Driving risk prediction method, apparatus and system, and storage medium

Abstract

A driving risk prediction method, apparatus and system, and a storage medium. The method comprises: on the basis of environment sensing data during the driving of a target vehicle, making a future risk prediction on an adjacent vehicle of the target vehicle, so as to determine whether the adjacent vehicle will encounter a safety risk from an ambient environment in the future; and, when a risk is predicted, predicting a risk avoidance measure, predicting a future driving risk of the target vehicle and, when there is a future driving risk, giving an early risk warning.

Description

Driving risk prediction method, device, system and storage medium

[0001] The present application claims priority to Chinese application No. 2024108710074, filed on June 27, 2024, entitled: Driving risk prediction method, device, system and storage medium, the entire contents of which are incorporated herein by reference. TECHNICAL FIELD

[0002] Embodiments of the present application relate to, but are not limited to, the technical field of vehicle control, and in particular to a driving risk prediction method, device, system and storage medium. BACKGROUND

[0003] With the increase of vehicle ownership and the expansion of the driving population, traffic accidents occur frequently, and the driving safety situation is severe. During vehicle driving, it is easy to be affected by other surrounding vehicles, for example, some dangerous driving behaviors of surrounding vehicles (such as sudden lane merging of vehicles in the adjacent lane, emergency braking of vehicles in front, etc.). In order to avoid accidents, the vehicle needs to take emergency measures (emergency braking, emergency steering), which greatly affects the comfort of users riding in the vehicle, and even may result in the failure of the execution of emergency measures and the occurrence of traffic accidents, resulting in low driving safety and user experience. TECHNICAL SOLUTION

[0004] In order to solve the above technical problems or at least partially solve the above technical problems, the present disclosure provides a driving risk prediction method, device, system and storage medium to solve the problem that the vehicle driving process is easily affected by surrounding vehicles, and emergency measures need to be taken, resulting in low driving safety and user experience.

[0005] The following is a summary of the subject matter described in detail herein. This summary is not intended to limit the scope of the claims.

[0006] To achieve the above-mentioned purpose, a driving risk prediction method of the present application comprises:

[0007] Based on the environmental perception data in the driving process of the target vehicle, the future risk of the adjacent vehicle of the target vehicle is predicted to determine whether the adjacent vehicle will encounter a safety risk from the surrounding environment in the future;

[0008] In response to predicting that the adjacent vehicle will encounter a safety risk in the future, the risk avoidance measures that the adjacent vehicle can take to avoid the safety risk in the future are predicted to obtain predicted adjacent vehicle risk avoidance measure information; and

[0009] Based on the predicted adjacent vehicle risk avoidance measure information, the driving risk of the target vehicle in the future is predicted, and in response to determining that the target vehicle has a driving risk in the future, a risk warning is performed.

[0010] Optionally, the environment perception data comprises state information of obstacles within a preset range of the target vehicle, and a future risk of the neighboring vehicle of the target vehicle is predicted to determine whether the neighboring vehicle will encounter a safety risk from the surrounding environment in the future, comprising:

[0011] Each target obstacle is classified and screened based on the state information of each obstacle to screen out the neighboring vehicle of the target vehicle and the neighboring obstacle of the neighboring vehicle; and

[0012] The future risk of the neighboring vehicle is predicted based on the state information of the neighboring vehicle and the state information of the neighboring obstacle to determine whether the neighboring vehicle will encounter a safety risk in the future.

[0013] Optionally, the future risk of the neighboring vehicle is predicted to determine whether the neighboring vehicle will encounter a safety risk in the future, comprising:

[0014] The collision time required for the neighboring vehicle to collide with the neighboring obstacle in the future is predicted based on the state information of the neighboring vehicle and the state information of the neighboring obstacle to obtain a first predicted collision time;

[0015] The time required for the neighboring vehicle to avoid the collision risk is determined to obtain a neighboring vehicle risk avoidance time; and

[0016] In response to the first predicted collision time being less than or equal to the neighboring vehicle risk avoidance time, it is determined that the neighboring vehicle will encounter a safety risk in the future.

[0017] Optionally, the driving risk of the target vehicle in the future is predicted based on the predicted neighboring vehicle risk avoidance measure information, comprising:

[0018] The risk avoidance risk of the target vehicle in the future is predicted based on the predicted neighboring vehicle risk avoidance measure information;

[0019] The environmental safety risk of the target vehicle in the future is predicted based on the environment perception data; and

[0020] In response to the prediction that the target vehicle has a risk avoidance risk in the future and has an environmental safety risk, it is determined that the target vehicle has a driving risk in the future.

[0021] Optionally, the risk avoidance risk of the target vehicle in the future is predicted based on the predicted neighboring vehicle risk avoidance measure information, comprising:

[0022] In response to the target vehicle and the neighboring vehicle being in the same lane, the neighboring vehicle being in front of the target vehicle, and the predicted neighboring vehicle risk avoidance measure information being braking and parking or turning and changing lanes, it is determined that the target vehicle has a risk avoidance risk in the future;

[0023] In response to the target vehicle and the neighboring vehicle being in different lanes and the predicted neighboring vehicle risk avoidance measure information being turning and changing lanes, it is determined that the target vehicle has a risk avoidance risk in the future.

[0024] Optionally, based on the environment perception data, predicting whether the target vehicle has an environmental safety risk in the future, comprises:

[0025] Based on the environment perception data, predicting the safety influence of the surrounding environment on the target vehicle to obtain a safety parameter of the target vehicle; and

[0026] Based on the safety parameter, predicting whether the target vehicle has an environmental safety risk in the future.

[0027] Optionally, the safety parameter comprises a collision time of the target vehicle colliding with the adjacent vehicle, and based on the environment perception data, predicting the safety influence of the surrounding environment on the target vehicle to obtain the safety parameter of the target vehicle, comprises:

[0028] Based on the environment perception data, determining whether the adjacent vehicle and the target vehicle are in the same lane; and

[0029] In response to determining that the adjacent vehicle and the target vehicle are in the same lane, predicting a collision time required for the target vehicle to collide with the adjacent vehicle in the future to obtain a second predicted collision time.

[0030] Optionally, based on the safety parameter, predicting whether the target vehicle has an environmental safety risk in the future, comprises:

[0031] Predicting a time required for the target vehicle to evade the adjacent vehicle to obtain a self-vehicle risk evasion time; and

[0032] Based on the second predicted collision time and the self-vehicle risk evasion time, determining whether the target vehicle has an environmental safety risk in the future, the second predicted collision time being a collision time required for the target vehicle to collide with the adjacent vehicle in the future when the target vehicle and the adjacent vehicle are in the same lane.

[0033] Optionally, based on the second predicted collision time and the self-vehicle risk evasion time, determining whether the target vehicle has an environmental safety risk in the future, comprises:

[0034] In response to the second predicted collision time being less than or equal to the self-vehicle risk evasion time, determining that the target vehicle has an environmental safety risk in the future;

[0035] Or,

[0036] In response to the second predicted collision time being less than the self-vehicle risk evasion time and a change rate of the second predicted collision time being greater than a preset change rate, determining that the target vehicle has an environmental safety risk in the future.

[0037] Optionally, before predicting the driving risk of the target vehicle in the future based on the predicted adjacent vehicle risk avoidance measure information, the method further comprises:

[0038] The validity of the environmental perception data and the driving data of the target vehicle is checked to obtain checking information; and

[0039] In response to the checking information being all valid, the future driving risk of the target vehicle is predicted based on the predicted neighboring vehicle risk avoidance measure information.

[0040] Optionally, in response to determining that the target vehicle has a future driving risk, the method further comprises:

[0041] The driving path of the target vehicle is optimized.

[0042] Optionally, the optimization of the driving path of the target vehicle comprises:

[0043] In response to the target vehicle and the neighboring vehicle being in the same lane, the target vehicle is controlled to brake and / or change lanes; and

[0044] In response to the target vehicle and the neighboring vehicle being in different lanes, and the predicted neighboring vehicle risk avoidance measure information being lane changing by turning, the target vehicle is controlled to change to a target lane, the target lane including a lane other than the lane where the neighboring vehicle is located and / or the lane where the target vehicle is located.

[0045] A driving risk prediction device is provided, comprising:

[0046] A neighboring vehicle risk predictor configured to predict a future risk of a neighboring vehicle of a target vehicle based on environmental perception data during driving of the target vehicle, to determine whether the neighboring vehicle will encounter a safety risk from the surrounding environment in the future; in response to predicting that the neighboring vehicle will encounter a safety risk in the future, predict an avoidance measure that the neighboring vehicle can take to avoid the safety risk in the future, to obtain predicted neighboring vehicle risk avoidance measure information; and

[0047] A target vehicle risk predictor configured to predict a future driving risk of the target vehicle based on the predicted neighboring vehicle risk avoidance measure information, and in response to determining that the target vehicle has a future driving risk, to perform a risk warning.

