shown, the front car is stuck in reverse, and the three-point turn is shown in the figure. In the third scenario, there are a first lane and a second lane, the first lane and the second lane are reverse lanes, and the front car needs to turn around and drive from the first lane to the second lane. At this time, the host car and the front car are driving in the same direction on the second lane. The front car drives forward on the second lane, stops at the B point position of the first lane after driving from the A point position of the second lane, then reverses to the C point position, and then continues to drive forward on the first lane. At this time, the front car changes from driving in the same direction as the host car to driving in the opposite direction.

[0071] As described above, the automatic driving system can determine that the obstacle in front of the host car is a front car using a preset deep learning model, and can predict the driving trajectory of the front car in the future. Under normal circumstances, the front car drives according to the predicted driving trajectory. However, when the actual driving trajectory of the front car is not consistent with the predicted driving trajectory of the front car, that is, the front car does not drive according to the predicted driving trajectory when actually driving, for example, the front car is about to drive onto the roadbed, or directly drive towards the illegally parked vehicle on the road, it is obvious that the front car will not be able to drive or will collide with the obstacle in front of it if it drives according to the actual driving trajectory, or the front car is parked in a strange posture on a normal road (with a large conflict with the driving direction of the road), it can be determined that the front car has a high probability of three-point turn. Data mining is performed on the conflict behavior of the front car. If the vehicle in the data exhibits a three-point turn, the data that may be or is three-point turning is labeled to form labeled data, and if the vehicle in the data does not exhibit a three-point turn, the data is excluded. Then, the labeled data formed after data mining is trained using a preset conflict logic model. In this way, after a large amount of labeled data is sent to the preset conflict logic model for deep learning training, the front car's three-point turn intention can be identified using the preset conflict logic model when the automatic driving system obtains the actual driving trajectory of the front car that is not consistent with the predicted driving trajectory of the front car.

[0072] In this embodiment, the three-point turn intention of the front car is identified using a preset conflict logic model when the actual driving trajectory of the front car is not consistent with the predicted driving trajectory of the front car, so that the automatic driving system can respond in time when the front car has a three-point turn intention, take the correct behavior strategy to avoid in time, give the front car enough operation space, greatly reduce the contradiction with the front car, solve the escape scene, improve the traffic efficiency and safety, and make the automatic driving more intelligent.

[0073] In one embodiment, in the step S3, the target position to be driven by the front vehicle after the three-point turn is determined according to the three-point turn intention of the front vehicle, and the driving track of the front vehicle to the target position is determined. Specifically, it comprises:

[0074] S31, according to the three-point turn intention of the front vehicle, the target position to be driven by the front vehicle after the three-point turn is determined by using a preset target lane model.

[0075] Specifically, after obtaining the three-point turn intention of the front vehicle, the target lane position to be driven by the front vehicle is determined by the complete driving behavior of the front vehicle, and then the position point is marked on the target lane position to be driven by the front vehicle, and then the position point is trained by using the preset target lane model. In this way, after a large number of the above-mentioned position points are sent to the preset target lane model for deep learning training, the lane position to be driven by the front vehicle in the future can be inferred by the automatic driving system when the three-point turn intention of the front vehicle is obtained, so as to determine the target position to be driven by the front vehicle after the three-point turn.

[0076] For example, as described above, by Figures 2 to 4 It can be seen that when the front vehicle is in the three-point turn intention, the target position to be driven by the front vehicle after the three-point turn can be roughly inferred from the first stage of the front vehicle according to the preset target lane model. Specifically, in the first stage, the front vehicle drives forward and exists the behavior of deviating from the lane and parking after driving to the position in front of the host vehicle, and it can be roughly inferred that the target position to be driven by the front vehicle after the three-point turn is the target position in the same direction as the host vehicle. Or, the front vehicle drives forward and exists the behavior of deviating from the lane and parking after driving to the position in the opposite lane, and it can be roughly inferred that the target position to be driven by the front vehicle after the three-point turn is the target position in the opposite direction of the host vehicle.

