Method and device for controlling vehicle

Abstract

The embodiment of the present invention discloses a method and device for controlling a vehicle. The method includes: when the front view of the vehicle is blocked and there is an intersection or a crosswalk behind the blocker, based on the blockage rate, the state of the vehicle, the position of the blocker, and the position of the intersection or the crosswalk, evaluating the risk of collision with a moving object when there is a blocked moving object at the intersection or the crosswalk; and controlling the vehicle based on the evaluated risk. The embodiment of the present invention can reduce the problem of reduced safety due to the blockage of the vehicle's field of view.

Description

Technical Field

[0001] The present invention relates to automatic driving technology, and in particular to a method and device for controlling a vehicle. Background Art

[0002] With the continuous development of autonomous driving technology, driving safety systems such as autonomous emergency braking (AEB) have become an important consideration for people to choose and assemble autonomous driving vehicles.

[0003] AEB is an auxiliary system that improves driving safety through technical means. It uses the radar or camera at the front of the vehicle to scan and detect other vehicles, pedestrians, obstacles, etc. on the road in real time. And according to the distance and relative speed of the detected objects, the algorithm calculates the potential collision risk. When there is a potential collision risk, the driver will be warned through warning lights, alarm sounds, etc. And when the driver does not take braking measures after receiving the warning, or when it is judged that the driver does not have time to brake, the AEB system will intervene forcibly and automatically brake the vehicle to reduce the probability of collision or reduce the collision effect.

[0004] The AEB system can effectively reduce the incidence of collision accidents, protect the safety of pedestrians and drivers, and help improve the overall safety of road traffic. However, in some special scenarios, the capabilities of the AEB system will be limited. For example, in scenarios such as "ghosting" and intersection obstruction (Straight Crossing Path with Obstruction, SCPO), the vehicle's field of view is blocked, and when the sensor detects the target, the vehicle may already be unable to avoid a collision. Summary of the invention

[0005] A method and device for controlling a vehicle according to an embodiment of the present invention can reduce the problem of reduced safety due to obstruction of the field of vision of the vehicle.

[0006] A method for controlling a vehicle in an embodiment of the present invention includes: when the field of view in front of the vehicle is blocked and there is an intersection or a crosswalk behind the obstruction: based on the obstruction rate, the state of the vehicle, the position of the obstruction, and the position of the intersection or the crosswalk, evaluating the risk of collision with the moving object assuming that there is an obstructed moving object at the intersection or the crosswalk; and controlling the vehicle based on the evaluated risk.

[0007] In some implementations, the state of the vehicle includes: the speed of the vehicle and the maximum braking acceleration of the vehicle; and

[0008] The risk is assessed by a risk ratio and is based on calculating the risk rate;

[0009] Among them, risk is the risk rate, and when vc 2 When -(pi-po)*2a<=0, risk is 0; k is the prior probability, ro is the occlusion rate, vc is the speed of the vehicle, po is the position of the occlusion, pi is the position of the intersection or crosswalk, and a is the maximum braking acceleration of the vehicle.

[0010] In some embodiments, controlling the vehicle based on the assessed risk includes:

[0011] When risks are assessed, multiple driving strategies are generated based on current environmental information and vehicle status information, as well as future scenarios previewed under each driving strategy;

[0012] Evaluate the future scenarios previewed under each of the driving strategies; and

[0013] Based on the evaluation result of the previewed future scenario, an optimal driving strategy for controlling the ego-vehicle is determined.

[0014] In some implementations, a deep learning network is used to generate multiple driving strategies and future scenarios previewed under each driving strategy based on the current environment information and vehicle state information.

[0015] In some embodiments, the method further comprises:

[0016] Determine whether there is an obstruction in the field of view in front of the vehicle;

[0017] When the obstruction exists, determining a magnitude relationship between the obstruction rate and a preset first threshold;

[0018] When the occlusion rate is greater than the first threshold, performing risk prevention control;

[0019] When the shielding rate is less than the first threshold, determining whether there is an intersection or a crosswalk behind the shielding object; and

[0020] When there is an intersection or crosswalk, perform actions to assess risk.

[0021] In some implementations, controlling the vehicle based on the assessed risk includes: performing risk prevention control on the vehicle when the risk rate is greater than a preset second threshold.

[0022] In some embodiments, the method further includes: detecting occluders based on a detection scheme based on deep learning and a bird's-eye view, and calculating the occlusion rate.

[0023] A computer device according to an embodiment of the present invention includes a memory, a processor, and a computer program / instruction stored in the memory, wherein the processor executes the computer program / instruction to implement the steps of the method according to the embodiment of the present invention.

