A method for emergency decision-making in autonomous driving

Abstract

The present invention relates to an automatic driving emergency decision-making method, including: collecting vehicle surrounding environment information, extracting potential risk targets based on the environment information; establishing a risk field model for intelligent networked vehicle drivers, and calculating the perceived risk index of the lane where the vehicle is traveling and the perceived risk index of the adjacent lane in combination with the potential risk targets; setting a perceived risk index threshold based on the driver's real driving data, comparing the perceived risk index of the lane where the vehicle is traveling and the perceived risk index of the adjacent lane with the perceived risk index threshold, obtaining the automatic driving emergency decision of the intelligent networked vehicle, and performing an automatic driving emergency response based on the automatic driving emergency decision of the intelligent networked vehicle. The present invention uses the driver risk field model to humanize the automatic driving decision, so as to be more in line with the driving habits of human drivers.

Description

Technical Field

[0001] The present invention relates to the fields of intelligent networked vehicles and autonomous driving technology, and in particular to an autonomous driving emergency decision-making method. Background Art

[0002] Accurate and timely perception of environmental information on the driving path is the basic guarantee for the driving safety of autonomous vehicles. Traditional autonomous driving solutions are single-vehicle intelligence based on on-board sensors to perceive the environment. However, when driving, it is difficult to respond to various emergencies due to the limitations of its own perception and decision-making (such as blocked vision). Due to reasons such as obstructed vision in rainy and foggy days and blocked vision by vehicles in front, vehicles cannot directly perceive emergency scenes, such as emergency braking of the pilot car in the convoy, which makes it difficult for vehicles to detect and predict actual road conditions in time, which is very likely to cause serious traffic accidents and irreparable economic losses. Intelligent connected vehicle technology can enable instant communication between vehicles (V2V, vehicle-to-vehicle), vehicles and infrastructure (V2I, vehicle-to-infrastructure), vehicles and networks (V2N, vehicle-to-network), and vehicles and pedestrians (V2P, vehicle-to-pedestrian), greatly increasing the amount of information obtained by vehicles and drivers, not only reducing the difficulty of intelligent driving, but also improving driving safety. Therefore, by combining the intelligent perception capabilities of smart connected vehicles and the V2X vehicle network information sharing capabilities, an information basis is provided for risk perception during vehicle driving and emergency decision-making for autonomous driving, which is of great significance for improving traffic safety. Summary of the invention

[0003] The purpose of the present invention is to address the deficiencies in the above-mentioned prior art, and to provide an autonomous driving emergency decision-making method that combines the risk field model of intelligent connected vehicle drivers, the on-board intelligent perception of intelligent connected vehicles, and the V2X vehicle network information sharing capabilities, so as to ensure driving safety while improving driving comfort through human-like driving decisions.

[0004] To achieve the above object, the present invention provides the following solutions:

[0005] An automatic driving emergency decision-making method, comprising:

[0006] Collecting environmental information about the vehicle, and extracting potential risk targets based on the environmental information;

[0007] Establishing a risk field model for intelligent connected vehicle drivers, and calculating the perceived risk index of the lane in which the vehicle is traveling and the perceived risk index of the adjacent lanes in combination with the potential risk target;

[0008] Based on the driver's actual driving data, a perception risk index threshold is set, the perception risk index of the lane in which the vehicle is traveling and the perception risk index of the adjacent lane are compared with the perception risk index threshold, the automatic driving emergency decision of the intelligent connected vehicle is obtained, and an automatic driving emergency response is performed based on the automatic driving emergency decision of the intelligent connected vehicle.

[0009] Preferably, extracting the potential risk target includes:

[0010] The vehicle-mounted intelligent sensor device of the intelligent networked vehicle is combined to perform intelligent perception, and information is shared with the V2X vehicle network to obtain the vehicle surrounding environment information, wherein the vehicle surrounding environment information includes the environmental information of the unobstructed area around the vehicle and the environmental information of the obstructed area.

[0011] Preferably, the environmental information of the unobstructed area around the vehicle includes: potential risk targets in front of the lane; wherein the potential risk targets include pedestrians, vehicles, non-motor vehicles and obstacles.

[0012] Preferably, obtaining the environmental information of the blocked area includes:

[0013] Information is shared through the V2X vehicle network to obtain environmental information of the obstructed area, and the vehicle-to-vehicle communication of the intelligent connected vehicle is used to obtain the speed and relative position of each vehicle in the fleet in front of the lane, the speed and relative position of the rear vehicle, and the speed and relative position of the vehicle in the adjacent lane; the speed and relative position of pedestrians in the lane are obtained based on the vehicle-to-pedestrian communication of the intelligent connected vehicle; and the information of non-motor vehicles, obstacles, traffic lights, and traffic signs is obtained based on the vehicle-to-infrastructure communication of the intelligent connected vehicle.