[0077] But because the front vehicle needs a certain operation time when performing the three-point turn, when the host vehicle discovers that the front vehicle is in the three-point turn, it is generally in the second stage, and at this time, the target position to be driven by the front vehicle after the three-point turn is inferred from the driving track of the second stage of the front vehicle according to the preset target lane model. Specifically, in the second stage, after the front vehicle reverses, the road pointed to by the front of the front vehicle is the target position to be driven by the front vehicle after the three-point turn.

[0078] S32, path planning is performed according to the current vehicle position of the front vehicle and the target position of the front vehicle, and the driving track of the front vehicle to the target position is determined.

[0079] Specifically, common path planning algorithms such as RC curve can be used to rule the path planning of the front vehicle, and the driving track of the front vehicle to the target position given by the preset target lane model is determined, which will not be described here.

[0080] In the embodiment, the target position to be driven by the front vehicle after the three-point turning is determined according to the three-point turning intention of the front vehicle, and the driving track of the front vehicle to the target position is determined, so that the automatic driving system of the host vehicle can respond in time when the front vehicle has a three-point turning intention, take correct behavior strategy in time to avoid, give the front vehicle enough operation space, greatly reduce the contradiction with the front vehicle, solve the escape scene, improve the traffic efficiency and safety, and make the automatic driving more intelligent.

[0081] In one embodiment, in the step S4, the driving track of the front vehicle to the target position is used to control the host vehicle to drive according to a preset behavior strategy to avoid the front vehicle.

[0082] Since the front vehicle needs a certain operation time when performing three-point turning, it is basically difficult to complete all operations in a short time. When reversing, the driver's field of view has a large blind area to the rear (especially for large vehicles), and needs to carefully observe the surrounding operations and give enough tips to the surrounding to perform the reversing. Therefore, the host vehicle should make a decision in combination with the overall road conditions when facing this scene, and should stop at what position to give way or timely bypass. The host vehicle should determine how to go according to its future possible driving track.

[0083] In the step S4, the driving track of the front vehicle to the target position is used to control the host vehicle to drive according to a preset behavior strategy to avoid the front vehicle. Specifically, it includes:

[0084] S41, identifying whether the front vehicle is a parked vehicle.

[0085] Since in the first stage, the front vehicle will stop at a suitable position. The host vehicle needs to further identify whether the front vehicle is a parked vehicle that needs to be parked at the current position or will continue to the second stage after stopping. If the front vehicle does not have further action after stopping at the current position for a preset time, it can be judged that the front vehicle is a parked vehicle, at this time, the host vehicle selects bypassing to avoid the front vehicle; if the front vehicle has further action (such as performing the second stage) after stopping at the current position, it can be judged that the front vehicle is a turning vehicle.

[0086] S42, taking a preset behavior strategy to avoid according to the distance between the host vehicle and the front vehicle and the road width. Specifically, it includes:

[0087] For example, Figure 5When vehicle 1 is close to vehicle 2 and the road is wide enough for a detour, vehicle 1 determines whether to yield or detour based on vehicle 2's current phase. In this case, vehicle 1's current phase is irrelevant; the decision about whether to yield or detour is based solely on the current phase prediction.

[0088] When the distance between the main vehicle 1 and the preceding vehicle 2 is far, and the road width is insufficient for detour, the main vehicle 1 needs to stop behind the preceding vehicle 2 that is making a three-point U-turn. The main vehicle 1 stops and waits for the preceding vehicle 2 to complete the three-point U-turn before continuing to move forward. In other words, the main vehicle 1 needs to further consider the space required for the preceding vehicle 2 to execute the second and third stages, calculate the space and time required for the preceding vehicle 2 to complete all stages, and only move forward when it is sure that the preceding vehicle 2 has sufficient space and speed to complete the final third stage. For example, if Figure 6 As shown, when there is an obstacle in the middle of the road, it is necessary to further consider the space required by the front vehicle 2 to execute the second stage, calculate the space and time required for the front vehicle 2 to complete all stages, and only when it is sure that the front vehicle 2 has enough space and speed to complete the final third stage, the main vehicle 1 will move forward. For example, if Figure 7 As shown, when there are no obstacles in the middle of the road, it is necessary to further consider the position space required for the front vehicle 2 to execute the third stage, calculate the space and time required for the front vehicle 2 to complete all stages, and only when it is certain that the front vehicle 2 has enough position space and speed to complete the final third stage, the main vehicle 1 will move forward.