[0024] A computer-readable storage medium according to an embodiment of the present invention stores a computer program / instruction thereon, and when the computer program / instruction is executed by a processor, the steps of the method according to the embodiment of the present invention are implemented.

[0025] A computer program product according to an embodiment of the present invention includes a computer program / instruction, and when the computer program / instruction is executed by a processor, the steps of the method according to the embodiment of the present invention are implemented.

[0026] Beneficial effects of the embodiments of the present invention:

[0027] When there is an obstruction in the field of view in front of the vehicle, and there is an intersection or crosswalk behind the obstruction, the risk of collision with the moving object when there is an obstructed moving object at the intersection or crosswalk is evaluated based on the obstruction rate, the state of the vehicle, the position of the obstruction, and the position of the intersection or crosswalk. And based on the evaluation results, the vehicle is controlled, such as warning the driver, reducing the speed of the vehicle, etc. In the above way, the risk of collision can be reduced and safety can be improved for scenes where the vehicle's field of view is limited, such as "ghosting" or passing through intersections. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Other details and advantages of the present invention will become apparent from the detailed description provided below. It should be understood that the following drawings are merely illustrative and thus cannot be considered as limiting the present invention, and the following will be described in detail with reference to the drawings, wherein:

[0029] Figure 1 is a flow chart of an embodiment of a method for controlling a vehicle of the present invention;

[0030] Figure 2 is a schematic diagram of a process of analyzing an image and data of a radar sensor to obtain a detection result according to an embodiment of the present invention;

[0031] Figure 3 is a schematic diagram of a process of determining an optimal driving strategy based on a deep learning network according to an embodiment of the present invention;

[0032] Figure 4 is a flow chart of another embodiment of a method for controlling a vehicle of the present invention;

[0033] Figure 5 is a bird's-eye view of an application scenario of an embodiment of the present invention;

[0034] Figure 6It is a structural schematic diagram of a computer device according to an embodiment of the present invention that can implement the vehicle control method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0035] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention clearer and more understandable, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only for explaining the present invention and are not intended to limit the present invention.

[0036] In the description of the present invention, it is to be understood that the terms "first" and "second" are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. Moreover, the terms "first", "second", etc. are applicable to distinguishing similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the terms used in this way are interchangeable where appropriate so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein.

[0037] like Figure 1 Shown is a flow chart of an embodiment of a method for controlling a vehicle of the present invention.

[0038] In some embodiments, the method can be used in a vehicle equipped with an autonomous driving function, such as a vehicle equipped with L2 to L4 autonomous driving functions. Specifically, the method can be integrated into the AEB (automatic emergency braking) function to increase the safety of the AEB function. In other embodiments, the method can be executed by a controller responsible for the autonomous driving function, such as an autonomous driving domain controller.

[0039] like Figure 1 As shown, the method for controlling a vehicle specifically includes:

[0040] Step S10: There is an obstruction in the front field of view of the vehicle and there is an intersection or a crosswalk behind the obstruction.

[0041] In step S10, based on data collected by a radar and / or image sensor (such as a camera) mounted in front of the vehicle, it can be used to identify whether the front view is blocked and there is an intersection or a crosswalk behind the blocker. Figure 5 That is, it shows such a scene. Figure 5 As shown, there is an obstacle 51 in the forward direction of the vehicle 50, and there is an intersection or a crosswalk 52 behind the obstacle.

[0042] In some embodiments, a multi-sensor fusion solution based on deep learning bird's-eye view (BEV) can be used to detect intersections, vehicles, pedestrians, and obstructions, etc., to identify whether the vehicle is in the scene shown in step S10. In addition, the detected data can also be used in the risk assessment in step S12 and the selection of the vehicle control strategy in step S14.

[0043] Specifically, Figure 2 As shown in FIG. 1 , it is a flowchart of a solution for environmental detection using a multi-sensor fusion solution based on deep learning bird's eye view (BEV). Figure 2 In the method, the image data collected by the image sensor and the radar data collected by the radar are respectively input into CNN (Convolutional Neural Networks) and MLP (Multilayer Perceptron) for processing, and the image features and radar features are respectively obtained. Then the image features and radar features are input into the LSS (Lift, Splat, Shoot) algorithm, so that the image features and radar features are converted into BEV (Bird's Eye View) features by the LSS algorithm, and finally processed by CNN to output the detection results. Among them, the detection results may include but are not limited to: the position, shape and speed of vehicles and pedestrians; the position and size of the obstruction; whether there is an intersection and the intersection position; and whether there is a crosswalk and the crosswalk position. Based on the detection results, it can be determined whether there is an obstruction in the field of view in front of the vehicle, and whether there is an intersection or a crosswalk behind the obstruction. For example, it can be determined whether there is an intersection or a crosswalk behind the obstruction based on the relative position of the obstruction and the intersection or crosswalk.