[0048] A driving risk prediction device is provided, comprising a memory, a processor, and computer instructions stored in the memory and executable on the processor, the processor executing the computer instructions to implement the steps of the above driving risk prediction method.

[0049] A driving risk prediction system is provided, comprising a target vehicle and the above driving risk prediction device.

[0050] A readable storage medium is provided, the readable storage medium storing computer instructions, the computer instructions being executed by a processor to implement the steps of the above driving risk prediction method.

[0051] In one scheme provided by the driving risk prediction method, device, system and storage medium, future risk of a neighboring vehicle of a target vehicle is predicted based on environment perception data of the target vehicle in a driving process, to determine whether the neighboring vehicle will encounter a safety risk from a surrounding environment in the future; when it is predicted that the neighboring vehicle will encounter a safety risk in the future, risk avoidance measures that can be taken by the neighboring vehicle to avoid the safety risk in the future are predicted, to obtain predicted neighboring vehicle risk avoidance measure information; future driving risk of the target vehicle is predicted based on the predicted neighboring vehicle risk avoidance measure information, and risk warning is performed after it is determined that the target vehicle has future driving risk. By identifying the safety risk caused by the surrounding environment of the neighboring vehicle to the neighboring vehicle, risk avoidance measures that can be taken by the neighboring vehicle are predicted, and then potential driving risk of the ego vehicle caused by the predicted neighboring vehicle risk avoidance measure information is predicted, so that the driver can be informed of the future driving risk of the ego vehicle in time, and then risk avoidance can be performed in advance, the adverse effects of the neighboring vehicle on the ego vehicle caused by emergency braking or turning of the neighboring vehicle due to subsequent discovery of the risk can be avoided, and the situation of the ego vehicle taking emergency risk avoidance measures can be reduced, thereby improving driving safety and user experience. BRIEF DESCRIPTION OF DRAWINGS

[0052] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the description of the embodiments of the present application will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0053] FIG. 1 is a structural schematic diagram of a driving risk prediction system in an embodiment of the present application;

[0054] FIG. 2 is a flow schematic diagram of a driving risk prediction method in an embodiment of the present application;

[0055] FIG. 3 is a distribution diagram of a target vehicle and a neighboring vehicle in an embodiment of the present application;

[0056] FIG. 4 is another flow schematic diagram of a driving risk prediction method in an embodiment of the present application;

[0057] FIG. 5 is a driving risk prediction effect diagram of a target vehicle in an embodiment of the present application;

[0058] FIG. 6 is another driving risk prediction effect diagram of a target vehicle in an embodiment of the present application;

[0059] FIG. 7 is an implementation flow schematic diagram of step S10 in FIG. 2 or FIG. 4;

[0060] FIG. 8 is an implementation flow schematic diagram of step S30 in FIG. 2 or FIG. 4;

[0061] FIG. 9 is a structural schematic diagram of a driving risk prediction device in an embodiment of the present application;

[0062] FIG. 10 is another structural schematic diagram of the driving risk prediction device in an embodiment of the present application.

[0063] Embodiments of the present application

[0064] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0065] It can be understood that the size of the serial number of each step in the following embodiments does not mean the order of execution. The execution order of each process should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0066] In order to illustrate the technical solutions of the present application, the following will be described by specific embodiments.

[0067] The driving risk prediction method provided by the embodiments of the present application can be applied to a driving risk prediction system as shown in FIG. 1, which includes a target vehicle and a driving risk prediction device. The driving risk prediction device communicates with the target vehicle through a bus.

[0068] It should be understood that the adjacent vehicles include vehicles in the same lane (front and rear vehicles in the same lane), vehicles in adjacent lanes (including side vehicles in adjacent lanes, front vehicles of side vehicles and rear vehicles of side vehicles). In the process of driving, due to the too close following distance or too fast speed of the adjacent vehicles, the adjacent vehicles may react too late to the dangers around them (such as sudden braking of the front vehicle, road construction, etc.), and take emergency avoidance measures (emergency braking, emergency steering), which may cause potential risks to the host vehicle (self-vehicle) and cause traffic accidents. In order to avoid accidents, the host vehicle needs to take emergency avoidance measures (emergency braking, emergency steering), which greatly affects the comfort of the user during driving, and even may cause the failure of the execution of the emergency avoidance measures and the occurrence of traffic accidents, resulting in low driving safety and user experience.

[0069] In the driving process of the target vehicle, the driving risk prediction device can perceive the environment around the target vehicle to obtain environment perception data, and then predict the future risk of the adjacent vehicle of the target vehicle based on the environment perception data in the driving process of the target vehicle, to determine whether the adjacent vehicle will encounter a safety risk from the surrounding environment in the future; when it is predicted that the adjacent vehicle will encounter a safety risk in the future, the risk avoidance measures that can be taken by the adjacent vehicle to avoid the safety risk in the future are predicted to obtain predicted adjacent vehicle risk avoidance measure information; the future driving risk of the target vehicle is predicted based on the predicted adjacent vehicle risk avoidance measure information, and a risk warning is performed after it is determined that the target vehicle has a future driving risk. In the embodiment, the safety risk caused by the environment around the adjacent vehicle is identified to predict the risk avoidance measures that can be taken by the adjacent vehicle, and then the potential driving risk caused by the adjacent vehicle risk avoidance measure information to the ego vehicle is predicted, so that the driver can timely know the future driving risk of the ego vehicle, and then take risk avoidance measures in advance, which can avoid the adverse effects of the subsequent discovery of the risk by the adjacent vehicle on the ego vehicle caused by emergency braking or turning, reduce the situation of taking emergency risk avoidance measures by the ego vehicle, and thus improve the driving safety and user experience.

[0070] In the embodiment, the driving risk prediction system includes a target vehicle and a driving risk prediction device, which is only illustrative, and in other embodiments, the driving risk prediction system also includes other devices. For example, the driving risk prediction system can also include a plurality of adjacent vehicles of the target vehicle, and an environment perception device installed on the target vehicle. The environment perception device is a sensor for image acquisition, including millimeter wave radar (such as 4D millimeter wave radar), laser radar, camera, and the like.

[0071] In an embodiment, as shown in FIG. 2, a driving risk prediction method is provided, and the driving risk prediction device in FIG. 1 is taken as an example to illustrate the method, which includes the following steps:

[0072] S10: predicting the future risk of the adjacent vehicle of the target vehicle based on the environment perception data in the driving process of the target vehicle, to determine whether the adjacent vehicle will encounter a safety risk from the surrounding environment in the future.

[0073] In the driving process of the target vehicle, the driving risk prediction device can perceive the environment around the target vehicle to obtain environment perception data. The environment perception data can be collected by the sensors installed in the target vehicle, such as 4D millimeter wave radar, laser radar, and camera.

[0074] Then, the driving risk prediction device predicts future risk of the adjacent vehicle (neighbor) of the target vehicle based on the environment perception data during driving of the target vehicle, to determine whether the adjacent vehicle will encounter safety risk from the surrounding environment in the future. For example, whether congestion occurs in front of the adjacent vehicle, whether the adjacent vehicle will collide with the obstacle (such as a vehicle) in front of it in the future, etc. If congestion occurs in front of the adjacent vehicle, or the adjacent vehicle will collide with the obstacle in front of it in the future, it is determined that the adjacent vehicle will encounter safety risk from the surrounding environment in the future.

[0075] The environment perception data includes road information (including lane information, road type), state information of the adjacent vehicle of the target vehicle, and state information of the obstacles (including vehicles) around the adjacent vehicle. The state information includes position information, speed information (including speed in multiple directions), acceleration information (including acceleration in multiple directions), heading information, etc. of the vehicle. The environment perception data also includes unique identification code (ID) of each obstacle (including the adjacent vehicle and the obstacles around the adjacent vehicle). The lane information includes ID, position information, curvature related parameters and information source (such as collected from sensors, or navigation information, or both fitted prediction, etc.) of each lane line.

[0076] The driving risk prediction device will classify and screen each target obstacle based on the lane information and the state information of each obstacle, to screen out the adjacent vehicle of the target vehicle, the adjacent obstacle of the adjacent vehicle, and further predict whether the adjacent vehicle will encounter safety risk from the surrounding environment (such as its neighbor) in the future. For example, it can be determined from the environment perception data whether there is road congestion in front of the lane where the adjacent vehicle is located, such as parking blockage caused by collision, construction site, traffic signal, etc. If there is congestion, it is determined that the adjacent vehicle will encounter safety risk in the future.

