A driving scene recognition method, device and equipment

Abstract

The present disclosure provides a driving scene recognition method, device and equipment, comprising: acquiring real-time parameter information of a vehicle in a driving process and road information monitored based on a high-definition map; determining a traffic action of the vehicle according to the real-time parameter information; and identifying a type of a driving scene of the vehicle according to the traffic action and the road information. The present disclosure can quickly and accurately identify the type of the driving scene of the vehicle, and improve the safety of driving.

Description

Technical Field

[0001] This disclosure belongs to the field of autonomous driving simulation technology, and specifically refers to a method, device and equipment for recognizing driving scenarios. Background Technology

[0002] Autonomous driving technology is currently undergoing rapid development, which has also brought about many important problems that urgently need to be solved. Among them, how to test autonomous driving algorithms more thoroughly at a lower cost is a difficult challenge.

[0003] One existing solution to this problem is to use autonomous driving simulation software to replace real vehicle testing, using driving cases collected by various data collection vehicles as test scenario inputs. However, there is currently no good solution for how to efficiently analyze, mine, classify, and identify dangerous and rare cases when faced with tens of thousands of real vehicle driving data.

[0004] Some big data systems can perform some degree of behavioral analysis, but they are usually limited to simple actions such as acceleration or deceleration, lane changing, and cutting in front of other vehicles; their ability to explore other complex scenarios is very limited. Summary of the Invention

[0005] The technical problem to be solved by this disclosure is to provide a method, device and equipment for identifying driving scenarios, so as to quickly and accurately identify the type of driving scenario of a vehicle and improve driving safety.

[0006] To address the aforementioned technical problems, embodiments of this disclosure provide a method for recognizing driving scenarios, including:

[0007] Acquire real-time parameter information of the vehicle during driving and road information monitored based on high-precision maps;

[0008] Based on real-time parameter information, determine the vehicle's traffic actions;

[0009] Based on traffic movements and road information, identify the type of driving scenario for the vehicle.

[0010] Optionally, based on real-time parameter information, determine the vehicle's traffic actions, including:

[0011] Based on real-time parameter information, determine the vehicle's atomic traffic actions and / or combinations of traffic actions formed by at least two atomic traffic actions.

[0012] Optionally, both atomic traffic actions and combinations of traffic actions include at least one of the following attribute information:

[0013] The subject of the action;

[0014] Action type;

[0015] The start and end times of the action;

[0016] Statistical attributes;

[0017] Identifier attributes.

[0018] Optionally, if traffic actions include atomic traffic actions, the type of vehicle driving scenario is identified based on the traffic actions and road information, including:

[0019] Based on the first preset screening criteria, the target atomic traffic action is determined;

[0020] Based on the target atom's traffic actions and road information, the vehicle's driving scenario is determined to be the first type of driving scenario.

[0021] Optionally, if traffic actions include combinations of traffic actions, the type of driving scenario for the vehicle is identified based on the traffic actions and road information, including:

[0022] Based on the second preset screening criteria, the target action pairs in the traffic action combination are determined;

[0023] Based on the target action pair and road information, the vehicle's driving scenario is determined to be the second type of driving scenario.

[0024] Optionally, based on the second preset filtering criteria, target action pairs in the traffic action combination are determined, including:

[0025] If a traffic action combination includes N following actions and M braking actions, and the first identification information of any vehicle in the N following actions is the same as the second identification information of any vehicle in the M braking actions, and the time intervals of the two actions overlap, then the current two actions are determined as the first target action pair, where N and M are both positive integers; or

[0026] If the traffic action combination includes: M braking actions and S lane changing actions; when the second identification information of any vehicle in the M braking actions is the same as the third identification information of any vehicle in the S lane changing actions, and the start time of the two actions is within a preset range, the current two actions are determined as the second target action pair, where S is a positive integer; or

[0027] If the traffic action combination includes Y straight-ahead actions and Z left-turn actions, when the fourth sign information of the intersection where any of the main vehicles in the Y straight-ahead actions are located is the same as the fifth sign information of the intersection where any of the main vehicles in the Z left-turn actions are located, and the time intervals of the two actions overlap, the current two actions are determined to be the third target action pair, where Y and Z are both positive integers.

[0028] Optional methods for recognizing driving scenarios also include:

[0029] The driving behavior is evaluated based on the type of driving scenario identified for the vehicle.

[0030] Embodiments of this disclosure also provide a driving scene recognition device, including:

[0031] The acquisition module is used to acquire real-time parameter information of the vehicle during driving and road information monitored based on high-precision maps;

[0032] The processing module is used to determine the vehicle's traffic actions based on real-time parameter information; and to identify the type of driving scenario based on the traffic actions and road information.

[0033] Optionally, the processing module determines the vehicle's traffic actions based on real-time parameter information, specifically for:

[0034] Based on real-time parameter information, determine the vehicle's atomic traffic actions and / or combinations of traffic actions formed by at least two atomic traffic actions.

[0035] Optionally, both atomic traffic actions and combinations of traffic actions include at least one of the following attribute information:

[0036] The subject of the action;

[0037] Action type;

[0038] The start and end times of the action;

[0039] Statistical attributes;

[0040] Identifier attributes.