[0014] Preferably, the environmental information obtained by intelligent perception by the on-board intelligent sensing device of the intelligent connected vehicle and information sharing by the V2X vehicle network is fused to obtain fused environmental information; based on the fused environmental information, the coordinate distance of the potential risk target and the speed relative to the vehicle are obtained, and the expected collision time of the vehicle and the potential risk target is calculated to obtain the perceived risk index.

[0015] Preferably, calculating the perceived risk index of the lane in which the vehicle is traveling and the perceived risk index of the adjacent lanes includes: taking the position of the geometric center of the vehicle as the coordinate origin, calculating the relative position (D x ,D y ) and relative speed V; Calculate the relative position of the nearest potential risk target in front of the adjacent lane of the vehicle Relative speed V nbr, based on the intelligent connected vehicle driver risk field model, calculate the perceived risk index R of the current lane d , the perceived risk index R of the adjacent lane nbr .

[0016] Preferably, the perceived risk index R of the lane where the vehicle is currently located is calculated based on the intelligent connected vehicle driver risk field model. d The method is:

[0017]

[0018] Among them, D x is the lateral position of the potential risk target in the lane where the vehicle is located relative to the vehicle's driving direction in the vehicle reference system, D y is the longitudinal position of the potential risk target in the lane where the vehicle is located relative to the vehicle's driving direction in the vehicle reference system, V is the relative speed of the potential risk target in the lane where the vehicle is located relative to the vehicle in the driving direction; t 1 ,t 2 ,t 3 ,t 4 is the model parameter and e is the natural exponent.

[0019] Preferably, the perceived risk index R of the adjacent lane is calculated based on the intelligent connected vehicle driver risk field model. nbr The method is:

[0020]

[0021] in, is the lateral position of the potential risk target in the adjacent lane of the vehicle relative to the vehicle's driving direction in the vehicle reference system, V is the longitudinal position of the potential risk target in the adjacent lane of the vehicle relative to the vehicle's driving direction in the vehicle reference system. nbr is the relative speed of the potential risk target in the adjacent lane to the vehicle in the driving direction; 1 ,t 2 ,t 3 ,t 4 is the model parameter and e is the natural exponent.

[0022] Preferably, the setting of the perceived risk index threshold based on the driver's actual driving data includes:

[0023] The real driving data of several drivers without accidents in any time period are collected, and the changes of the braking deceleration of the vehicle and the risk index of the lane where the vehicle is located are recorded. Based on the multivariate nonlinear regression of the IRLS weighted iterative least squares method, the braking deceleration is calculated. The perceived risk index of the lane where the vehicle is located Perform function fitting to obtain the braking deceleration-perception risk index function Based on the braking deceleration-perception risk index function Set the perceived risk index threshold; wherein the perceived risk index threshold includes: the gentle braking threshold R s , emergency braking threshold R e , Super emergency braking threshold R ee , Lane change threshold R c .

[0024] Preferably, the autonomous driving emergency decision-making of the intelligent networked vehicle includes:

[0025] When the perceived risk index of the lane where the vehicle is located is greater than or equal to the gentle braking threshold R s But less than the emergency braking threshold R e When the vehicle takes a gentle braking decision;

[0026] When the perceived risk index of the lane where the vehicle is located is greater than or equal to the emergency braking threshold R e and is less than the super emergency braking threshold R ee When the vehicle takes emergency braking decision;

[0027] When the perceived risk index of the lane where the vehicle is located is greater than or equal to the super emergency braking threshold R ee And the adjacent road perception risk index is less than the lane change threshold R c When the lane change decision is made;

[0028] When the perceived risk index of the lane where the vehicle is located is greater than or equal to the super emergency braking threshold R ee And the adjacent road perception risk index is greater than or equal to the lane change threshold R c When the vehicle is in emergency, an emergency braking decision is made.