[0077] As shown in FIG. 3, a distribution diagram of the target vehicle and its adjacent vehicles is shown. The center area of FIG. 3 represents the target vehicle (Subject vehicle), and the arrow direction represents the driving direction (Travel direction) of the target vehicle. The inner circle areas 1-8 represent the adjacent vehicles of the target vehicle. The outer circle areas 9-24 represent the adjacent vehicles (neighbors) of the adjacent vehicles. As can be seen from the figure, in this embodiment, at least the state information of at least 24 area points around can be identified, and the risk influence of the vehicles in the surrounding near circle (i.e. the adjacent vehicles) on the ego vehicle can be predicted.

[0078] The state information of the adjacent vehicle of the target vehicle and the state information of the obstacles around the adjacent vehicle are detected by means of multipath propagation, and the position information and motion parameter information of the adjacent vehicle and the peripheral objects (including vehicles) are perceived, and then the future risk of the adjacent vehicle of the target vehicle is predicted based on the position information and motion parameter information of the adjacent vehicle and the peripheral objects (including vehicles), so as to determine whether the adjacent vehicle will encounter a safety risk from the surrounding environment in the future. The environmental perception data in the embodiment only relies on the equipment of the own vehicle to complete detection, and does not require the surrounding environment vehicles to also have special communication devices, so that the surrounding vehicles can be detected, the cost of obtaining environmental perception data is reduced, the real-time and safety problems that may occur in the collection and transmission of data of other vehicles are reduced, the accuracy, real-time and safety of the collection of environmental perception data are improved, and the speed of risk prediction is improved.

[0079] S20: When it is predicted that the adjacent vehicle will encounter a safety risk in the future, the risk avoidance measures that can be taken by the adjacent vehicle to avoid the safety risk in the future are predicted, and adjacent vehicle risk avoidance measure information is obtained.

[0080] When it is predicted that the adjacent vehicle will encounter a safety risk in the future, the risk avoidance measures that can be taken by the adjacent vehicle to avoid the safety risk in the future are predicted by the driving risk prediction device, that is, the risk avoidance operations that can be taken by the adjacent vehicle when the safety risk is identified are pre-judged, so that the adjacent vehicle risk avoidance measure information is obtained.

[0081] S30: The driving risk of the target vehicle in the future is predicted based on the adjacent vehicle risk avoidance measure information, and risk warning is performed when it is determined that the target vehicle has a driving risk in the future.

[0082] Finally, the driving risk prediction device predicts the driving risk of the target vehicle in the future based on the adjacent vehicle risk avoidance measure information, and performs risk warning when it is determined that the target vehicle has a driving risk in the future, so that the driver can learn the driving risk of the vehicle in the future in time and take response operations in advance to avoid risks.

[0083] For example, the driving risk of the target vehicle in the future is predicted based on the predicted neighboring vehicle risk avoidance measure information, including: when the target vehicle and the neighboring vehicle are in the same lane, and the neighboring vehicle is in front of the target vehicle, and the predicted neighboring vehicle risk avoidance measure information is that the neighboring vehicle brakes to stop or turns to change lanes, the neighboring vehicle will affect the normal driving of the target vehicle, and it is determined that the target vehicle has driving risk in the future. When the target vehicle and the neighboring vehicle are in different lanes, and the neighboring vehicle is in front of the target vehicle, and the predicted neighboring vehicle risk avoidance measure information is that the neighboring vehicle turns to change lanes, the neighboring vehicle may turn to change lanes to the lane where the target vehicle is located in the future, thereby affecting the normal driving of the target vehicle, and it is determined that the target vehicle has driving risk in the future. When the target vehicle and the neighboring vehicle are in different lanes, if the predicted neighboring vehicle risk avoidance measure information is that the neighboring vehicle brakes to stop, the neighboring vehicle cannot affect the target vehicle, and it is determined that the target vehicle has no driving risk in the future.

[0084] The risk warning includes the neighboring vehicle with safety risk and the safety risk that the neighboring vehicle may encounter. In other embodiments, the risk warning can also include basic information of the neighboring vehicle (such as position information, heading angle information, and speed information, and state information such as vehicle type and color) so that the driver can identify and perform risk avoidance operations subsequently.

[0085] In this embodiment, by identifying the safety risk caused by the surrounding environment of the neighboring vehicle, the risk avoidance measure that the neighboring vehicle may take is predicted, and the potential driving risk caused by the neighboring vehicle risk avoidance measure information to the ego vehicle is predicted, so that the driver can know the future driving risk of the ego vehicle in time, and then perform risk avoidance in advance, which can avoid the adverse effects of the neighboring vehicle on the ego vehicle caused by the subsequent discovery of risk and emergency braking or turning, reduce the situation of the ego vehicle taking emergency risk avoidance measures, and thus improve the driving safety and user experience.

[0086] In an embodiment, as shown in FIG. 4, after it is determined that the target vehicle has driving risk in the future, the method further includes the following steps:

[0087] S40: optimizing the driving path of the target vehicle.

[0088] After it is determined that the target vehicle has driving risk in the future, the driving risk prediction device needs to optimize the driving path of the target vehicle while performing risk warning. Specifically, the driving path of the target vehicle can be optimized based on the position condition of the target vehicle and the neighboring vehicle with safety risk. For example, if the target vehicle and the neighboring vehicle with safety risk are in the same lane, the target vehicle is directly controlled to turn to change lanes to change lanes, and even if the neighboring vehicle subsequently identifies the safety risk and performs emergency braking or turning, it cannot affect the target vehicle.

[0089] Wherein, after determining that the target vehicle has a driving risk in the future, the driving risk prediction device sends the state information of the adjacent vehicle with safety risk and the predicted adjacent vehicle risk avoidance measure information to the path planning device of the target vehicle, so as to optimize the driving path through the path planning device.

[0090] In this embodiment, the driving path of the target vehicle is optimized based on the predicted adjacent vehicle risk avoidance measure information, so that the driving path of the ego vehicle can be changed in advance according to the predicted potential risk, the risk is avoided in advance, the situation of taking emergency risk avoidance measures by the ego vehicle is reduced, and the driving safety and user experience are improved. The driving path is automatically optimized, and the driver does not need to operate automatically, and the intelligence is higher.

[0091] In an embodiment, in step S40, the driving path of the target vehicle is optimized, including:

[0092] S41: When the target vehicle and the adjacent vehicle are in the same lane, the target vehicle is controlled to brake and / or change lanes.

[0093] After determining that the target vehicle has a driving risk in the future, the driving risk prediction device needs to determine the position situation (specific lane) of the target vehicle and the adjacent vehicle, so as to optimize the driving path of the target vehicle according to the predicted adjacent vehicle risk avoidance measure information of the adjacent vehicle and the position situation of the two vehicles, improve the path optimization effect, and further improve the driving safety.

[0094] Wherein, when the target vehicle and the adjacent vehicle are in the same lane, no matter whether the predicted adjacent vehicle risk avoidance measure information of the adjacent vehicle is braking and parking or turning and changing lanes, the driving risk prediction device controls the target vehicle to brake and / or change lanes, reduces the situation of taking emergency risk avoidance measures by the target vehicle in the future, and ensures the driving safety and improves the user experience.

[0095] As shown in FIG. 5, the whole process of risk judgment and path optimization of the target vehicle during driving is shown, and the target vehicle, the adjacent vehicle of the target vehicle, and the adjacent vehicle of the adjacent vehicle are sequentially arranged from bottom to top in the figure, and the three are located in the same lane, and the target vehicle (the ego vehicle) follows the vehicle. When there is congestion in front of the lane, such as a collision accident of the adjacent vehicle of the adjacent vehicle, the driving risk prediction device of the target vehicle predicts in advance that the adjacent vehicle may have a safety risk, and predicts that the adjacent vehicle may take risk-avoiding measures such as (emergency) braking or (emergency) steering, and before the adjacent vehicle identifies the safety risk, the target vehicle judges that the ego vehicle has a potential risk (risk), the target vehicle predicts that the ego vehicle has a driving risk, and performs early warning, and controls the target vehicle to brake or change lanes to brake and stop or change to the left lane or the right lane. After that, even if the adjacent vehicle performs emergency braking (stop) when it finds its own safety risk, the target vehicle and the adjacent vehicle have a long collision time (TTC), and it is difficult for the two vehicles to send a rear-end collision, and the rear-end collision of the target vehicle and the adjacent vehicle can be avoided, thereby ensuring the driving safety of the target vehicle.