[0041] Optionally, if traffic actions include atomic traffic actions, the type of vehicle driving scenario is identified based on the traffic actions and road information, including:

[0042] Based on the first preset screening criteria, the target atomic traffic action is determined;

[0043] Based on the target atom's traffic actions and road information, the vehicle's driving scenario is determined to be the first type of driving scenario.

[0044] Optionally, if traffic actions include combinations of traffic actions, the type of driving scenario for the vehicle is identified based on the traffic actions and road information, including:

[0045] Based on the second preset screening criteria, the target action pairs in the traffic action combination are determined;

[0046] Based on the target action pair and road information, the vehicle's driving scenario is determined to be the second type of driving scenario.

[0047] Optionally, based on the second preset filtering criteria, target action pairs in the traffic action combination are determined, specifically for:

[0048] If a traffic action combination includes N following actions and M braking actions, and the first identification information of any vehicle in the N following actions is the same as the second identification information of any vehicle in the M braking actions, and the time intervals of the two actions overlap, then the current two actions are determined as the first target action pair, where N and M are both positive integers; or

[0049] If the traffic action combination includes: M braking actions and S lane changing actions; when the second identification information of any vehicle in the M braking actions is the same as the third identification information of any vehicle in the S lane changing actions, and the start time of the two actions is within a preset range, the current two actions are determined as the second target action pair, where S is a positive integer; or

[0050] If the traffic action combination includes Y straight-ahead actions and Z left-turn actions, when the fourth sign information of the intersection where any of the main vehicles in the Y straight-ahead actions are located is the same as the fifth sign information of the intersection where any of the main vehicles in the Z left-turn actions are located, and the time intervals of the two actions overlap, the current two actions are determined to be the third target action pair, where Y and Z are both positive integers.

[0051] Optionally, the processing module is also specifically used to: evaluate driving behavior based on the type of driving scenario of the identified vehicle.

[0052] Embodiments of this disclosure also provide a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the methods described above.

[0053] Embodiments of this disclosure also provide a computer-readable storage medium including instructions that, when executed on a computer, cause the computer to perform the methods described above.

[0054] The above-described solution disclosed herein includes at least the following beneficial effects:

[0055] By acquiring real-time parameter information of the vehicle during driving and road information monitored based on high-precision maps, and determining the vehicle's traffic actions based on the real-time parameter information, the type of driving scenario of the vehicle can be identified based on the traffic actions and real-time parameter information. This enables the rapid and accurate identification of the type of driving scenario, and further analysis can be performed based on the identified driving scenario type to improve driving safety. Attached Figure Description

[0056] Figure 1 This is a flowchart of a driving scenario recognition method provided in an embodiment of this disclosure;

[0057] Figure 2This is a schematic diagram of a target action pair provided in an optional embodiment of this disclosure;

[0058] Figure 3 This is a schematic diagram of another target action pair provided in an optional embodiment of this disclosure;

[0059] Figure 4 This is a schematic diagram of another target action pair provided in an optional embodiment of this disclosure;

[0060] Figure 5 This is a schematic diagram of the module block of the driving scene recognition device provided in the embodiments of this disclosure. Detailed Implementation

[0061] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0062] In the field of autonomous driving scenario simulation, there are related terms for driving scenarios. In the following embodiments of this disclosure, the technical terms are explained as follows:

[0063] 1. Traffic participants (hereinafter referred to as Actors): refers to the subjects involved in the driving scenario, such as motor vehicles, bicycles, or pedestrians.

[0064] 2. Opposite vehicle: refers to other motor vehicles besides the main vehicle.

[0065] 3. Traffic Action (hereinafter referred to as Action): This describes an action completed by one or more traffic participants, either independently or collaboratively, within a certain period of time. For example, if vehicle number 1 completes a left lane change within 2-5 seconds, this is a left lane change traffic action. Or, if vehicle number 1 is in motion when a pedestrian crosses the road in front of it within 1-10 seconds, this is a "pedestrian crossing the road in front of the vehicle in motion" traffic action.

[0066] In the following embodiments of this disclosure, traffic actions are divided into two categories: one is basic, atomic traffic actions, such as lane changing and vehicle movement, while the other is traffic actions that can be composed of combinations of other traffic actions, such as the front vehicle braking being composed of following and emergency braking.

[0067] To address the problem of existing technologies being unable to quickly and accurately identify the type of vehicle driving scenario, such as... Figure 1 As shown, embodiments of this disclosure provide a method for recognizing driving scenarios, including:

[0068] Step 11: Obtain real-time parameter information of the vehicle during driving and road information monitored based on high-precision maps;

[0069] Step 12: Determine the vehicle's traffic actions based on real-time road parameter information;

[0070] Step 13: Identify the type of driving scenario for the vehicle based on traffic movements and road information.

[0071] In this embodiment, real-time parameter information of the vehicle during driving is obtained. Here, the real-time parameter information may include: the GPS location information, speed, direction, acceleration, distance to preset marker objects, and other real-time parameter information of the current autonomous vehicle. It should be noted that each type of parameter information in the real-time parameter information corresponds to corresponding time information.