[0044] by Figure 5 For example, based on the above detection results, the obstruction 51 and the intersection or crosswalk 52 within the field of view 53 can be identified, so it can be determined that there is an obstruction in the field of view in front of the vehicle, and there is an intersection or crosswalk 52 in the field of view. In addition, based on the above detection results, the position of the obstruction 51 and the position of the intersection or crosswalk 52 can also be obtained, and based on the relative positions of the two, it can be determined that the intersection or crosswalk 52 is behind the obstruction 51. Therefore, based on the detection results, it can be determined that the vehicle is in the scene described in step S10.

[0045] In similar Figure 5In the scenario, when the vehicle 50 is moving forward, if a moving object passes through the crosswalk or intersection 52, the AEB function may not detect the moving object and brake in time because the radar or image sensor's field of view is blocked by the obstruction 51, resulting in a dangerous situation where the vehicle and the moving object collide. In this embodiment, when it is recognized that the vehicle is in such a scenario, a risk assessment can be performed based on the method of step S12, and based on the result of the risk assessment, the vehicle can be controlled to improve the safety problem caused by the obstruction of the field of view.

[0046] Step S12: Assuming that there is an obstructed moving object at the intersection or on the crosswalk, the risk of collision with the moving object is evaluated.

[0047] In some implementations, the risk may be assessed based on the occlusion rate, the state of the vehicle, the location of the occlusion object, and the location of the intersection or crosswalk.

[0048] The occlusion rate can be obtained by projecting the vehicle's field of view onto the BEV plane and then dividing the projected area of ​​the occluded area by the total area. Figure 5 As shown, the area of ​​region 54 (ie, the size of the blocked visual field area) divided by the area of ​​visual field 53 is the blocking rate.

[0049] The state of the vehicle may include: the speed of the vehicle and the maximum braking acceleration of the vehicle. The speed of the vehicle may be measured by a speed sensor mounted on the vehicle, and the maximum braking acceleration of the vehicle may be pre-stored in the vehicle.

[0050] The position of the obstruction and the position of the intersection or crosswalk refer to the distance between the vehicle and the obstruction and the position of the vehicle and the intersection or crosswalk when the line is drawn along the forward direction of the vehicle. The position of the obstruction and the position of the intersection or crosswalk can be obtained from the previous detection results.

[0051] Generally speaking, the greater the occlusion rate (indicating a larger blind spot in the field of vision), the faster the vehicle's speed, the smaller the maximum braking acceleration, and the closer the position, the greater the possibility of risk.

[0052] In some implementations, step S10 may quantify the risk based on the following formula (1):

[0053]

[0054] Among them, risk is the risk rate, and when vc 2 When -(pi-po)*2a<=0, risk is 0; k is the prior probability, ro is the occlusion rate, vc is the speed of the vehicle, po is the position of the occlusion, pi is the position of the intersection or crosswalk, and a is the maximum braking acceleration of the vehicle.

[0055] In this risk rate calculation formula, parameters such as the obstruction rate, the speed of the vehicle, the maximum braking acceleration of the vehicle, the location of the obstruction, and the location of the intersection or crosswalk are comprehensively considered, so the risk can be assessed more accurately.

[0056] In some implementations, after the risk rate is obtained, the risk may be divided into different levels, such as high risk, medium risk, and low risk, based on the relationship between the risk rate and the threshold.

[0057] Step S14: Based on the risk evaluated in step S12, control the vehicle.

[0058] In some embodiments, different risk levels may correspond to different self-vehicle control strategies. For example, low risk may not perform any additional operations. Medium risk may perform operations to warn the driver, such as prompting the driver to be careful when passing through intersections or crosswalks through sound, light, vibration, etc. High risk may perform operations such as warning and actively reducing the speed of the vehicle. For example, if the user does not brake after the warning, actively brake to reduce the speed of the vehicle.

[0059] In other implementations, the optimal control (driving) strategy for the vehicle may be determined based on a deep learning network.