[0096] S42: When the target vehicle and the adjacent vehicle are in different lanes, if the predicted adjacent vehicle risk-avoiding measure information is steering and lane changing, the target vehicle is controlled to change to the target lane.

[0097] When the target vehicle and the adjacent vehicle are in different lanes, if the predicted adjacent vehicle risk-avoiding measure information is steering and lane changing, the target vehicle is controlled to change to the target lane, wherein the target lane includes a lane other than the lane where the adjacent vehicle is located and / or the lane where the target vehicle is located.

[0098] As shown in FIG. 6, the whole process of risk judgment and path optimization of the target vehicle during driving is shown, and the target vehicle (the ego vehicle) is in a constant speed cruise, and at this time, there is a vehicle in the adjacent lane of the target vehicle (there is a vehicle in the adjacent lane); the target vehicle, the adjacent vehicle of the target vehicle, and the adjacent vehicle of the adjacent vehicle are sequentially arranged from bottom to top in the figure. When there is congestion in front of the lane of the adjacent vehicle, the driving risk prediction device of the target vehicle predicts in advance that the adjacent vehicle may have a safety risk, and predicts that the adjacent vehicle may take risk-avoiding measures such as (emergency) braking or (emergency) steering, and before the adjacent vehicle identifies the safety risk, the target vehicle judges that the ego vehicle has a potential risk (risk), the target vehicle predicts that the ego vehicle has a driving risk, and performs early warning, and controls the target vehicle to change lanes to change to the target lane, that is, to a lane other than the lane where the two vehicles are located. After that, even if the adjacent vehicle performs steering and lane changing (emergency steering) when it finds its own safety risk, it cannot affect the target vehicle, and it is difficult for the two vehicles to send a rear-end collision, and the collision of the target vehicle and the adjacent vehicle can be avoided, thereby ensuring the driving safety of the target vehicle.

[0099] In the embodiment, when the target vehicle and the adjacent vehicle are in the same lane, the target vehicle is controlled to brake and / or change lanes; when the target vehicle and the adjacent vehicle are in different lanes, if the predicted adjacent vehicle safety measure information is lane changing by turning, the target vehicle is controlled to change to the target lane, the target lane including the lane where the adjacent vehicle is located and / or the lane outside the lane where the target vehicle is located, and the driving path of the target vehicle is optimized according to the predicted adjacent vehicle safety measure information of the adjacent vehicle and the position condition of the two vehicles, so as to improve the path optimization effect and further improve the driving safety.

[0100] In an embodiment, the environment perception data includes lane information and state information of obstacles within a preset range of the target vehicle. As shown in FIG. 7, in step S10, the future risk of the adjacent vehicle of the target vehicle is predicted to determine whether the adjacent vehicle will encounter a safety risk from the surrounding environment in the future, which specifically includes the following steps:

[0101] S11: Based on the lane information and the state information of each obstacle, each target obstacle is classified and screened to screen out the adjacent vehicle of the target vehicle and the adjacent obstacle of the adjacent vehicle.

[0102] In the embodiment, the environment perception data includes lane information, state information of the adjacent vehicle of the target vehicle, and state information of obstacles within a preset range of the adjacent vehicle. The state information includes position information, speed information (including speed in multiple directions), acceleration information (including acceleration in multiple directions), heading information, etc. of the vehicle. The environment perception data further includes unique identification codes (IDs) of each obstacle (including the adjacent vehicle and obstacles around the adjacent vehicle). The lane information includes IDs, position information, curvature-related parameters, and information sources (such as collected from sensors, or navigation information, or both fitted and predicted) of each lane line.

[0103] After the environment perception data is collected by the sensor, the driving risk prediction device classifies and screens each target obstacle based on the lane information and the state information of each obstacle to screen out the adjacent vehicle of the target vehicle and the adjacent obstacle of the adjacent vehicle.

[0104] S12: Based on the state information of the adjacent vehicle and the state information of the adjacent obstacle, the future risk of the adjacent vehicle is predicted to determine whether the adjacent vehicle will encounter a safety risk in the future.

[0105] Then, the driving risk prediction device predicts, based on the state information of the adjacent vehicle and the state information of the adjacent obstacle, a future risk of the adjacent vehicle to determine whether the adjacent vehicle will encounter a safety risk in the future. For example, based on the state information of the adjacent vehicle and the state information of the adjacent obstacle, a possibility of a collision between the adjacent vehicle and the adjacent obstacle in front of the adjacent vehicle is predicted. If it is predicted that the adjacent vehicle and the adjacent obstacle in front of the adjacent vehicle will collide in the future, it is determined that the adjacent vehicle will encounter a safety risk in the future.

[0106] In this embodiment, based on the road information and the state information of each obstacle, each target obstacle is classified and screened to screen the adjacent vehicle of the target vehicle and the adjacent obstacle of the adjacent vehicle. Then, based on the state information of the adjacent vehicle and the state information of the adjacent obstacle, a future risk of the adjacent vehicle is predicted to determine whether the adjacent vehicle will encounter a safety risk in the future, which refines the future safety risk prediction step of the adjacent vehicle. By predicting the future risk of the adjacent vehicle based on the state information of the adjacent vehicle and the state information of the adjacent obstacle, the risk that has not occurred can be effectively predicted, and the accuracy is higher.

[0107] In an embodiment, in step S12, that is, predicting a future risk of the adjacent vehicle to determine whether the adjacent vehicle will encounter a safety risk in the future, the following steps are specifically included:

[0108] S121: predicting, based on the state information of the adjacent vehicle and the state information of the adjacent obstacle, a collision time required for a collision between the adjacent vehicle and the adjacent obstacle in the future to obtain a first predicted collision time.

[0109] The driving risk prediction device predicts, based on the state information of the adjacent vehicle and the state information of the adjacent obstacle, a collision time required for a collision between the adjacent vehicle and the adjacent obstacle in the future to obtain a first predicted collision time. The state information includes position information, speed information, and acceleration information.

[0110] The vehicle dynamics model can be used to predict the collision time required for a collision between the adjacent vehicle and the adjacent obstacle (such as the adjacent vehicle) in the future. Specifically, based on the lane information, the position information of the adjacent vehicle, and the position information of the adjacent obstacle, it is determined whether the adjacent vehicle and the adjacent obstacle are located in the same lane. If they are located in the same lane, based on the respective speed information and acceleration information, the respective dynamics models are determined to obtain the dynamics model of the adjacent vehicle and the dynamics model of the adjacent obstacle in the same lane. Meanwhile, the theoretical trajectory function of the adjacent vehicle is determined based on the lane line function. The median interpolation method can be used to median interpolate the lane line functions of the two lane lines of the lane where the adjacent vehicle is located to obtain the theoretical trajectory function of the adjacent vehicle.

[0111] The dynamics model of the adjacent vehicle is represented by the following formula:

[0112] wherein D1 represents the dynamics model of the adjacent vehicle; v 1x , v 1y , v 1z respectively represent the speed of the adjacent vehicle in the x direction (vehicle driving direction), y direction (horizontal direction perpendicular to the x direction), and z direction (vertical direction); a 1x a 1y a 1z respectively represent the acceleration of the adjacent vehicle in the x direction, y direction, and z direction; j 1x j 1y j 1z respectively represent the jerk of the adjacent vehicle in the x direction, y direction, and z direction, i.e., the acceleration of the acceleration.

[0113] The dynamics model of the adjacent obstacle of the adjacent vehicle is represented by the following formula:

[0114] wherein D0 represents the dynamics model of the adjacent vehicle; v 0x , v 0y , v 0z respectively represent the speed of the adjacent vehicle in the x direction (vehicle driving direction), y direction (horizontal direction perpendicular to the x direction), and z direction (vertical direction); a 0x , a 0y , a 0z respectively represent the acceleration of the adjacent vehicle in the x direction, y direction, and z direction; j 0x , j 0y , j 0z respectively represent the jerk of the adjacent vehicle in the x direction, y direction, and z direction, i.e., the acceleration of the acceleration.

[0115] The lane line function is as follows:

[0116] L i = L(x, y, z); i = 1, 2…n

[0117] wherein L i represents the lane line function of the lane line i, and n is the number of lane lines; x, y, and z respectively represent the position information (spatial coordinates) of each point on the lane line.

[0118] Based on the above, the first predicted collision time can be determined according to the following formula:

[0119] wherein TTC 1-0represents the first predicted collision time; P1 represents the spatial coordinates of the adjacent vehicle; P0 represents the position information (spatial coordinates) of the adjacent obstacle of the adjacent vehicle; L i-(i+1) represents the theoretical trajectory function of the adjacent vehicle, which is obtained by fitting the lane line function L i of the lane where the adjacent vehicle is located to the position information (spatial coordinates) of the adjacent obstacle of the adjacent vehicle. i+1 is obtained by median interpolation; D1 represents the dynamics model of the adjacent vehicle; and D0 represents the dynamics model of the adjacent obstacle of the adjacent vehicle.