[0072] Road information monitored based on high-precision maps can include: the type of road the current autonomous vehicle is on (e.g., whether it is a highway, urban road or rural road), the shape of the road (e.g., whether there are curves), the condition of the road (e.g., whether there is construction or traffic congestion), information on surrounding traffic participants (e.g., other vehicles, pedestrians, bicycles, etc.), and the condition of traffic signs and traffic lights.

[0073] Furthermore, based on real-time vehicle parameter information, the vehicle's traffic actions are determined. For example, if the vehicle's speed suddenly decreases, it may mean that the vehicle is braking; if the vehicle's direction changes, it may mean that the vehicle is turning or changing lanes. These traffic actions can be actions completed by one or more vehicles within a preset time period. For example, if vehicle number 1 completes a left lane change within 2-5 seconds, this is a left lane change traffic action; or if vehicle number 1 is driving smoothly when a pedestrian crosses the road ahead, this is a "pedestrian crossing the road ahead while the vehicle is in motion" traffic action.

[0074] Furthermore, based on the determined traffic actions of the current vehicle, and taking into account current road information (e.g., whether there are pedestrians crossing the road in front of the vehicle, whether other vehicles nearby will brake), the driving scenario of the current vehicle can be accurately identified. For example, if the vehicle makes continuous lane changes on a highway and exceeds the speed limit, the current driving scenario can be identified as "dangerous driving on a highway"; if the vehicle is on an urban road and there are pedestrians crossing the road in front, and the vehicle brakes, the current driving scenario can be identified as "emergency braking on an urban road".

[0075] The above method can be used to define the vehicle's traffic actions based on real-time parameter information. Furthermore, based on traffic actions and road information, the type of the vehicle's current driving scenario can be accurately identified, thereby improving road driving safety, optimizing the autonomous driving experience, and promoting the development of autonomous driving technology.

[0076] In an optional embodiment of this disclosure, step 12 above may include:

[0077] Step 121: Based on real-time parameter information, determine the vehicle's atomic traffic actions and / or traffic action combinations formed by at least two atomic traffic actions.

[0078] Here, atomic traffic actions include at least one of the following: changing lanes, following another vehicle, accelerating, decelerating, turning, and going straight.

[0079] Furthermore, both atomic traffic actions and combinations of traffic actions include at least one of the following attribute information: action subject; action type; start and end time of action completion; statistical attribute; and identification attribute.

[0080] Here, the subject of the action refers to the entity participating in the driving scenario: such as a motor vehicle, bicycle, or pedestrian; in this embodiment, the subject of the corresponding atomic traffic action and the action combination of traffic actions can both refer to a vehicle.

[0081] Action type represents different atomic traffic actions or different combinations of traffic actions formed by at least two atomic traffic actions;

[0082] The start and end times of an action represent the start and end times of any atomic traffic action or combination of traffic actions. Here, the start and end times of an action can be calculated as the time interval during which the corresponding atomic traffic action or combination of traffic actions continuously meets specific conditions. For example, the time interval during which the vehicle's acceleration is continuously less than -5 is the time interval for emergency braking.

[0083] Statistical attributes can represent statistical data of vehicles within a preset monitoring period, such as average speed and average acceleration within the preset monitoring period; they can also represent the actual data of current traffic actions at the corresponding moment, such as real-time speed and real-time acceleration.

[0084] Identification attributes can represent attributes specific to any atomic traffic action or combination of traffic actions, such as lane changing having an identification attribute for the direction of lane change, and following another vehicle having an identification attribute for the average following distance, etc.

[0085] In this embodiment, the traffic actions of the corresponding vehicle may include: atomic traffic actions and / or combinations of traffic actions formed by at least two atomic traffic actions; here, an atomic action is the most basic component of an action generated during vehicle driving, and cannot be further subdivided; during vehicle driving, atomic traffic actions may include: lane changing, following, acceleration, deceleration, steering, etc.

[0086] Based on the real-time parameter information of the vehicle during the driving process and the preset judgment conditions, the real-time parameter information is analyzed to determine the traffic action of the current vehicle, thereby ensuring the accuracy of the identification of the driving scenario type of subsequent vehicles.

[0087] For the actions of vehicles when passing through intersections, based on road information obtained from high-precision maps, information on each road connected to each intersection is first extracted. By calculation, if there are 4 roads connected to an intersection, and they are arranged at 90-degree intervals (with an error of less than 10 degrees), it can be considered a crossroads. Then, the roads that the vehicle passes through when entering and leaving the intersection are judged to obtain the angle between the entering and leaving roads, and then the various atomic traffic actions of the vehicle at the intersection are identified, such as left turn, right turn, straight, U-turn, etc.

[0088] In one feasible example of this disclosure, the lane-changing behavior of a vehicle is judged based on real-time parameter information: taking left lane changing as an example, if the distance between the detected vehicle and the left lane line (that is, the distance to the preset marker object) continues to decrease within a certain period of time, the vehicle is moving to the left; otherwise, the detection process ends; and if the lane where the main vehicle is currently located is different from the lane it was in last detection while the vehicle is moving to the left, and the two lanes are adjacent lanes, then it is considered that the vehicle has crossed the lane, and this behavior is a lane-changing behavior.