[0029] The beneficial effects of the present invention are:

[0030] The present invention first obtains environmental information such as vehicles, pedestrians or obstacles in front of and around the vehicle's driving path by combining the intelligent perception capability of the intelligent connected vehicle and the V2X vehicle network information sharing capability, then calculates the perceived risk index through the driver risk field model, and finally makes an autonomous driving decision corresponding to the perceived risk index based on the threshold obtained by analyzing the driving data of real skilled drivers. Since the driver risk field is derived from the driver risk perception experiment, the decision obtained is closer to the actual driving decision of the real driver, and has human-like characteristics compared to general driving decisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0032] Figure 1 The figure is a flow chart of a method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0033] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0034] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0035] As attached Figure 1 As shown, this embodiment provides an automatic driving emergency decision-making method, including the following steps:

[0036] (1) Acquiring environmental information by combining the intelligent connected vehicle’s onboard intelligent sensor device and V2X vehicle network information sharing to extract information about potential risk targets includes the following steps:

[0037] (1-1) Through intelligent sensing devices on the intelligent connected vehicle (self-vehicle), intelligent perception is performed to obtain real-time environmental information of the surrounding unobstructed areas, identify potential risk targets (pedestrians, vehicles, non-motor vehicles, obstacles, etc.) in front of the lane, and track their trajectories.

[0038] (1-2) Use the V2X vehicle-to-vehicle (V2X) network of intelligent connected vehicles to share information and obtain environmental information of obstructed areas. Use the vehicle-to-vehicle (V2V) communication of intelligent connected vehicles to obtain the speed and relative position of each vehicle in the front convoy of the lane, the speed and relative position of the rear vehicle, and the speed and relative position of vehicles in adjacent lanes, regardless of whether the line of sight is obstructed. Based on the vehicle-to-pedestrian (V2P) communication of intelligent connected vehicles, the speed and relative position of nearby pedestrians can be obtained. Based on the vehicle-to-infrastructure (V2I) communication of intelligent connected vehicles, the information of traffic lights and traffic signs ahead can be obtained.

[0039] (1-3) The environmental information obtained by the on-board intelligent sensor device of the intelligent connected vehicle and the V2X vehicle network information sharing are integrated to obtain integrated environmental information.

[0040] (1-4) Taking the geometric center of the ego vehicle as the coordinate origin, the coordinate distance of the potential risk target and the speed relative to the ego vehicle are obtained from the fused environmental information, and the expected collision time between the ego vehicle and the potential risk target is calculated.

[0041] (2) Based on the environmental information obtained in steps (1-4), the driver risk field model is used to calculate the perceived risk index R of the current lane. d , the perceived risk index R of the adjacent lane nbr and other risk indicators.

[0042] (2-1) Based on the environmental information obtained in step (1-4), the relative position (D) of the nearest potential risk target (vehicle, pedestrian, obstacle, etc.) in front of the lane where the vehicle is located is calculated with the position of the geometric center of the vehicle as the coordinate origin. x ,D y ), relative speed V; calculate the relative position of the nearest potential risk target (vehicle, pedestrian, obstacle, etc.) in front of the adjacent lane Relative speed V nbr ;

[0043] (2-2) Calculate the perceived risk index R of the lane where the vehicle is located based on the driver risk field model d :

[0044]

[0045] Among them, D x is the lateral position of the potential risk target (vehicle, pedestrian, obstacle, etc.) in the lane where the vehicle is located relative to the vehicle's driving direction in the reference system of the vehicle, D y is the longitudinal position of the potential risk target (vehicle, pedestrian, obstacle, etc.) in the lane where the ego vehicle is located relative to the vehicle's driving direction in the ego vehicle reference system, and V is the relative speed of the potential risk target in the lane where the ego vehicle is located relative to the ego vehicle in the driving direction; t 1 ,t 2 ,t 3 ,t 4 is the model parameter, and the parameter size is obtained by least square fitting the data measured by the driver risk field experiment.

[0046] (2-3) Calculate the perceived risk index R of the adjacent lane based on the driver risk field model nbr :

[0047]

[0048] in, is the lateral position of the potential risk target (vehicle, pedestrian, obstacle, etc.) in the adjacent lane relative to the vehicle's driving direction in the vehicle reference frame. V is the longitudinal position of the potential risk target (vehicle, pedestrian, obstacle, etc.) in the adjacent lane relative to the vehicle's driving direction in the vehicle reference frame. nbr is the relative speed of the potential risk target in the adjacent lane relative to the vehicle in the driving direction; t 1 ,t 2 ,t 3 ,t 4 are model parameters. The parameter size is obtained by least square fitting the data measured in the driver risk field experiment and is consistent with that in (2-2).