[0120] In other embodiments, a pre-trained collision time model is obtained, the state information of the adjacent vehicle and the state information of the adjacent obstacle are taken as inputs of the collision time model, the collision possibility is predicted by the collision time model, and the collision time required for the adjacent vehicle to collide with the adjacent obstacle in the future is predicted to obtain the first predicted collision time.

[0121] S122: determining the time required for the adjacent vehicle to avoid the collision risk to obtain the adjacent vehicle risk avoidance time.

[0122] Then, the driving risk prediction device determines the time required for the adjacent vehicle to avoid the collision risk to obtain the adjacent vehicle risk avoidance time. The adjacent vehicle risk avoidance time or the ego vehicle risk avoidance time is the reaction time required for the driver in the vehicle to take emergency measures to avoid the safety risk when encountering the safety risk. The adjacent vehicle risk avoidance time can be a fixed value pre-calibrated based on real vehicle test data.

[0123] In other embodiments, a driver reaction model can also be introduced to predict the time required for the adjacent vehicle to avoid the collision risk to improve the accuracy of the adjacent vehicle risk avoidance time. The driver reaction model is calibrated or trained based on real vehicle test data and is a model for predicting the time required for the vehicle to avoid risks under different working conditions.

[0124] The driver reaction model includes a first standard avoidance time under low-speed working conditions and a second standard avoidance time under medium-to-high-speed working conditions. The second standard avoidance time is less than the first standard avoidance time. The speed of the adjacent vehicle (or the relative speed of the adjacent vehicle and the adjacent obstacle) can be determined. The speed in this embodiment is the speed of the vehicle in the driving direction (y-axis speed). Then, it is determined whether the speed of the adjacent vehicle (or the relative speed) is greater than the calibrated speed, which is a calibrated value or an empirical value based on real vehicle test data. If it is less than or equal to the calibrated speed, it indicates that the speed is low and the collision risk is not great, and the first standard avoidance time is selected as the time required for the adjacent vehicle to avoid the collision risk, i.e., the adjacent vehicle risk avoidance time. If it is greater than the calibrated speed, it indicates that the speed of the adjacent vehicle is fast and there is a certain safety risk, and the second standard avoidance time is selected as the standard collision time corresponding to the current working condition of the target vehicle.

[0125] In other embodiments, the driver reaction model includes standard evasion times in different working conditions, each of which is calibrated by real vehicle test data. When predicting the future risk of the adjacent vehicle, the road type, the relative distance between the adjacent vehicle and the adjacent obstacle, and the speed of the adjacent vehicle and the adjacent obstacle can be determined based on the road information, the state information of the adjacent vehicle, and the state information of the adjacent obstacle of the adjacent vehicle, to obtain the current working condition of the adjacent vehicle. Then, the current working condition of the adjacent vehicle is input into the driver reaction model, matched with each working condition of the driver reaction model, and the standard evasion time of the matched working condition pair is taken as the adjacent vehicle risk evasion time. By pre-calibrating the standard evasion time in different working conditions, the time required for the adjacent vehicle to evade the collision risk is determined based on the current working condition of the adjacent vehicle in actual application, so that the adjacent vehicle risk evasion time is more in line with the current working condition, and the accuracy of the adjacent vehicle risk evasion time is improved.

[0126] In other embodiments, the driver reaction model can also be a deep learning model trained based on real vehicle test data in different working conditions. When predicting the future risk of the adjacent vehicle, the road information, the state information of the adjacent vehicle, and the state information of the adjacent obstacle of the adjacent vehicle can be taken as the input of the driver reaction model, so as to predict the time required for the current adjacent vehicle to evade the collision risk through the driver reaction model, and obtain the adjacent vehicle risk evasion time, which is simple and convenient, and has high accuracy.

[0127] S123: When the first predicted collision time is less than or equal to the adjacent vehicle risk evasion time, it is determined that the adjacent vehicle will encounter a safety risk in the future.

[0128] After obtaining the adjacent vehicle risk evasion time and the first predicted collision time, the driving risk prediction device needs to determine whether the first predicted collision time is greater than the adjacent vehicle risk evasion time, and then determine whether the adjacent vehicle will encounter a safety risk from the surrounding environment in the future based on the judgment result.

[0129] When the first predicted collision time is less than or equal to the adjacent vehicle risk evasion time, it indicates that the adjacent vehicle and its adjacent obstacle may collide in the short term (such as within the adjacent vehicle risk evasion time), and it is determined that the adjacent vehicle will encounter a safety risk in the future. When the first predicted collision time is greater than the adjacent vehicle risk evasion time, it indicates that the adjacent vehicle and its adjacent obstacle will not collide in the short term, and it is determined that the adjacent vehicle will not encounter a safety risk in the future.

[0130] In this embodiment, based on the state information of the adjacent vehicle and the state information of the adjacent obstacle, the collision time required for the adjacent vehicle and the adjacent obstacle to collide in the future is predicted to obtain a first predicted collision time, and the time required for the adjacent vehicle to avoid collision risk is determined to obtain an adjacent vehicle risk avoidance time. When the first predicted collision time is less than or equal to the adjacent vehicle risk avoidance time, it is determined that the adjacent vehicle will encounter a safety risk in the future, and the specific process of determining whether the adjacent vehicle will encounter a safety risk in the future is clarified. The possibility of the adjacent vehicle encountering a safety risk in the future can be effectively predicted, the accuracy of the adjacent vehicle risk prediction is improved, and the accuracy of the self-vehicle driving risk judgment is further improved.

[0131] In an embodiment, as shown in FIG. 8, in step S30, the driving risk of the target vehicle in the future is predicted based on the predicted adjacent vehicle risk avoidance measure information, specifically including the following steps:

[0132] S31: Based on the predicted adjacent vehicle risk avoidance measure information, it is predicted whether there is a risk avoidance risk for the target vehicle in the future.

[0133] After obtaining the predicted adjacent vehicle risk avoidance measure information, the driving risk prediction device predicts whether there is a risk avoidance risk for the target vehicle in the future based on the predicted adjacent vehicle risk avoidance measure information.

[0134] Specifically, when the target vehicle and the adjacent vehicle are in the same lane, and the adjacent vehicle is located in front of the target vehicle, and the predicted adjacent vehicle risk avoidance measure information is that the adjacent vehicle brakes and stops or turns and changes lanes, it is determined that the target vehicle has a risk avoidance risk in the future. When the target vehicle and the adjacent vehicle are in different lanes, and the predicted adjacent vehicle risk avoidance measure information is that the adjacent vehicle turns and changes lanes, it is determined that the target vehicle has a risk avoidance risk in the future.

[0135] S32: Based on the environment perception data, it is predicted whether there is an environmental safety risk for the target vehicle in the future.

[0136] At the same time, based on the environment perception data, it is predicted whether there is an environmental safety risk for the target vehicle in the future. For example, whether the road conditions (pits, construction), pedestrians, etc. around the target vehicle will affect the normal driving of the target vehicle, that is, whether there is an environmental safety risk.

[0137] S33: When it is predicted that the target vehicle has a risk avoidance risk in the future and there is an environmental safety risk, it is determined that the target vehicle has a driving risk in the future.

[0138] If it is predicted that the target vehicle has a risk avoidance risk in the future and it is predicted that the target vehicle has an environmental safety risk in the future, it is determined that the target vehicle has a driving risk in the future. In other embodiments, if it is predicted that the target vehicle has a risk avoidance risk in the future or there is an environmental safety risk, it is determined that the target vehicle has a driving risk in the future.

[0139] In this embodiment, based on the predicted neighboring vehicle risk avoidance measure information, it is predicted whether there is a risk avoidance risk for the target vehicle in the future, and based on the environment perception data, it is predicted whether there is an environmental safety risk for the target vehicle in the future. When it is predicted that there is a risk avoidance risk for the target vehicle in the future and there is an environmental safety risk, it is determined that there is a driving risk for the target vehicle in the future. The driving risk of the target vehicle is judged from two dimensions of the predicted neighboring vehicle risk avoidance measure information and the environment perception data, thereby improving the accuracy of the judgment.

[0140] In an embodiment, in step S32, based on the environment perception data, it is predicted whether there is an environmental safety risk for the target vehicle in the future, which specifically includes the following steps:

[0141] S321: Based on the environment perception data, the safety influence of the surrounding environment on the target vehicle is predicted to obtain a safety parameter of the target vehicle.