[0089] In one feasible example of this disclosure, the following behavior of a vehicle is judged based on real-time parameter information: if, within a preset monitoring period, preferably longer than 3 seconds, the vehicle in front of vehicle S is always the same vehicle T, and the distance between vehicle S and vehicle T is always less than a preset distance, preferably 50 meters, then vehicle S is in following behavior within the preset monitoring period.

[0090] In an optional embodiment of this disclosure, if the traffic action includes an atomic traffic action, step 13 above may include:

[0091] Step 131a: Determine the target atom traffic action according to the first preset screening conditions;

[0092] Step 132a: Based on the target atom traffic actions and road information, determine that the vehicle's driving scenario is the first type of driving scenario.

[0093] In this embodiment, the first type of driving scenario represents a set of simple driving scenarios that only contain atomic traffic actions, which may include a variety of different simple driving scenarios; the preset filtering conditions may include: action type, statistical attributes, etc. It should be understood that the corresponding first preset filtering conditions should be different for different simple driving scenarios; for example, for a specific simple driving scenario, only atomic traffic actions that conform to the type of the scenario are filtered out, such as vehicle lane changing, vehicle acceleration, etc. At the same time, whether the statistical attributes of the atomic traffic actions in the scenario, such as vehicle speed, meet the requirements of the preset threshold.

[0094] Furthermore, based on the target atomic traffic action determined by the first preset screening conditions, as well as other traffic participants in the road information (such as the position and status of pedestrians and other vehicles, such as whether pedestrians are on the zebra crossing, whether the vehicle in front is braking, whether the speed of surrounding vehicles has changed significantly, etc.), road conditions (such as whether it is near a school, sidewalk, zebra crossing, etc., the smoothness or roughness of the road surface, etc.), the status of traffic signs and traffic lights, etc., the current driving scenario of the vehicle is determined to be the first type of driving scenario.

[0095] Taking the first type of driving scenario as "dangerous driving scenario on highway" as an example, the first preset screening conditions may include: action type: the vehicle suddenly changes lanes or brakes suddenly; vehicle speed: the vehicle speed exceeds the speed limit of the highway; the status of other vehicles: the vehicle in front is braking or its speed has changed significantly; the driving environment in the new road information is: on highway, busy traffic, etc.; the driving scenario of the vehicle corresponding to the atomic traffic action that meets all the above conditions is regarded as "dangerous driving scenario on highway".

[0096] In an optional embodiment of this disclosure, if the traffic action includes a combination of traffic actions, step 13 above may include:

[0097] Step 131b: Determine the target action pair in the traffic action combination according to the second preset screening conditions;

[0098] Step 132b: Based on the target action pair and road information, determine that the vehicle's driving scenario is the second type of driving scenario.

[0099] In this embodiment, the second type of driving scenario represents a set of complex driving scenarios that include combinations of traffic actions, which may include a variety of different complex driving scenarios; when at least two atomic traffic actions in the combination of traffic actions are paired to form a target action pair, the two atomic traffic actions in the target action pair should satisfy the second preset screening condition; it should be understood that for different complex driving scenarios, the corresponding target action pair and the corresponding second preset screening condition should be different.

[0100] Here, the second preset filtering criteria may include: the temporal relationship, spatial relationship, and vehicle status between atomic traffic actions;

[0101] Regarding the temporal relationship between two atomic traffic actions: Generally speaking, the temporal sequence and interval between two atomic traffic actions are very important for determining whether they constitute a more complex target action pair; for example, a braking action immediately following a following action can be interpreted as "the car in front brakes suddenly"; if the interval between two lane-changing actions is short, it can be interpreted as "continuous lane changing"; for a target action pair, the time interval between the two atomic traffic actions should meet a preset interval threshold.

[0102] Regarding the spatial relationship between two atomic traffic actions: the combination of some target action pairs may depend on the spatial relationship, such as the distance and relative position of the two vehicles; for example, when the distance between two vehicles is too close, the combination of braking and following actions may be regarded as "emergency braking"; for a target action pair, the distance between the vehicles corresponding to the two atomic traffic actions should meet the preset distance threshold.

[0103] The vehicle states corresponding to two atomic traffic actions, such as the vehicle's speed, direction, and braking system status, can all affect the combination of the target action pairs.

[0104] Preferably, the target conditions may also include: driving environment conditions, driver behavior and state; among them, driving environment conditions, for example, in rainy or snowy weather, even at relatively slow speeds, braking and following actions may be interpreted as "emergency braking"; driver behavior and state may also affect the combination of target action pairs; for example, a driver frequently checking the rearview mirror when changing lanes continuously may be judged as "nervous driving".

[0105] Furthermore, based on the determined target action pair and other traffic participants in the road information (such as the position and status of pedestrians and other vehicles, such as whether pedestrians are on the zebra crossing, whether the vehicle in front is braking, whether the speed of surrounding vehicles has changed significantly, etc.), road conditions (such as whether it is near a school, sidewalk, zebra crossing, etc., the smoothness or roughness of the road surface, etc.), the status of traffic signs and traffic lights, the current driving scenario of the vehicle is determined to be the second type of driving scenario.

[0106] By combining and matching atomic traffic actions with the second preset screening conditions and corresponding road information, more complex second-type driving scenarios can be identified and understood.