[0060] like Figure 3 As shown, the input of the deep learning network includes: environmental signals, vehicle status signals, risk decision algorithms and risk situations. Among them, environmental signals can be provided by environmental sensors such as image sensors and radars, which may include: obstacles in the environment, lane lines, pedestrians, other vehicles, speeds, and the like. Examples of vehicle status signals may include: wheel speed signals, steering wheel signals, and the like. The risk situation is derived from the output of step S12, and may be, for example, indication information of different risk levels, indication information of whether there is a risk, or a risk rate. In some embodiments, when the risk situation indicates that there is no risk or a low risk, the deep learning network is not started, that is, the best driving strategy is not determined, but the current driving plan is maintained.

[0061] Based on these input signals, the deep learning network outputs multiple driving strategies (i.e., vehicle control signals) and future scenarios previewed under each driving strategy. In some embodiments, the deep learning network can determine multiple driving strategies and preview future scenarios based on environmental data, vehicle status data, and risk rate. Then, the previewed future scenarios and driving strategies are input into the scenario evaluation module, and the previewed future scenarios are evaluated. For example, a scoring algorithm can be used to score the previewed future scenarios, and the best driving strategy is determined based on the scoring results.

[0062] Specifically, the scoring algorithm can evaluate the previewed future scenarios based on the following perspectives: driving safety under different driving strategies, the position and relative speed of the vehicle with respect to other objects in the future scenario, the vehicle's own speed and driving direction, the complexity of the control signal operation, and the impact on the driving experience. Among them, the complexity of the control signal can be measured by the number, frequency, and amplitude of the controlled objects. The impact on the driving experience can be measured by the acceleration size and turning radius, etc.

[0063] In an embodiment of the present invention, when there is an obstruction in the field of view in front of the vehicle, and there is an intersection or a crosswalk behind the obstruction, the risk of collision with the moving object when there is an obstructed moving object at the intersection or the crosswalk is evaluated based on the obstruction rate, the state of the vehicle, the position of the obstruction, and the position of the intersection or the crosswalk. And based on the evaluation result, the vehicle is controlled, for example, the driver is warned, the speed of the vehicle is reduced, etc. In the above manner, the risk of collision can be reduced and safety can be improved for scenes where the field of view of the vehicle is limited, such as "ghosting" or passing through intersections.

[0064] In addition, the embodiment of the present invention provides a method for evaluating risk based on parameters such as occlusion rate, such as formula (1). This method for evaluating risk considers comprehensive parameters and can improve the accuracy of the evaluated risk.

[0065] In addition, when evaluating risks, the embodiments of the present invention can determine the best driving strategy based on a deep learning network, thereby achieving a balance between risk control and driving experience.

[0066] The following combination Figure 4 , a specific example of an embodiment of the present invention is described.

[0067] like Figure 4As shown, based on the data of the radar and / or image sensor, the surrounding environment is detected to obtain the detection result (step S40). Then, based on the detection result, it is determined whether there is any obstruction in the front field of view (step S41). If there is no obstruction, there is no risk to be concerned about in this embodiment, and it continues to detect whether there is any obstruction. If there is obstruction, it is first determined whether the obstruction rate is greater than the preset threshold value th1 (step S42). If the obstruction rate is greater than the threshold value th1, it can be directly considered that there is a risk, and operations related to the risk are performed, such as early warning or controlling the vehicle to slow down (step S45). If the obstruction rate is less than the threshold value th1, it is continued to determine whether there is an intersection or a crosswalk behind the obstruction (step S43). If there is no intersection or crosswalk, it can be considered that the risk is low and no additional control is performed. If there is an intersection or crosswalk, it is determined whether the risk rate is greater than the preset threshold value th2 (step S44). If the risk rate is less than the threshold value th2, it can be considered that the risk is low, and the current control can be maintained without performing additional control operations. If the risk rate is greater than th2, risk-related operations may be performed, such as early warning, controlling the vehicle to slow down, or even controlling the vehicle to stop (step S44).

[0068] In this embodiment, the occlusion rate is used as a guide to evaluate the risk of a vehicle passing through an obstructed intersection or crosswalk, and the vehicle is controlled based on the evaluation result to improve driving safety.

[0069] like Figure 6 , is a schematic diagram of the structure of an embodiment of a computer device 6 of the present invention, which includes a memory 60, a processor 62 and a computer program / instruction stored in the memory 60, and the processor 62 executes the computer program / instruction to implement the method described in the embodiment of the present invention.

[0070] In this embodiment, the computer device 6 may be, for example, a controller responsible for the automatic driving function, such as a central computing unit, a domain controller, a regional controller or other ECU (electronic unit).

[0071] The embodiment of the present invention further provides a computer-readable storage medium on which a computer program / instruction is stored. When the computer program / instruction is executed by a processor, the method described in the embodiment of the present invention is implemented.