[0142] S322: Based on the safety parameter, it is predicted whether there is an environmental safety risk for the target vehicle in the future.

[0143] The safety parameter can be a risk potential field of the target vehicle, which can be obtained based on the environment perception data and the risk influence of each obstacle in the surrounding environment on the target vehicle, thereby obtaining the risk potential field of the target vehicle. Then, based on the risk potential field, it is predicted whether there is an environmental safety risk for the target vehicle in the future.

[0144] The preset safety model can be a safety force field (SFF) model or a responsibility-sensitive safety (RSS) model.

[0145] In other embodiments, other safety parameters can also be used to predict whether there is an environmental safety risk for the target vehicle in the future, which will not be described here.

[0146] In this embodiment, based on the environment perception data, the safety influence of the surrounding environment on the target vehicle is predicted to obtain a safety parameter of the target vehicle, and based on the safety parameter, it is predicted whether there is an environmental safety risk for the target vehicle in the future. The step of predicting whether there is an environmental safety risk for the target vehicle in the future is refined, the environment perception data is converted into a safety parameter that can be quantified, and then the risk is predicted based on the safety parameter of the target vehicle, which is simple and has high accuracy.

[0147] In an embodiment, the safety parameter comprises a collision time of the target vehicle colliding with the adjacent vehicle. In step S321, the safety parameter of the target vehicle is obtained by predicting the safety influence of the surrounding environment on the target vehicle based on the environment perception data, specifically including the following steps:

[0148] In step S3211, it is determined whether the adjacent vehicle and the target vehicle are in the same lane based on the environment perception data.

[0149] In the embodiment, the safety parameter comprises a collision time of the target vehicle colliding with the adjacent vehicle.

[0150] The driving risk prediction device needs to determine whether the adjacent vehicle and the target vehicle are in the same lane based on the environment perception data. For example, the environment perception data comprises lane information (position information of lane lines), position information of the target vehicle, and position information of the adjacent vehicle. Based on the lane information and the position information of the two vehicles, it can be determined whether the adjacent vehicle and the target vehicle are in the same lane.

[0151] In step S3212, when it is determined that the adjacent vehicle and the target vehicle are in the same lane, a collision time required for the target vehicle to collide with the adjacent vehicle in the future is predicted to obtain a second predicted collision time.

[0152] When it is determined that the adjacent vehicle and the target vehicle are in the same lane, a collision time required for the target vehicle to collide with the adjacent vehicle in the future is predicted to obtain a second predicted collision time. The prediction process of the collision time is as described above and will not be repeated here.

[0153] The dynamics model of the target vehicle is represented by the following formula:

[0154] wherein D e represents the dynamics model of the target vehicle; v ex , v ey , v ez respectively represent the speed of the target vehicle in the x direction (vehicle driving direction), y direction (horizontal direction perpendicular to the x direction), and z direction (vertical direction); a ex , a ey , a ez respectively represent the acceleration of the target vehicle in the x direction, y direction, and z direction; j ex , j ey , j ez respectively represent the jerk of the target vehicle in the x direction, y direction, and z direction, i.e., the acceleration of the acceleration.

[0155] The second predicted collision time can be determined according to the following formula:

[0156] wherein TTCe-1 represents the second predicted collision time; P1 represents the position information (spatial coordinates) of the adjacent vehicle; P e represents the position information (spatial coordinates) of the target vehicle; L i-(i+1) represents the theoretical trajectory function of the target vehicle, which is obtained by fitting the lane line function L i with the lane line function L i+1 is obtained by median interpolation; D e represents the dynamics model of the target vehicle; D1 represents the dynamics model of the adjacent vehicle.

[0157] In this embodiment, whether the adjacent vehicle and the target vehicle are in the same lane is determined based on the environment perception data; when it is determined that the adjacent vehicle and the target vehicle are in the same lane, the collision time required for the target vehicle to collide with the adjacent vehicle in the future is predicted, and the second predicted collision time is obtained. Compared with the risk potential field data, the collision time is more accurate as a safety parameter to predict the environmental safety risk, which improves the accuracy of the environmental safety risk prediction.

[0158] In an embodiment, the safety parameter includes the collision time of the target vehicle colliding with the adjacent vehicle, which is based on the environment perception data. In step S322, the environmental safety risk of the target vehicle in the future is predicted based on the safety parameter, which specifically includes the following steps:

[0159] S3221: predicting the time required for the target vehicle to avoid the adjacent vehicle to obtain the self-vehicle risk avoidance time.

[0160] The driving risk prediction device needs to predict the time required for the target vehicle to avoid the adjacent vehicle to obtain the self-vehicle risk avoidance time. The prediction process of the self-vehicle risk avoidance time is consistent with the prediction process of the adjacent vehicle risk avoidance time, which will not be repeated here.

[0161] S3222: determining whether the target vehicle has an environmental safety risk in the future based on the second predicted collision time and the self-vehicle risk avoidance time.

[0162] Then, the driving risk prediction device determines whether the target vehicle has an environmental safety risk in the future based on the second predicted collision time and the self-vehicle risk avoidance time. The second predicted collision time is the collision time required for the target vehicle to collide with the adjacent vehicle in the future when the adjacent vehicle and the target vehicle are in the same lane.

[0163] Specifically, when the second predicted collision time is less than or equal to the self-vehicle risk avoidance time, it is determined that the target vehicle has an environmental safety risk in the future; when the second predicted collision time is greater than the self-vehicle risk avoidance time, it is determined that the target vehicle does not have an environmental safety risk in the future, and the judgment process is simple and relatively accurate.

[0164] In other embodiments, in order to improve the accuracy of the environmental safety risk judgment, a second predicted collision time of the target vehicle at different times can also be obtained to determine a change rate of the second predicted collision time; then, a preset change rate calibrated in advance is obtained, and when the second predicted collision time is less than the ego vehicle risk avoidance time and the change rate of the second predicted collision time is greater than the preset change rate, it is determined that the target vehicle has a future environmental safety risk.

[0165] In this embodiment, the ego vehicle risk avoidance time is obtained by predicting the time required for the target vehicle to avoid the adjacent vehicle, and then whether the target vehicle has a future environmental safety risk is determined based on the second predicted collision time and the ego vehicle risk avoidance time, thereby improving the accuracy of the environmental safety risk prediction and further improving the accuracy of the subsequent target vehicle driving risk prediction.

[0166] In an embodiment, before step S30, i.e., before predicting the future driving risk of the target vehicle based on the predicted adjacent vehicle risk avoidance measure information, the method further includes the following steps:

[0167] S01: Validity check is performed on the environmental perception data and the driving data of the target vehicle to obtain check information.

[0168] Before predicting the future driving risk of the target vehicle based on the predicted adjacent vehicle risk avoidance measure information, or before predicting the future risk of the adjacent vehicle of the target vehicle, the driving risk prediction device needs to perform validity check on the environmental perception data and the driving data of the target vehicle to obtain check information.

[0169] The environmental perception data includes lane information, state information of the adjacent vehicle of the target vehicle, and state information of the adjacent vehicle and the adjacent obstacle. The driving data of the target vehicle includes state information of the target vehicle.

[0170] The validity check on the environmental perception data and the driving data of the target vehicle includes: validity check on the state information of the target vehicle; validity check on the state information of the adjacent vehicle; validity check on the state information of the adjacent obstacle; and validity check on the lane information, so as to obtain the validity check result of each information data and obtain the check information.

[0171] The validity check includes check bit check, information range check, and information jump check of each information data to determine whether each information data is valid information. If any of the check bit check, information range check, and information jump check of a certain information data fails, the validity check result of the information data is invalid data; if all of the check bit check, information range check, and information jump check of the information data pass, the validity check result of the information data is valid data.

[0172] In other embodiments, the driving data of the target vehicle further includes intelligent driving system state information of the target vehicle. When the validity of the environment perception data and the driving data of the target vehicle is verified, the validity of the intelligent driving system state information also needs to be verified. Since the information output by the intelligent driving system is fault identification information, the validity verification of the intelligent driving system state information can only be bit verification, reducing the amount of verification processing.

[0173] S02: If the verification information is that all data is valid, predict the future driving risk of the target vehicle based on the predicted neighboring vehicle risk avoidance measure information.

[0174] After the validity of the environment perception data and the driving data of the target vehicle is verified, if the verification information indicates that all data is valid, the driving risk prediction device predicts the future driving risk of the target vehicle based on the predicted neighboring vehicle risk avoidance measure information.