[0107] In an optional embodiment of this disclosure, step 131b described above may include:

[0108] Step 131b-1: If the traffic action combination includes: N following actions and M braking actions, when the first identification information of any preceding vehicle in the N following actions is the same as the second identification information of any leading vehicle in the M braking actions, and the time intervals of the two actions overlap, determine the current two actions as the first target action pair, where N and M are both positive integers; or

[0109] Step 131b-2: If the traffic action combination includes: M braking actions and S lane changing actions; when the second identification information of any primary vehicle in the M braking actions is the same as the third identification information of any primary vehicle in the S lane changing actions, and the start time of the two actions is within a preset range, determine the current two actions as the second target action pair, where S is a positive integer; or

[0110] Step 131b-13: If the traffic action combination includes Y straight-ahead actions and Z left-turn actions, when the fourth sign information of the intersection where any of the main vehicles in the Y straight-ahead actions are located is the same as the fifth sign information of the intersection where any of the main vehicles in the Z left-turn actions are located, and the time intervals of the two actions overlap, the current two actions are determined to be the third target action pair, where Y and Z are both positive integers.

[0111] In this embodiment, at least two atomic traffic actions in the traffic action combination are sequentially paired one by one to obtain at least one target action pair, and at least one target action pair satisfies the second preset screening condition; specifically:

[0112] If a traffic action combination includes M braking actions and N following actions, then each of the N following actions and M braking actions is paired one by one. Only when the first identification information of the preceding vehicle in the following action and the second identification information of the leading vehicle in the braking action are the same, and the time intervals of the two actions overlap, are these two actions considered to form a new first target action pair, i.e., a following-braking target action pair. Here, both the first and second identification information can be vehicle ID information. Figure 2 As shown, vehicle A is braking and vehicle B is following. When the second identification information of vehicle A is the same as the first identification information of vehicle B, and the time intervals of their actions overlap, the braking action of vehicle A and the following action of vehicle B can be combined into a following-braking target action pair.

[0113] If the traffic action combination includes: M braking actions and S lane-changing actions; pair the M braking actions and S lane-changing actions in the traffic action combination one by one. Only when the third identification information of the host vehicle in the lane-changing action is the same as the second identification information of the host vehicle in the preceding vehicle's braking action; and the start time point t1 of the lane-changing action and the start time point t2 of the braking action satisfy t1 > t2, and at the same time t1 - t2 < t0, it is confirmed that these two actions can form a new second target action pair, that is, a braking-lane-changing target action pair; here, the third identification information can be the vehicle's ID information, and t0 is a preset time threshold (which can be set according to the actual situation); for example Figure 3 As shown, vehicle C is in a braking action and vehicle D is in a lane-changing action. When the second identification information of vehicle C is the same as the third identification information of vehicle D, and the starting time t1 corresponding to vehicle D's lane-changing action is greater than the starting time t2 corresponding to vehicle C's braking action, and the difference between their corresponding starting times is less than a preset time threshold, then the braking action of vehicle C and the lane-changing action of vehicle D can be combined into: a braking-lane-changing target action pair;

[0114] If the traffic action combination includes: Y straight-moving actions and Z left-turning actions, pair the Y straight-moving actions and Z left-turning actions in the traffic action combination one by one. Only when any fourth identification information of the intersection where the host vehicle is located in the straight-moving action is the same as any fifth identification information of the intersection where the host vehicle is located in the left-turning action, and the time intervals of the two actions overlap, it is determined that the current two actions are the third target action pair, that is, a straight-moving-turning target action pair; here, both the fourth identification information and the fifth identification information can be the ID information of the intersection; for example Figure 4 As shown, vehicle E is in a straight-moving action and vehicle F is in a left-turning action. When the fourth identification information of the intersection where vehicle E is located is the same as the fifth identification information of the intersection where vehicle F is to turn into, and the time intervals of their actions overlap, then the straight-moving action of vehicle E and the left-turning action of vehicle F can be combined into: a straight-moving-turning target action pair;

[0115] It should be known that the target action pairs are not limited to the target action pairs described in the above embodiments. For the atomic traffic actions in the traffic action combination that meet the second preset screening conditions, different target action pairs can be formed to identify different and complex driving scenario types; for example: combine a following action and a reverse action to form a target action pair of "a vehicle in front in this lane is reversing". During the time periods of these two atomic traffic actions, it is required that when the host vehicle in the following action is driving, there is at least one vehicle with the same ID information reversing in front for at least N consecutive seconds, where N is a positive integer;

[0116] The above embodiments of this disclosure pair different atomic traffic actions in a traffic action combination one by one and combine them into target action pairs that meet the second preset screening conditions, so as to identify more complex driving scenarios, so as to make warnings or decisions based on complex driving scenarios and improve driving safety.

[0117] Furthermore, step 132b above may include:

[0118] Step 132b-1: Based on at least one first target action pair and road information, determine that the vehicle's driving scenario is a second type of driving scenario; or

[0119] Step 132b-2: Based on at least one second target action pair and road information, determine that the vehicle's driving scenario is a second type of driving scenario; or

[0120] Step 132b-3: Based on at least one third target action pair and road information, determine that the vehicle's driving scenario is a second type of driving scenario.

[0121] In this embodiment, the first type of driving scenario, the second type of driving scenario, and the third type of driving scenario are all complex driving scenarios; different types of complex driving scenarios are identified according to different target actions to ensure the accuracy of complex driving scenario identification.