[0089] When the road is narrow (e.g., with a double yellow line on the left or a road boundary obstructing it) and there is insufficient safe space to maneuver around, the driver stops and waits, observing the trajectory of the vehicle 2 ahead that needs to complete a three-point turn. The driver then determines whether the braking position of the driver 1 will affect the space required for the vehicle 2 ahead to execute the three-point turn. If not, the driver continues to stop and wait for the vehicle 2 ahead to complete the three-point turn. If so, the driver reverses to a suitable position to clear the obstruction while ensuring the safety of the vehicle behind the driver 1. When the driver 1 reverses, the driver sets a distance safety threshold to determine whether the distance to the vehicles ahead and behind is too close. If the vehicle 2 ahead intrudes into the safety threshold distance of the driver 1, the driver needs to reverse to avoid it. At the same time, the driver needs to maintain a safe distance from the rear obstacle and the rear vehicle must give the driver enough space to continue reversing. If the vehicle 2 ahead is too close to the driver 1 and enters the dangerous distance within the safety threshold distance, the driver 1 will actively honk the horn to remind the vehicle 2 ahead to pay attention to the operation.

[0090] When the road is wide, based on the driving trajectory of the front vehicle to the target location, if the driving trajectory and impact range of the front vehicle 2 will block the passage of the main vehicle 1 in the future, the priority is to brake and wait, no longer moving forward, and wait until the front vehicle 2 completes a three-point turn before continuing to move forward.

[0091] In the embodiment, when it is determined that the front vehicle is going to make a three-point turn, a preset behavior strategy is adopted for avoidance according to the distance between the host vehicle and the front vehicle and the road width, so that the automatic driving system of the vehicle can timely respond when the front vehicle has a three-point turn intention, and timely adopt a correct behavior strategy for avoidance, thus giving the front vehicle enough operation space, greatly reducing the contradiction with the front vehicle, solving the escape scene, improving the traffic efficiency and safety, and making the automatic driving more intelligent.

[0092] Based on the same concept, in one embodiment, the present application also provides an automatic driving vehicle, such as Figure 8 As shown in the figure, the automatic driving vehicle 900 comprises a memory 902, a processor 901, and one or more computer programs stored in the memory 902 and movable on the processor 901, the memory 902 and the processor 901 are coupled together through a bus system 903, and the one or more computer programs are executed by the processor 901 to implement the following steps of a vehicle automatic control method provided by the embodiment of the present application:

[0093] S1, determining that the front obstacle of the host vehicle is a front vehicle by using a preset deep learning model, and obtaining a predicted driving track of the front vehicle, wherein the front vehicle is a vehicle closest to the host vehicle in front of the host vehicle;

[0094] S2, identifying a three-point turn intention of the front vehicle by using a preset conflict logic model according to an actual driving track of the front vehicle and the predicted driving track of the front vehicle;

[0095] S3, determining a target position to be driven by the front vehicle after the three-point turn of the front vehicle, and determining a driving track of the front vehicle to the target position according to the three-point turn intention of the front vehicle;

[0096] S4, controlling the host vehicle to drive according to a preset behavior strategy to avoid the front vehicle according to the driving track of the front vehicle to the target position.

[0097] The method disclosed by the embodiments of the present application can be applied to the processor 901 or implemented by the processor 901. The processor 901 can be an integrated circuit chip having a signal processing capability. In the implementation process, each step of the above method can be completed by an integrated logic circuit or an instruction in a software form of the processor 901. The processor 901 can be a general-purpose processor, a DSP, or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component, etc. The processor 901 can implement or execute the disclosed methods, steps and logic block diagrams in the embodiments of the present application. The general-purpose processor can be a microprocessor or any conventional processor, etc. In combination with the steps of the method disclosed in the embodiments of the present application, the hardware decoding processor can be directly embodied to complete the execution, or the hardware and software modules in the decoding processor can be combined to complete the execution. The software module can be located in a storage medium, and the storage medium is located in the memory 902. The processor 901 reads the information in the memory 902 and combines the hardware to complete the steps of the above method.