[0175] In this embodiment, by verifying the validity of the environment perception data and the driving data of the target vehicle, the verification information is obtained. If the verification information is that all data is valid, the future driving risk of the target vehicle is predicted based on the predicted neighboring vehicle risk avoidance measure information. Before predicting the future driving risk of the target vehicle, the data validity is verified first to ensure that all data is valid before predicting the driving risk of the target vehicle. This can also avoid invalid risk prediction due to invalid data, improve the accuracy of driving risk prediction, and reduce the vehicle data processing load.

[0176] It can be understood that the size of the serial number of each step in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0177] In an embodiment, a driving risk prediction device is provided, which corresponds one-to-one to the driving risk prediction method in the above embodiments. As shown in FIG. 9, the driving risk prediction device includes a neighboring vehicle risk predictor 901 and a target vehicle risk predictor 902. The functions of each function are described in detail as follows:

[0178] The neighboring vehicle risk predictor 901 is configured to predict the future risk of a neighboring vehicle of a target vehicle based on environment perception data in a driving process of the target vehicle, to determine whether the neighboring vehicle will encounter a safety risk from the surrounding environment in the future; when it is predicted that the neighboring vehicle will encounter a safety risk in the future, predict the risk avoidance measures that can be taken by the neighboring vehicle to avoid the safety risk in the future, to obtain predicted neighboring vehicle risk avoidance measure information.

[0179] The target vehicle risk predictor 902 is configured to predict a future driving risk of the target vehicle based on the predicted neighboring vehicle risk avoidance measure information, and to perform a risk warning when it is determined that the target vehicle will encounter a driving risk in the future.

[0180] Optionally, the environment perception data comprises state information of obstacles within a preset range of the target vehicle, and the neighboring vehicle risk predictor 901 is specifically configured to:

[0181] classify and screen each target obstacle based on the road information and the state information of each obstacle, to screen out a neighboring vehicle of the target vehicle and a neighboring obstacle of the neighboring vehicle;

[0182] predict a future risk of the neighboring vehicle based on the state information of the neighboring vehicle and the state information of the neighboring obstacle, to determine whether the neighboring vehicle will encounter a safety risk in the future.

[0183] Optionally, the neighboring vehicle risk predictor 901 is further configured to:

[0184] predict a collision time required for the neighboring vehicle to collide with the neighboring obstacle in the future based on the state information of the neighboring vehicle and the state information of the neighboring obstacle, to obtain a first predicted collision time;

[0185] determine a time required for the neighboring vehicle to avoid the collision risk, to obtain a neighboring vehicle risk avoidance time;

[0186] determine that the neighboring vehicle will encounter a safety risk in the future when the first predicted collision time is less than or equal to the neighboring vehicle risk avoidance time.

[0187] Optionally, the target vehicle risk predictor 902 is specifically configured to:

[0188] predict whether there is a risk avoidance risk for the target vehicle in the future based on the predicted neighboring vehicle risk avoidance measure information;

[0189] predict whether there is an environmental safety risk for the target vehicle in the future based on the environment perception data;

[0190] determine that the target vehicle will encounter a driving risk in the future if it is predicted that there is a risk avoidance risk for the target vehicle in the future and there is an environmental safety risk.

[0191] Optionally, the target vehicle risk predictor 902 is further configured to:

[0192] determine that the target vehicle will encounter a risk avoidance risk in the future when the target vehicle and the neighboring vehicle are in the same lane, the neighboring vehicle is located in front of the target vehicle, and the predicted neighboring vehicle risk avoidance measure information is braking and parking or turning and changing lanes;

[0193] When the target vehicle and the adjacent vehicle are in different lanes, and the adjacent vehicle is located in front of the target vehicle, and the predicted adjacent vehicle risk avoidance measure information is that the adjacent vehicle is turning to change lanes, it is determined that the target vehicle has future risk avoidance risk.

[0194] Optionally, the target vehicle risk predictor 902 is specifically further configured to:

[0195] based on the environment perception data, predict the safety influence of the surrounding environment on the target vehicle to obtain a safety parameter of the target vehicle;

[0196] based on the safety parameter, predict whether the target vehicle has future environmental safety risk.

[0197] Optionally, the safety parameter includes a collision time of the target vehicle colliding with the adjacent vehicle, and based on the environment perception data, the target vehicle risk predictor 902 is specifically further configured to:

[0198] based on the environment perception data, determine whether the adjacent vehicle and the target vehicle are in the same lane;

[0199] When it is determined that the adjacent vehicle and the target vehicle are in the same lane, predict a collision time required for the target vehicle and the adjacent vehicle to have a future collision to obtain a second predicted collision time.

[0200] Optionally, the target vehicle risk predictor 902 is specifically further configured to:

[0201] predict a time required for the target vehicle to avoid the adjacent vehicle to obtain a self-vehicle risk avoidance time;

[0202] based on the second predicted collision time and the self-vehicle risk avoidance time, determine whether the target vehicle has future environmental safety risk.

[0203] Optionally, the target vehicle risk predictor 902 is specifically further configured to:

[0204] when the second predicted collision time is less than or equal to the self-vehicle risk avoidance time, it is determined that the target vehicle has future environmental safety risk; or,

[0205] when the second predicted collision time is less than the self-vehicle risk avoidance time, and the change rate of the second predicted collision time is greater than a preset change rate, it is determined that the target vehicle has future environmental safety risk.

[0206] Optionally, before predicting the future driving risk of the target vehicle based on the predicted adjacent vehicle risk avoidance measure information, the target vehicle risk predictor 902 is further configured to:

[0207] perform validity verification on the environment perception data and the driving data of the target vehicle to obtain verification information;

[0208] If the check information is all data valid, the driving risk of the target vehicle in the future is predicted based on the predicted neighboring vehicle risk avoidance measure information.

[0209] Optionally, the driving risk prediction device further comprises a path planner 903, after determining that there is a driving risk in the future of the target vehicle, the path planner 903 is configured to:

[0210] Optimize the driving path of the target vehicle.

[0211] Optionally, the path planner 903 is specifically configured to:

[0212] When the target vehicle and the neighboring vehicle are in the same lane, the target vehicle is controlled to brake and / or change lanes;

[0213] When the target vehicle and the neighboring vehicle are in different lanes, if the predicted neighboring vehicle risk avoidance measure information is a lane change by turning, the target vehicle is controlled to change to the target lane, and the target lane includes a lane other than the lane where the neighboring vehicle is located and / or the lane where the target vehicle is located.

[0214] In one embodiment, a driving risk prediction device is provided, which can be a vehicle controller. The driving risk prediction device comprises a processor, a memory and a database connected through a system bus. Among them, the processor of the driving risk prediction device is configured to provide computing and control capabilities. The memory of the driving risk prediction device comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer instructions and a database. The internal memory provides an environment for the operating system and computer instructions in the non-volatile storage medium to run. The database of the driving risk prediction device is used to store the data used or generated by the above-mentioned driving risk prediction method, such as environment perception data, etc.

[0215] In one embodiment, as shown in FIG. 10, a driving risk prediction device is provided, comprising a memory, a processor and computer instructions stored on the memory and executable on the processor, when the processor executes the computer instructions, the following steps are implemented:

[0216] Based on the environment perception data in the driving process of the target vehicle, the future risk of the neighboring vehicle of the target vehicle is predicted to determine whether the neighboring vehicle will encounter a safety risk from the surrounding environment in the future;

[0217] When it is predicted that the neighboring vehicle will encounter a safety risk in the future, the risk avoidance measures that can be taken by the neighboring vehicle to avoid the safety risk in the future are predicted, and the predicted neighboring vehicle risk avoidance measure information is obtained;

[0218] The driving risk of the target vehicle in the future is predicted based on the predicted neighboring vehicle risk avoidance measure information, and after determining that there is a driving risk in the future of the target vehicle, a risk warning is performed.

[0219] In one embodiment, a readable storage medium is provided, and computer instructions are stored on the readable storage medium, and the computer instructions are executed by a processor to implement the following steps:

[0220] Based on the environment perception data during the driving of the target vehicle, a future risk prediction is performed on the adjacent vehicle of the target vehicle to determine whether the adjacent vehicle will encounter a safety risk from the surrounding environment in the future;

[0221] When it is predicted that the adjacent vehicle will encounter a safety risk in the future, a risk avoidance measure that can be taken by the adjacent vehicle to avoid the safety risk in the future is predicted to obtain predicted adjacent vehicle risk avoidance measure information;

[0222] Based on the predicted adjacent vehicle risk avoidance measure information, a driving risk of the target vehicle in the future is predicted, and a risk warning is performed after it is determined that there is a driving risk of the target vehicle in the future.