[0049] (3) A driver risk field model was established by collecting real driving data of multiple drivers within three months (without accidents). The changes in the braking deceleration of the vehicle and the risk index of the lane where the vehicle was located were recorded in the real driving data of the drivers. The braking deceleration and the risk index of the lane where the vehicle was located were analyzed using multivariate nonlinear regression based on the IRLS weighted iterative least squares method. The perceived risk index of the lane where the vehicle is located Perform function fitting to obtain the braking deceleration-perception risk index function

[0050] (4) Based on the braking deceleration-perception risk index function obtained in step (3) Setting the perceived risk indicator threshold: Smooth braking threshold R s , emergency braking threshold R e , Super emergency braking threshold R ee In the braking deceleration-perception risk index function, the braking deceleration (such as 0m / s 2 ) corresponds to the lane perception risk index of the smooth braking threshold R s , braking deceleration (e.g. -4m / s 2 ) corresponds to the lane perception risk index of the emergency braking threshold R e , braking deceleration (e.g. -8m / s 2 ) is the corresponding lane perception risk index is the emergency braking threshold R ee Lane changing threshold R of adjacent lane perception risk index c Equal to the emergency braking threshold R e : R c =R e .

[0051] (5) When the vehicle is driving, the lane perception risk index R obtained in step (3) isd , the risk index R of the adjacent lane nbr The thresholds of the perceived risk indicators obtained in step (4) are used to make judgments and make emergency decisions for autonomous driving of intelligent connected vehicles.

[0052] (5-1) If the perceived risk index of the lane where the vehicle is located is R d Greater than or equal to the gentle braking decision threshold R s and is less than the emergency braking threshold R e , the vehicle executes a gentle braking decision;

[0053] (5-2) If the perceived risk index of the lane where the vehicle is located is R d Greater than or equal to the emergency braking decision threshold R e and is less than the super emergency braking threshold R ee , the vehicle executes emergency braking decision;

[0054] (5-3) If the perceived risk index of the lane where the vehicle is located is R d Greater than or equal to the super emergency braking threshold R ee , but the perceived risk index R of the adjacent lane nbr Less than the lane change decision threshold R c , the vehicle executes the lane-changing decision;

[0055] (5-4) If the perceived risk index of the lane where the vehicle is located is R d Greater than or equal to the super emergency braking threshold R ee , and the risk index of the adjacent lane is R nbr Greater than or equal to the lane change decision threshold R c , the vehicle executes an ultra-emergency braking decision.

[0056] An automatic driving emergency decision-making method proposed in the present invention is implemented on the basis of the intelligent connected vehicle's perception of the environment and the driver's risk field model. The driver's risk field model is a generalized driver risk perception model that is independent of specific scenarios. It describes the perceived risk brought to the driver by obstacles at different positions in front during driving and has strong adaptability. First, by combining the intelligent perception capability of the intelligent connected vehicle and the V2X vehicle network information sharing capability, environmental information such as vehicles, pedestrians or obstacles in front of and around the vehicle's driving path is obtained, and then the perceived risk index is calculated through the driver's risk field model. Finally, the threshold obtained based on the driving data analysis of real skilled drivers makes an automatic driving decision corresponding to the perceived risk index. Since the driver's risk field comes from the driver's risk perception experiment, the decision obtained is closer to the actual driving decision of the real driver, and has human-like characteristics compared to general driving decisions.

[0057] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. An autonomous driving emergency decision-making method, It is characterized in that include: Collecting environmental information about the vehicle, and extracting potential risk targets based on the environmental information, wherein the vehicle is an intelligent network-connected vehicle; Establishing a risk field model for intelligent connected vehicle drivers, and calculating the perceived risk index of the lane in which the vehicle is traveling and the perceived risk index of the adjacent lanes in combination with the potential risk target; Based on the driver's actual driving data, a perception risk index threshold is set, the perception risk index of the lane in which the vehicle is traveling and the perception risk index of the adjacent lane are compared with the perception risk index threshold, an automatic driving emergency decision of the vehicle is obtained, and an automatic driving emergency response is performed based on the automatic driving emergency decision of the vehicle; The perceived risk index of the lane in which the vehicle is traveling is calculated based on the intelligent connected vehicle driver risk field model. The method is: （1） in, is the lateral position of the potential risk target in the lane where the vehicle is located relative to the vehicle's driving direction in the vehicle reference system, is the longitudinal position of the potential risk target in the lane where the vehicle is located relative to the vehicle's driving direction in the vehicle reference system, The relative speed of the potential risk target in the lane where the vehicle is located relative to the vehicle in the driving direction; , , , are model parameters, The natural index.

