Ghost head recognition method and device in driving process and electronic equipment

Abstract

This application discloses a method, device, and electronic device for identifying "ghost pedestrians" (vehicles lurking near obstacles) during driving, relating to the field of driver assistance technology. The method includes: acquiring driving behavior data of adjacent vehicles in the adjacent lane of the target vehicle, driving status data of the target vehicle, and current road environmental data; determining the behavior type of adjacent vehicles based on the driving behavior data and preset abnormal behavior rules; determining the ghost pedestrian risk level of the target vehicle during driving based on the behavior type, driving status data, environmental data, and preset ghost pedestrian risk weight ratio; and determining the corresponding graded control strategy for the target vehicle based on the ghost pedestrian risk level. This application, through data collected by multi-source sensing devices, can determine different ghost pedestrian risk levels and different levels of control strategies, thereby reducing the probability of false emergency braking warnings and accidents, lowering deployment costs, and improving the user's driving experience.

Description

Technical Field

[0001] This application relates to the field of driver assistance technology, and in particular to a method, device and electronic device for identifying "ghost pedestrians" during driving. Background Technology

[0002] Currently, when a pedestrian or two-wheeled vehicle suddenly crosses an intersection while the vehicle is in motion, the driver usually applies the brakes immediately to stop the vehicle. However, if the pedestrian or two-wheeled vehicle appears too suddenly or the braking distance is too short, the driver's manual braking may not be able to prevent an accident in time.

[0003] However, when pedestrians or two-wheeled vehicles suddenly cross the road at intersections and are obscured by adjacent vehicles (a phenomenon known as "ghost peek"), drivers and even automatic braking systems cannot react in time, leading to accidents. In some special scenarios, even when pedestrians or two-wheeled vehicles are far from the autonomous vehicle, the vehicle's autonomous driving system may still brake suddenly, causing false warnings of sudden braking and affecting the user's driving experience. Furthermore, traditional ghost peek detection methods rely on additional road infrastructure, resulting in high deployment costs. Summary of the Invention

[0004] This application provides a method, device, and electronic device for identifying "ghost pedestrians" during driving. The embodiments provided in this application solve the technical problems in the prior art where, in scenarios where pedestrians or two-wheeled vehicles suddenly cross intersections and are obscured by adjacent vehicles (referred to as "ghost pedestrian phenomenon"), drivers and even automatic braking systems cannot react in time, leading to accidents, affecting the user's driving experience, causing false warnings for emergency braking, and incurring high deployment costs. The embodiments provided in this application, through data collected by multi-source sensing devices, can determine different risk levels of ghost pedestrians and different levels of control strategies, thereby reducing the probability of false warnings for emergency braking and the probability of accidents, lowering deployment costs, and improving the user's driving experience.

[0005] In a first aspect, this application provides a method for identifying "ghost pedestrians" (protruding vehicles) during driving, the method comprising: Acquire driving behavior data of adjacent vehicles in the adjacent lane of the target vehicle, driving status data of the target vehicle, and environmental data of the current road, wherein the current road is used to characterize the road where the target vehicle is located at the current position; Based on the driving behavior data and preset abnormal behavior rules, the behavior type of the adjacent vehicles is determined, wherein the behavior type includes normal driving behavior or abnormal driving behavior; Based on the behavior type, the driving status data, the environmental data, and the preset ghost peep risk weight ratio, the ghost peep risk level of the target vehicle during the driving process is determined; Based on the risk level of the "ghost peek" (protruding vehicle), a graded control strategy corresponding to the target vehicle is determined.

[0006] In one feasible implementation, determining the behavior type of the adjacent vehicles based on the driving behavior data and preset abnormal behavior rules includes: The driving behavior data is used to extract change features to determine the driving behavior change features of the adjacent vehicles; When the driving behavior change characteristics meet the preset behavior anomaly rules, the behavior type of the adjacent vehicle is determined to be abnormal behavior, and the behavior confidence score is output. When the driving behavior change characteristics do not meet the preset abnormal behavior rules, the behavior type of the adjacent vehicle is determined to be normal behavior.

[0007] In one feasible implementation, determining the behavior type of the adjacent vehicles based on the driving behavior data and preset abnormal behavior rules includes: The driving behavior data is used to extract change features to determine the driving behavior change features of the adjacent vehicles; When the driving behavior change characteristics meet the preset behavior anomaly rules, the behavior type of the adjacent vehicle is determined to be abnormal behavior, and the behavior confidence score is output. When the driving behavior change characteristics do not meet the preset abnormal behavior rules, the behavior type of the adjacent vehicle is determined to be normal behavior.

[0008] In one feasible implementation, the driving state data of the target vehicle includes the target vehicle speed, and the environmental data of the current road includes the current road environment type. Determining the risk level of a pedestrian appearing unexpectedly during the driving process based on the behavior type, the driving state data, the environmental data, and a preset risk weight ratio for such incidents includes: Based on the confidence level of the behavior corresponding to the behavior type, determine the confidence coefficient; Based on the target vehicle speed, determine the corresponding speed coefficient for the target vehicle; Based on the current road environment type, determine the environmental coefficient corresponding to the target vehicle; Based on the confidence coefficient, the vehicle speed coefficient, the environmental coefficient, and the preset ghost peep risk weight ratio, the ghost peep risk level of the target vehicle during driving is determined.