[0122] In an optional embodiment of this disclosure, based on steps 11 to 13 above, the driving scenario recognition method may further include:

[0123] Step 14: Evaluate the driving behavior based on the type of driving scenario of the identified vehicle.

[0124] In this embodiment, subsequent driving behaviors are evaluated or analyzed based on the identified type of driving scenario, and the current driving behavior can be readjusted or decided based on the evaluation or analysis results to ensure driving safety.

[0125] Here, readjusting or deciding on current driving behavior based on the results of the assessment or analysis may include at least one of the following:

[0126] Driving behavior scoring: Based on the identified type of driving scenario, the driver's behavior can be further scored; for example, if the identified driving scenario is "dangerous driving on a highway", the driver's behavior can be assessed as dangerous driving and a corresponding driving score can be given.

[0127] Hazard warning: Based on the type of driving scenario identified, real-time hazard warnings can be provided to the driver; for example, if the type of driving scenario identified indicates a dangerous driving situation that the driver may face, a warning can be issued through the vehicle information system to remind the driver to slow down, increase the distance from the vehicle in front, etc.

[0128] Driving guidance: Based on the identified driving scenario type, driving guidance suggestions can be provided to help drivers improve their driving habits; for example, if a driver does not react quickly enough in certain driving scenarios (such as highways, busy urban roads, etc.), corresponding driving training can be provided, such as earlier warnings and improved driving attention.

[0129] Autonomous driving decision support: In autonomous driving scenarios, the identified driving scenarios can help the autonomous driving system make more accurate decisions; for example, when an emergency braking scenario on an urban road is identified, the autonomous driving system can decide to immediately slow down or stop to avoid potential dangers.

[0130] The above embodiments of this disclosure, starting with the vehicle's traffic actions determined based on the acquired basic road information and real-time parameter information, can identify complex driving scenarios for the vehicle; furthermore, based on the identified driving scenarios, driving behavior is evaluated and autonomous driving decisions are supported, which can be applied to a driving behavior monitoring and management system.

[0131] Furthermore, the monitoring and management system for driving behavior can be further optimized to further improve driving safety.

[0132] The monitoring and management system for driving behavior has been further optimized, as detailed below:

[0133] Driving habit modeling and prediction: By collecting and analyzing driving behavior data, driving habit models can be established. These models can predict possible driver behaviors in certain scenarios, enabling earlier warnings and interventions to ensure driving safety.

[0134] Driving Behavior Anomaly Detection: In daily driving, if the driver's behavior is abnormal (such as being more aggressive or overly cautious than usual), it may indicate certain problems, such as driver fatigue, illness, or distraction. The system can monitor the driver's behavior in real time, and once an anomaly is detected, it will immediately issue a reminder or intervene to ensure driving safety.

[0135] Intelligent analysis of road and traffic conditions: In addition to analyzing driver behavior, intelligent analysis of road and traffic conditions can also be performed; for example, by analyzing road characteristics (such as curves, slopes, etc.) and real-time traffic information, potential dangerous situations can be predicted and prepared in advance to ensure driving safety.

[0136] Data sharing and cloud analytics: The system can upload collected data (such as driving behavior data, vehicle status data, road condition data, etc.) to the cloud for storage and analysis, conduct in-depth analysis of driving behavior, and provide drivers with more personalized and accurate services.

[0137] The above embodiments of this disclosure improve driving safety by acquiring real-time parameter information of the vehicle during driving; acquiring road information monitored based on high-precision maps; determining the vehicle's traffic actions based on the real-time parameter information; and quickly and accurately identifying the type of driving scenario based on the traffic actions and road information.

[0138] like Figure 5 As shown, embodiments of this disclosure also provide a driving scene recognition device 50, including:

[0139] The acquisition module 51 is used to acquire real-time parameter information of the vehicle during the driving process and road information monitored based on high-precision maps;

[0140] The processing module 52 is used to determine the vehicle's traffic actions based on real-time parameter information; and to identify the type of driving scenario of the vehicle based on the traffic actions and road information.

[0141] Optionally, the processing module 52 determines the vehicle's traffic actions based on real-time parameter information, including: determining the vehicle's atomic traffic actions and / or combinations of traffic actions formed by at least two atomic traffic actions based on real-time parameter information.

[0142] Optionally, both atomic traffic actions and combinations of traffic actions include at least one of the following attribute information: action subject; action type; start and end time of action completion; statistical attribute; and identification attribute.

[0143] Optionally, if the traffic action includes atomic traffic actions, the processing module 52 identifies the type of vehicle driving scenario based on the traffic action and road information, including:

[0144] Based on the first preset screening criteria, the target atomic traffic action is determined;

[0145] Based on the target atom's traffic actions and road information, the vehicle's driving scenario is determined to be the first type of driving scenario.

[0146] Optionally, if the traffic actions include combinations of traffic actions, the processing module 52 identifies the type of driving scenario for the vehicle based on the traffic actions and road information, including:

[0147] Based on the second preset screening criteria, the target action pairs in the traffic action combination are determined;

[0148] Based on the target action pair and road information, the vehicle's driving scenario is determined to be the second type of driving scenario.