2. The autonomous driving emergency decision-making method according to claim 1, It is characterized in that Extracting the potential risk targets includes: The vehicle's on-board intelligent sensor device is combined to perform intelligent perception, and information is shared with the V2X vehicle network to obtain the vehicle's surrounding environment information, wherein the vehicle's surrounding environment information includes environmental information of unobstructed areas around the vehicle and environmental information of obstructed areas.

3. The autonomous driving emergency decision-making method according to claim 2, It is characterized in that The environmental information of the unobstructed area around the vehicle includes: potential risk targets in front of the lane; wherein the potential risk targets include pedestrians and vehicles.

4. The autonomous driving emergency decision-making method according to claim 3, It is characterized in that Obtaining environmental information of the blocked area includes: Information is shared through the V2X vehicle network to obtain environmental information of the obstructed area, and the vehicle-to-vehicle communication of the vehicle is used to obtain the speed and relative position of each vehicle in the front team of the lane, the speed and relative position of the rear vehicle, and the speed and relative position of the vehicle in the adjacent lane; the speed and relative position of pedestrians in the lane are obtained based on the vehicle-to-pedestrian communication of the vehicle; and the information of non-motor vehicles, traffic lights, and traffic signs is obtained based on the vehicle-to-infrastructure communication of the vehicle.

5. The autonomous driving emergency decision-making method according to claim 2, It is characterized in that The environmental information obtained by intelligent perception of the vehicle's onboard intelligent sensor device and information sharing of the V2X vehicle network is integrated to obtain integrated environmental information; based on the integrated environmental information, the coordinate distance of the potential risk target and the speed relative to the vehicle are obtained, and the expected collision time between the vehicle and the potential risk target is calculated to obtain the perception risk index.

6. The autonomous driving emergency decision-making method according to claim 1, It is characterized in that Calculating the perceived risk index of the lane in which the vehicle is traveling and the perceived risk index of the adjacent lanes includes: taking the position of the geometric center of the vehicle as the coordinate origin, calculating the relative position of the nearest potential risk target in front of the lane in which the vehicle is located ( ) and relative speed ; Calculate the relative position of the nearest potential risk target in front of the adjacent lane of the vehicle ( ), relative speed , based on the intelligent connected vehicle driver risk field model, calculate the perceived risk index of the current lane , the perceived risk index of the adjacent lane .

7. The autonomous driving emergency decision-making method according to claim 6, It is characterized in that Calculate the perceived risk index of adjacent lanes based on the intelligent connected vehicle driver risk field model The method is: （2） in, is the lateral position of the potential risk target in the adjacent lane of the vehicle relative to the vehicle's driving direction in the vehicle reference system, is the longitudinal position of the potential risk target in the adjacent lane of the vehicle relative to the vehicle's driving direction in the vehicle reference system, The relative speed of the potential risk target in the adjacent lane to the vehicle in the driving direction; , , , are model parameters, The natural index.

8. The autonomous driving emergency decision-making method according to claim 1, It is characterized in that The setting of the perceived risk index threshold based on the driver's actual driving data includes: The real driving data of several drivers without accidents in any time period are collected, and the changes of the braking deceleration of the vehicle and the perceived risk index of the lane where the vehicle is located are recorded. Based on the multivariate nonlinear regression of the IRLS weighted iterative least squares method, the braking deceleration is analyzed. The perceived risk index of the lane where the vehicle is located Perform function fitting to obtain the braking deceleration-perception risk index function , based on the braking deceleration-perception risk index function , set the perceived risk index threshold; wherein the perceived risk index threshold includes: smooth braking threshold , Emergency braking threshold , Super emergency braking threshold , Lane change threshold .

9. The autonomous driving emergency decision-making method according to claim 8, It is characterized in that The vehicle automatic driving emergency decision-making includes: When the perceived risk index of the lane where the vehicle is located is greater than or equal to the gentle braking threshold but less than the emergency braking threshold When the vehicle takes a gentle braking decision; When the perceived risk index of the lane where the vehicle is located is greater than or equal to the emergency braking threshold and less than the super emergency braking threshold When the vehicle takes emergency braking decision; When the perceived risk index of the lane where the vehicle is located is greater than or equal to the super emergency braking threshold and the adjacent road perception risk index is less than the lane change threshold When the lane change decision is made; When the perceived risk index of the lane where the vehicle is located is greater than or equal to the super emergency braking threshold And the adjacent road perception risk index is greater than or equal to the lane change threshold When the vehicle is in emergency, an emergency braking decision is made.

Images