[0223] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by computer instructions instructing related hardware, and the computer instructions can be stored in a non-volatile computer readable storage medium. When the computer instructions are executed, the processes of the above-mentioned embodiments can be included. Any reference to a memory, storage, database or other medium used in the embodiments provided in the present application can include non-volatile and / or volatile memory.

[0224] The above-described embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, a person of ordinary skill in the art can understand that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A method for predicting driving risks, wherein, include: Based on environmental perception data during the target vehicle's driving process, future risk prediction is performed on the target vehicle's neighboring vehicles to determine whether the neighboring vehicles will encounter safety risks from the surrounding environment in the future. In response to the prediction that the adjacent vehicle will encounter the safety risk in the future, the risk avoidance measures that the adjacent vehicle can take to avoid the safety risk in the future are predicted, and the information on the predicted adjacent vehicle risk avoidance measures is obtained. as well as Based on the information on the predicted neighbor vehicle avoidance measures, the future driving risk of the target vehicle is predicted, and in response to the determination that the target vehicle has a future driving risk, a risk warning is issued.

2. The driving risk prediction method as described in claim 1, wherein, The environmental perception data includes the state information of obstacles within a preset range of the target vehicle. The step of predicting future risks to neighboring vehicles of the target vehicle to determine whether these neighboring vehicles will encounter safety risks from the surrounding environment in the future includes: Based on the state information of each obstacle, the target obstacles are classified and filtered to identify adjacent vehicles of the target vehicle and adjacent obstacles of the adjacent vehicles; and Based on the status information of the adjacent vehicles and the status information of the adjacent obstacles, the future risks of the adjacent vehicles are predicted to determine whether the adjacent vehicles will encounter safety risks in the future.

3. The driving risk prediction method as described in claim 1 or 2, wherein, The prediction of future risks of adjacent vehicles to determine whether adjacent vehicles will encounter safety risks in the future includes: Based on the state information of the adjacent vehicles and the state information of the adjacent obstacles, the collision time required for the adjacent vehicles to collide with the adjacent obstacles in the future is predicted, and a first predicted collision time is obtained. Determine the time required for the adjacent vehicles to avoid collision risk, thus obtaining the adjacent vehicle risk avoidance time; and In response to the first predicted collision time being less than or equal to the neighbor vehicle risk avoidance time, it is determined that the neighbor vehicle will encounter a safety risk in the future.

4. The driving risk prediction method as described in any one of claims 1-3, wherein, The prediction of future driving risks of the target vehicle based on the predicted neighbor vehicle avoidance measures information includes: Based on the predicted neighbor vehicle risk avoidance measures information, predict whether there is a risk of risk avoidance for the target vehicle in the future; Based on the environmental perception data, predict whether the target vehicle will pose a future environmental safety risk; and In response to the prediction that the target vehicle will face future risks of hazard avoidance and environmental safety, it is determined that the target vehicle will face future driving risks.

5. The driving risk prediction method as described in claim 4, wherein, The prediction of whether there is a risk of danger to the target vehicle in the future based on the predicted neighbor vehicle avoidance measures information includes: In response to the target vehicle and the adjacent vehicle being in the same lane, and the adjacent vehicle being in front of the target vehicle, and the predicted adjacent vehicle avoidance measure information being braking to a stop or turning to change lanes, it is determined that the target vehicle has a future risk of danger.

6. The driving risk prediction method as described in claim 4 or 5, wherein, The prediction of whether there is a risk of danger to the target vehicle in the future based on the predicted neighbor vehicle avoidance measures information includes: In response to the target vehicle and the adjacent vehicle being in different lanes, and the predicted adjacent vehicle avoidance measure information being a lane change, it is determined that the target vehicle faces a future risk of danger.

7. The driving risk prediction method according to any one of claims 4-6, wherein, The prediction of whether the target vehicle will face future environmental safety risks based on the environmental perception data includes: Based on the environmental perception data, the safety impact of the surrounding environment on the target vehicle is predicted, and the safety parameters of the target vehicle are obtained; and Based on the aforementioned safety parameters, it is predicted whether the target vehicle will pose any future environmental safety risks.

8. The driving risk prediction method as described in claim 7, wherein, The safety parameters include the collision time between the target vehicle and the adjacent vehicle. The process of predicting the impact of the surrounding environment on the safety of the target vehicle based on the environmental perception data, to obtain the safety parameters of the target vehicle, includes: Based on the environmental perception data, determine whether the adjacent vehicle and the target vehicle are in the same lane; and In response to determining that the adjacent vehicle and the target vehicle are in the same lane, the collision time required for a future collision between the target vehicle and the adjacent vehicle is predicted to obtain a second predicted collision time.

9. The driving risk prediction method as described in claim 7 or 8, wherein, The prediction of whether the target vehicle poses a future environmental safety risk based on the aforementioned safety parameters includes: Predict the time required for the target vehicle to avoid the adjacent vehicle, and obtain the risk avoidance time of the vehicle itself; and Based on the second predicted collision time and the vehicle's risk avoidance time, it is determined whether the target vehicle will face environmental safety risks in the future. The second predicted collision time is the time required for the target vehicle and the adjacent vehicle to collide in the future when the target vehicle and the adjacent vehicle are in the same lane.

10. The driving risk prediction method as described in claim 9, wherein, The step of determining whether the target vehicle poses a future environmental safety risk based on the second predicted collision time and the vehicle's risk avoidance time includes: In response to the second predicted collision time being less than or equal to the vehicle risk avoidance time, it is determined that the target vehicle will face environmental safety risks in the future.

11. The driving risk prediction method as described in claim 9 or 10, wherein, The step of determining whether the target vehicle poses a future environmental safety risk based on the second predicted collision time and the vehicle's risk avoidance time includes: In response to the second predicted collision time being less than the vehicle's risk avoidance time, and the rate of change of the second predicted collision time being greater than a preset rate of change, it is determined that the target vehicle will face environmental safety risks in the future.

12. The driving risk prediction method according to any one of claims 1-11, wherein, Before predicting the future driving risk of the target vehicle based on the predicted neighbor vehicle avoidance measures information, the method further includes: The validity of the environmental perception data and the driving data of the target vehicle are verified to obtain verification information; and In response to the verification information confirming that all data is valid, the future driving risk of the target vehicle is predicted based on the predicted neighbor vehicle avoidance measures information.

13. The driving risk prediction method according to any one of claims 1-12, wherein, In response to determining that the target vehicle poses a future driving risk, the method further includes: The driving path of the target vehicle is optimized.

14. The driving risk prediction method as described in claim 13, wherein, The optimization of the target vehicle's travel path includes: In response to the target vehicle being in the same lane as the adjacent vehicle, the target vehicle is controlled to brake.

15. The driving risk prediction method as described in claim 13 or 14, wherein, The optimization of the target vehicle's travel path includes: In response to the target vehicle being in the same lane as the adjacent vehicle, the target vehicle is controlled to change lanes.

16. The driving risk prediction method according to any one of claims 13-15, wherein, The optimization of the target vehicle's travel path includes: In response to the target vehicle and the adjacent vehicle being in different lanes, and the predicted adjacent vehicle avoidance measure information being a lane change, the target vehicle is controlled to change to a target lane, the target lane including the lane where the adjacent vehicle is located and / or a lane other than the lane where the target vehicle is located.

17. A driving risk prediction device, wherein, include: The neighbor vehicle risk predictor is used to predict the future risks of neighboring vehicles based on environmental perception data during the driving process of the target vehicle, so as to determine whether the neighboring vehicles will encounter safety risks from the surrounding environment in the future; in response to predicting that the neighboring vehicles will encounter the safety risks in the future, it predicts the avoidance measures that the neighboring vehicles can take to avoid the safety risks in the future, and obtains the predicted neighboring vehicle avoidance measures information. as well as The target vehicle risk predictor is used to predict the future driving risk of the target vehicle based on the predicted neighbor vehicle avoidance measures information, and to issue a risk warning in response to determining that the target vehicle has a future driving risk.

18. A driving risk prediction device, comprising a memory, a processor, and computer instructions stored in the memory and executable on the processor, wherein, When the processor executes the computer instructions, it implements the steps of the driving risk prediction method as described in any one of claims 1 to 16.

19. A driving risk prediction system applied to the driving risk prediction method as described in any one of claims 1-16, wherein, Includes the target vehicle and the driving risk prediction device as described in claim 17 or 18.

20. A readable storage medium for use in the driving risk prediction method as described in any one of claims 1-16, the readable storage medium storing computer instructions, wherein, When the computer instructions are executed by the processor, they implement the steps of the driving risk prediction method as described in any one of claims 1 to 16.

Images