[0149] Optionally, the processing module 23 determines the target action pair in the traffic action combination according to the second preset filtering conditions, specifically for:

[0150] If a traffic action combination includes N following actions and M braking actions, and the first identification information of any vehicle in the N following actions is the same as the second identification information of any vehicle in the M braking actions, and the time intervals of the two actions overlap, then the current two actions are determined as the first target action pair, where N and M are both positive integers; or

[0151] If the traffic action combination includes: M braking actions and S lane changing actions; when the second identification information of any vehicle in the M braking actions is the same as the third identification information of any vehicle in the S lane changing actions, and the start time of the two actions is within a preset range, the current two actions are determined as the second target action pair, where S is a positive integer; or

[0152] If the traffic action combination includes Y straight-ahead actions and Z left-turn actions, when the fourth sign information of the intersection where any of the main vehicles in the Y straight-ahead actions are located is the same as the fifth sign information of the intersection where any of the main vehicles in the Z left-turn actions are located, and the time intervals of the two actions overlap, the current two actions are determined to be the third target action pair, where Y and Z are both positive integers.

[0153] Optionally, the processing module 52 is also specifically used to: evaluate driving behavior based on the type of driving scenario of the identified vehicle.

[0154] It should be noted that this device is a device corresponding to the above-mentioned driving scenario recognition method. All implementation methods in the above method embodiments are applicable to the embodiments of this device and can achieve the same technical effect.

[0155] Embodiments of this disclosure also provide a computing device, including: a processor and a memory storing a computer program, wherein the computer program, when executed by the processor, performs the method described above. All implementations in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0156] Embodiments of this disclosure also provide a computer-readable storage medium including instructions that, when executed on a computer, cause the computer to perform the methods described above. All implementations in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0157] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this disclosure.

[0158] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0159] In the embodiments provided in this disclosure, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0160] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0161] In addition, the functional units in the various embodiments of this disclosure can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0162] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this disclosure, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this disclosure. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0163] Furthermore, it should be noted that in the apparatus and method of this disclosure, it is obvious that the components or steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered equivalent solutions of this disclosure. Moreover, the steps performing the above series of processes can naturally be executed in the order described, but are not necessarily required to be executed in chronological order; some steps can be executed in parallel or independently of each other. Those skilled in the art will understand that all or any step or component of the method and apparatus of this disclosure can be implemented in any computing device (including processors, storage media, etc.) or network of computing devices, in hardware, firmware, software, or a combination thereof, which can be achieved by those skilled in the art using their basic programming skills after reading the description of this disclosure.

[0164] Therefore, the object of this disclosure can also be achieved by running a program or a set of programs on any computing device. The computing device can be a known general-purpose device. Therefore, the object of this disclosure can also be achieved simply by providing a program product containing program code implementing the method or apparatus. That is, such a program product also constitutes this disclosure, and a storage medium storing such a program product also constitutes this disclosure. Obviously, the storage medium can be any known storage medium or any storage medium developed in the future. It should also be noted that in the apparatus and methods of this disclosure, it is obvious that the components or steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered equivalent solutions of this disclosure. Furthermore, the steps performing the above series of processes can naturally be performed in the order described and in chronological order, but are not necessarily required to be performed in chronological order. Some steps can be performed in parallel or independently of each other.

[0165] The above description represents the preferred embodiments of this disclosure. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles described herein, and these improvements and modifications should also be considered within the scope of protection of this disclosure.

Claims

1. A method for recognizing driving scenarios, characterized in that, including: obtaining real-time parameter information of the vehicle during driving and road information monitored based on a high-precision map; the road information is used to identify the connection relationship of intersections and determine the steering action of the vehicle according to the road angle; determining the traffic action of the vehicle according to the real-time parameter information; identifying the type of the driving scenario of the vehicle according to the traffic action and the road information; wherein, determining the traffic action of the vehicle according to the real-time parameter information includes: determining an atomic traffic action of the vehicle and / or a traffic action combination formed by at least two atomic traffic actions according to the real-time parameter information; wherein, the atomic traffic action includes at least one of the following: lane change, following a vehicle, accelerating, decelerating, steering, going straight; wherein, identifying the type of the driving scenario of the vehicle according to the traffic action and the road information includes: if the traffic action includes an atomic traffic action, determining a target atomic traffic action according to a first preset screening condition; determining the driving scenario of the vehicle as a first type of driving scenario according to the target atomic traffic action and the road information; the first preset screening condition includes: action type and statistical attribute; if the traffic action includes a traffic action combination, determining a target action pair in the traffic action combination according to a second preset screening condition, pairing at least two atomic traffic actions in the traffic action combination one by one in sequence to obtain at least one target action pair, and at least one target action pair satisfies the second preset screening condition; if the traffic action combination includes: N following-vehicle actions and M braking actions, when the first identification information of any leading vehicle in the N following-vehicle actions is the same as the second identification information of any host vehicle in the M braking actions, and the time intervals of the two actions overlap, determining the current two actions as a first target action pair, that is, a following-vehicle - braking target action pair, where N and M are both positive integers; or if the traffic action combination includes: M braking actions and S lane change actions; when the second identification information of any host vehicle in the M braking actions is the same as the third identification information of any host vehicle in the S lane change actions, and the start time point t1 of the lane change action and the start time point t2 of the braking action satisfy t1>t2, and at the same time t1 - t2 < t0, where t0 is a preset time threshold, determining the current two actions as a second target action pair, that is, a braking - lane change target action pair, where S is a positive integer; or if the traffic action combination includes: Y going-straight actions and Z left-turn actions, when the fourth identification information of any host vehicle at the intersection where the Y going-straight actions are located is the same as the fifth identification information of any host vehicle at the intersection where the Z left-turn actions are located, and the time intervals of the two actions overlap, determining the current two actions as a third target action pair, that is, a going-straight - steering target action pair, where Y and Z are both positive integers; or If the following action and the reverse action are combined to form the target action pair "the vehicle in front of this lane is going against the flow", then within the time period of the two atomic traffic actions, it is necessary to satisfy that when the main vehicle of the following action is traveling, there is a vehicle with the same ID information going against the flow for at least N0 consecutive seconds in front of it, where N0 is a positive integer. Based on the target action pair and the road information, the driving scenario of the vehicle is determined to be a second type of driving scenario; wherein, the second preset screening conditions include: the temporal relationship, spatial relationship and vehicle status between atomic traffic actions.