[0072] The embodiment of the present invention further provides a computer program product, including a computer program / instruction, which implements the method described in the embodiment of the present invention when executed by a processor.

[0073] The description of the above device, storage medium and program product embodiments is similar to the description of the above method embodiment, and has similar beneficial effects as the method embodiment. For technical details not disclosed in the device, storage medium and program product embodiments of this application, please refer to the description of the method embodiment of this application for understanding.

[0074] The processor may be at least one of an application specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field programmable gate array (FPGA), a central processing unit (CPU), a controller, a microcontroller, a microprocessor, etc. It is understandable that the electronic device that implements the function of the processor may also be other, and the embodiments of the present application are not specifically limited.

[0075] The above-mentioned computer storage medium / memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a magnetic random access memory (FRAM), a flash memory (Flash Memory), a magnetic surface memory, an optical disk, or a compact disc read-only memory (CD-ROM) and the like; it can also be various terminals including one or any combination of the above-mentioned memories, such as mobile phones, computers, tablet devices, personal digital assistants, etc.

[0076] It should be noted that the above description is only an example and not a limitation of the present invention. In other embodiments of the present invention, the method may have more, fewer, or different steps, and the relationships such as the order, inclusion, and functions between the steps may be different from those described and illustrated. For example, multiple steps can usually be combined into a single step, and a single step can also be split into multiple steps. For those of ordinary skill in the art, without creative efforts, the changes in the sequence of each step are also within the protection scope of the present invention.

[0077] Essentially, the technical solution of the present invention, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.), a processor, or a microcontroller to execute all or part of the steps of the methods described in various embodiments of the present invention.

[0078] Those of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps including the above method embodiments.

[0079] Although the present invention has been disclosed above with preferred embodiments, the present invention is not limited thereto. Any changes and modifications made by those skilled in the art within the spirit and scope of the present invention should be incorporated into the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the scope defined by the claims.

Claims

1. A method for controlling a vehicle, characterized in that: include: When the vehicle's front view is blocked and there is an intersection or crosswalk behind the obstruction: Based on the occlusion rate, the state of the vehicle, the position of the occluding object, and the position of the intersection or the crosswalk, assessing the risk of collision with the moving object when the occluded moving object is present at the intersection or the crosswalk; as well as Based on the assessed risk, controlling the ego-vehicle; The state of the vehicle includes: the speed of the vehicle and the maximum braking acceleration of the vehicle; The risk is assessed by a risk ratio, and based on calculating the risk rate; Among them, risk is the risk rate, and when vc 2 When -(pi-po)*2a<=0, risk is 0; k is the prior probability, ro is the occlusion rate, vc is the speed of the vehicle, po is the position of the occlusion, pi is the position of the intersection or crosswalk, and a is the maximum braking acceleration of the vehicle.

2. The method for controlling a vehicle according to claim 1, characterized in that: The controlling of the vehicle based on the assessed risk includes: When risks are assessed, multiple driving strategies are generated based on current environmental information and vehicle status information, as well as future scenarios previewed under each driving strategy; Evaluate the future scenarios previewed under each of the driving strategies; and Based on the evaluation result of the previewed future scenario, an optimal driving strategy for controlling the vehicle is determined.

3. The method for controlling a vehicle according to claim 2, characterized in that: A deep learning network is used to generate multiple driving strategies and preview future scenarios under each driving strategy based on the current environment information and vehicle status information.

4. The method for controlling a vehicle according to claim 1, characterized in that: The method further comprises: Determine whether there is an obstruction in the field of view in front of the vehicle; When the obstruction exists, determining a magnitude relationship between the obstruction rate and a preset first threshold; When the occlusion rate is greater than the first threshold, performing risk prevention control; When the occlusion rate is less than the first threshold, determining whether there is an intersection or a crosswalk behind the obstruction; When there is an intersection or crosswalk, perform actions to assess risk.

5. The method for controlling a vehicle according to claim 1, characterized in that: The controlling of the vehicle based on the assessed risk includes: When the risk rate is greater than a preset second threshold, risk prevention control is performed on the vehicle.

6. The method for controlling a vehicle according to claim 1, characterized in that: The method further comprises: A detection scheme based on deep learning and a bird's-eye view is used to detect occluders and calculate the occlusion rate.

7. A computer device comprising a memory, a processor and a computer program / instruction stored in the memory, characterized in that: The processor executes the computer program / instructions to implement the steps of the method according to any one of claims 1 to 6.

8. A computer-readable storage medium having a computer program / instruction stored thereon, characterized in that: When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images