[0098] It is to be understood that the memory 902 of the embodiments of this application can be volatile or nonvolatile memory, or both.Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable read-only memory (EEPROM), a ferromagnetic random access memory (FRAM), a flash memory or other memory technology, a compact disc read-only memory (CD-ROM), a digital video disc (DVD) or other optical disc storage, a magnetic cassette, a magnetic tape, a magnetic disk storage or other magnetic storage device; the volatile memory can be a random access memory (RAM), and many forms of RAM can be used by way of example but not limitation, such as static random access memory (SRAM), synchronous static random access memory (SSRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), sync link dynamic random access memory (SLDRAM), direct rambus random access memory (DRRAM).The memory described in the embodiments of the present application is intended to include, but not limited to, these and any other suitable types of memory.

[0099] It should be noted that the above automatic driving vehicle embodiment and the method embodiment belong to the same concept, the specific implementation process is detailed in the method embodiment, and the technical features in the method embodiment are all applicable to the automatic driving vehicle embodiment, which will not be repeated here.

[0100] In addition, in the exemplary embodiments, the embodiments of the present application also provide a computer storage medium, specifically a computer readable storage medium, for example, including the memory 902 storing the computer program, one or more programs of a vehicle automatic control method are stored on the computer storage medium, and the one or more programs of the vehicle automatic control method are executed by the processor 901 to implement the following steps of a vehicle automatic control method provided by the embodiments of the present application:

[0101] S1, determining that the obstacle in front of the host vehicle is a preceding vehicle by using a preset deep learning model, and obtaining a predicted driving trajectory of the preceding vehicle, wherein the preceding vehicle is a vehicle closest to the host vehicle in front of the host vehicle;

[0102] S2, identifying a three-point U-turn intention of the preceding vehicle by using a preset conflict logic model according to an actual driving trajectory of the preceding vehicle and the predicted driving trajectory of the preceding vehicle;

[0103] S3, determining a target position to be driven by the preceding vehicle after the three-point U-turn of the preceding vehicle, and determining a driving trajectory of the preceding vehicle to the target position according to the three-point U-turn intention of the preceding vehicle;

[0104] S4, controlling the host vehicle to drive according to a preset behavior strategy to avoid the preceding vehicle according to the driving trajectory of the preceding vehicle to the target position.

[0105] It should be noted that the above vehicle automatic control method program embodiment on the computer readable storage medium and the method embodiment belong to the same concept, the specific implementation process is detailed in the method embodiment, and the technical features in the method embodiment are all applicable to the above computer readable storage medium embodiment, which will not be repeated here.

[0106] It should be noted that, in the present text, the terms "comprises", "comprising", or any other variant thereof, are intended to cover the non-exclusive inclusion, such that processes, methods, articles, or apparatuses that comprise a list of elements are not limited to those elements, but can also include other elements not expressly listed, or even other elements inherent to such processes, methods, articles, or apparatuses. Without further limitation, an element defined by the phrase "comprising a" does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0107] The above examples are only used to illustrate the technical solutions of the present application, and are not intended to limit the present application; under the idea of the present application, the technical features in the above examples or different examples can also be combined, the steps can be implemented in any order, and there are many other changes of different aspects of the present application as described above, which are not provided in details for simplicity; although the present application has been described in detail with reference to the foregoing examples, those skilled in the art should understand that the technical solutions recorded in the foregoing examples can still be modified, or some technical features therein can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A vehicle automatic control method characterized by comprising: The vehicle automatic control method comprises: determining that the obstacle in front of the host vehicle is a preceding vehicle by using a preset deep learning model, and obtaining a predicted driving track of the preceding vehicle; identifying a three-point turn intention of the preceding vehicle by using a preset conflict logic model according to an actual driving track of the preceding vehicle and the predicted driving track of the preceding vehicle; judging a target position to be driven by the preceding vehicle after the three-point turn of the preceding vehicle according to the three-point turn intention of the preceding vehicle, and determining a driving track of the preceding vehicle to the target position; controlling the host vehicle to drive according to a preset behavior strategy to avoid the preceding vehicle according to the driving track of the preceding vehicle to the target position.