2. The driving scene recognition method according to claim 1, characterized in that, The atomic traffic action and the combination of traffic actions each include at least one of the following attribute information: The subject of the action; Action type; The start and end times of the action; Statistical attributes; Identifier attributes.

3. The driving scene recognition method according to claim 1, characterized in that, Also includes: The driving behavior is evaluated based on the type of driving scenario identified for the vehicle.

4. A driving scene recognition device, characterized in that, include: The acquisition module is used to acquire real-time parameter information of the vehicle during driving and road information monitored based on high-precision maps; The road information is used to identify the connection relationship of intersections and to determine the vehicle's turning action based on the road angle; The processing module is configured to determine the vehicle's traffic actions based on the real-time parameter information; and to identify the type of the vehicle's driving scenario based on the traffic actions and the road information. Determining the vehicle's traffic actions based on the real-time parameter information includes: determining the vehicle's atomic traffic actions and / or combinations of traffic actions formed by at least two atomic traffic actions based on the real-time parameter information; wherein the atomic traffic actions include at least one of the following: lane change, following another vehicle, acceleration, deceleration, steering, and straight driving; and identifying the type of the vehicle's driving scenario based on the traffic actions and the road information includes: if the traffic actions include atomic traffic actions, determining a target atomic traffic action based on a first preset filtering condition; determining the vehicle's driving scenario as a first type of driving scenario based on the target atomic traffic action and the road information; the first preset filtering condition includes: action type and statistical attributes; if the traffic actions include combinations of traffic actions, determining a target action pair in the combination of traffic actions based on a second preset filtering condition; and sequentially pairing at least two atomic traffic actions in the combination to obtain at least one target action pair, wherein at least one target action pair satisfies the second preset filtering condition. If the traffic action combination includes N following actions and M braking actions, when the first identification information of any vehicle in the N following actions and the second identification information of any vehicle in the M braking actions are the same, and the time intervals of the two actions overlap, the current two actions are determined as the first target action pair, i.e., the following-braking target action pair, where N and M are both positive integers; or If the traffic action combination includes: M braking actions and S lane-changing actions; when the second identification information of any host vehicle in the M braking actions is the same as the third identification information of any host vehicle in the S lane-changing actions, and the start time point t1 of the lane-changing action and the start time point t2 of the braking action satisfy t1 > t2, and at the same time t1 - t2 < t0, where t0 is a preset time threshold, determine that the current two actions are the second target action pair, that is, the braking-lane-changing target action pair, where S is a positive integer; or If the traffic action combination includes: Y straight-ahead actions and Z left-turn actions, when the fourth identification information of the intersection where any host vehicle in the Y straight-ahead actions is located is the same as the fifth identification information of the intersection where any host vehicle in the Z left-turn actions is located, and the time intervals of the two actions overlap, determine that the current two actions are the third target action pair, that is, the straight-ahead-turning target action pair, where Y and Z are both positive integers; or If a target action pair of "a vehicle in front in the same lane is逆行" is combined with a following action and a逆行action, within the time periods of the two atomic traffic actions, it is required that when the host vehicle in the following action is driving, there is at least one vehicle with the same ID information逆行in front for at least N0 consecutive seconds, where N0 is a positive integer; according to the target action pair and the road information, determine that the driving scenario of the vehicle is a second-type driving scenario; where the second preset screening condition includes: the time relationship, spatial relationship, and vehicle state between atomic traffic actions.

5. A computing device, characterized in that, Includes: A processor and a memory storing a computer program, when the computer program is run by the processor, execute the method according to any one of claims 1 to 3.

6. A computer-readable storage medium, characterized in that, Includes instructions that, when run on a computer, cause the computer to execute the method according to any one of claims 1 to 3. It should be noted that the word "逆行" in the original text seems to be an incorrect or non-standard term. It might be a misspelling or an inappropriate expression. If it has a specific meaning in the relevant context, it should be accurately translated according to that meaning. Here, a literal translation is provided for the sake of following the translation requirements.

Images