2. The vehicle automatic control method according to claim 1, characterized by, The determination of the obstacle in front of the host vehicle as the preceding vehicle and the determination of the driving track of the preceding vehicle by using the preset deep learning model comprise: controlling a sensor of the host vehicle to capture point clouds of the obstacle in front of the host vehicle in real time; after obtaining the point clouds of the obstacle in front of the host vehicle, determining whether the point clouds of the obstacle in front of the host vehicle are a vehicle contour and a vehicle orientation by using the preset deep learning model, and calculating a vehicle speed according to a change in movement of the vehicle orientation and the vehicle position according to the vehicle contour; obtaining the predicted driving track of the preceding vehicle according to the vehicle orientation, the vehicle position and the vehicle speed.

3. The vehicle automatic control method according to claim 2, characterized by, The determination of the obstacle in front of the host vehicle as the preceding vehicle and the determination of the driving track of the preceding vehicle by using the preset deep learning model further comprise: superimposing the vehicle contour, the vehicle orientation, the vehicle position, the vehicle speed and the driving track of the preceding vehicle in the display screen of the host vehicle.

4. The vehicle automatic control method according to claim 2, characterized by, The identification of the three-point turn intention of the preceding vehicle by using the preset conflict logic model according to the actual driving track of the preceding vehicle and the predicted driving track of the preceding vehicle comprises: when the actual driving track of the preceding vehicle is inconsistent with the predicted driving track of the preceding vehicle, identifying the three-point turn intention of the preceding vehicle by using the preset conflict logic model.

5. The vehicle automatic control method according to claim 1, characterized by, The judgment of the target position to be driven by the preceding vehicle after the three-point turn of the preceding vehicle according to the three-point turn intention of the preceding vehicle, and the determination of the driving track of the preceding vehicle to the target position comprise: judging the target position to be driven by the preceding vehicle after the three-point turn of the preceding vehicle by using a preset target lane model according to the three-point turn intention of the preceding vehicle; planning a path according to the current vehicle position of the preceding vehicle and the target position of the preceding vehicle, and determining the driving track of the preceding vehicle to the target position.

6. The vehicle automatic control method according to claim 5, characterized by The control of the host vehicle to drive according to the preset behavior strategy to avoid the preceding vehicle according to the driving track of the preceding vehicle to the target position comprises: identifying whether the preceding vehicle is a parked vehicle; controlling the host vehicle to drive according to the preset behavior strategy to avoid the preceding vehicle according to the distance between the host vehicle and the preceding vehicle and the road width.

7. The vehicle automatic control method according to claim 6, characterized by, The identification of whether the preceding vehicle is the parked vehicle comprises: if the preceding vehicle has no further action after parking at a current position for a preset time, judging that the preceding vehicle is the parked vehicle; if the preceding vehicle has further action after parking at the current position, judging that the preceding vehicle is a three-point turn vehicle.

8. The vehicle automatic control method according to claim 6, characterized by, The method comprises the following steps: When the distance between the host vehicle and the front vehicle is short and there is space to bypass on the road, it is determined whether to yield or bypass directly according to the current stage of the front vehicle; When the distance between the host vehicle and the front vehicle is far, if the road width is not enough to bypass and the host vehicle needs to stop behind the front vehicle, the host vehicle stops and waits until the front vehicle completes a three-point turn before continuing to move forward; When the road is narrow and there is not enough safe space to bypass, the host vehicle stops and waits, observes the entire trajectory of the three-point turn of the front vehicle, and further determines whether the stopping position of the host vehicle will affect the space required by the front vehicle to perform the three-point turn; When the road is wide, according to the driving trajectory of the front vehicle to the target position, if the driving trajectory of the front vehicle will block the host vehicle in the future, the host vehicle stops and waits until the front vehicle completes a three-point turn before continuing to move forward.

9. An autonomous vehicle, comprising: The method comprises the following steps: A memory, a processor, and a computer program stored on the memory and running on the processor, wherein the computer program is executed by the processor to implement the vehicle automatic control method according to any one of claims 1 to 8.

10. A computer storage medium, characterized in that, The computer storage medium stores a vehicle automatic control method program, and the vehicle automatic control method program is executed by the processor to implement the vehicle automatic control method according to any one of claims 1 to 8.

Images