[0009] In one feasible implementation, the driving state data of the target vehicle includes the target vehicle speed, and the environmental data of the current road includes the current road environment type. Determining the risk level of a pedestrian appearing unexpectedly during the driving process based on the behavior type, the driving state data, the environmental data, and a preset risk weight ratio for such incidents includes: Based on the confidence level of the behavior corresponding to the behavior type, determine the confidence coefficient; Based on the target vehicle speed, determine the corresponding speed coefficient for the target vehicle; Based on the current road environment type, determine the environmental coefficient corresponding to the target vehicle; Based on the confidence coefficient, the vehicle speed coefficient, the environmental coefficient, and the preset ghost peep risk weight ratio, the ghost peep risk level of the target vehicle during driving is determined.

[0010] In one feasible implementation, determining the graded control strategy corresponding to the target vehicle based on the risk level of the "ghost peek" includes: When the risk level of the "ghost peek" is low, the target vehicle is controlled to maintain its current driving state, and the driving behavior data of the adjacent vehicles in the adjacent lane are monitored in real time. When the risk level of the "ghost peek" is the medium risk level, the graded control strategy corresponding to the target vehicle is determined to be the first-level warning strategy, and the target vehicle is controlled to enter the pre-deceleration preparation state. When the risk level of the "ghost peek" is high, the corresponding graded control strategy for the target vehicle is determined to be a level two warning strategy, and a graded braking operation is triggered based on the speed of the target vehicle.

[0011] In one feasible implementation, the method further includes: Acquire the working status of vehicle-mounted multi-source sensing devices; When a malfunction is detected in the vehicle-mounted multi-source sensing device, a malfunction prompt message is generated, and a manual driving prompt message is output.

[0012] In a second aspect, this application provides a device for identifying "ghost pedestrians" (protruding vehicles) during driving, the device comprising: The acquisition module is used to acquire driving behavior data of adjacent vehicles in the adjacent lane of the target vehicle, driving status data of the target vehicle, and environmental data of the current road, wherein the current road is used to characterize the road where the target vehicle is located at the current position; The first determining module is used to determine the behavior type of the adjacent vehicles based on the driving behavior data and preset abnormal behavior rules, wherein the behavior type includes normal driving behavior or abnormal driving behavior. The second determining module is used to determine the risk level of a "ghost peep" during the driving process of the target vehicle based on the behavior type, the driving state data, the environmental data, and a preset "ghost peep" risk weight ratio. The third determining module is used to determine the graded control strategy corresponding to the target vehicle based on the risk level of the "ghost peephole".

[0013] In a third aspect of this application, an electronic device is provided, including a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the memory via the bus, and the machine-readable instructions are executed by the processor to perform the steps of the above-described method for identifying ghost pedestrians during vehicle operation.

[0014] In a fourth aspect of this application, an embodiment of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs the steps of the above-described method for identifying ghost pedestrians during vehicle operation.

[0015] Compared with the prior art, the ghost peep detection method, device, and electronic equipment provided in this application obtain driving behavior data of adjacent vehicles in the adjacent lane of the target vehicle, and determine the behavior type of the adjacent vehicles based on the driving behavior data and preset abnormal behavior rules. This enables real-time monitoring and analysis of the dynamic driving behavior of adjacent vehicles in adjacent lanes. Based on the behavior type, driving status data, environmental data, and preset ghost peep risk weight ratio, the ghost peep risk level of the target vehicle during driving is determined. This allows for the prediction of the ghost peep risk level in areas obscured by adjacent vehicles, thereby triggering graded control strategies and intervention operations in advance to reduce the probability of false emergency braking warnings and accidents. This reduces deployment costs while improving the user's driving experience. Attached Figure Description

[0016] Figure 1 A flowchart illustrating a method for identifying "ghost pedestrians" during driving, provided in an embodiment of this application, is shown. Figure 2 This paper shows a structural block diagram of a method and apparatus for identifying ghost pedestrians during driving, provided in an embodiment of this application. Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown.

[0017] Figure 2 and Figure 3 The correspondence between the figure labels and figure titles in the accompanying drawings is as follows: 200 Ghost protrusion detection device during vehicle operation; 210 Acquisition module; 220 First determination module; 230 Second determination module; 240 Third determination module; 250 Acquisition module; 260 Generation module; 300 Electronic device; 310 Processor; 320 Memory; 330 Bus. Detailed Implementation

[0018] To better understand the technical solutions provided in the embodiments of this specification, the technical solutions of the embodiments of this specification will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of this specification and the specific features in the embodiments are detailed descriptions of the technical solutions of the embodiments of this specification, rather than limitations on the technical solutions of this specification. In the absence of conflict, the embodiments of this specification and the technical features in the embodiments can be combined with each other.

[0019] In this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, without necessarily requiring or implying any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element. The term "two or more" includes two or more cases.

[0020] First, the applicable application scenarios of this application will be introduced. The embodiments provided in this application are applicable to the field of assisted driving technology, and in particular, relate to a method, device and electronic device for ghost peek-out during driving.

[0021] Currently, when a pedestrian or two-wheeled vehicle suddenly crosses an intersection while the vehicle is in motion, the driver usually applies the brakes immediately to stop the vehicle. However, if the pedestrian or two-wheeled vehicle appears too suddenly or the braking distance is too short, the driver's manual braking may not be able to prevent an accident in time.

[0022] However, when pedestrians or two-wheeled vehicles suddenly cross the road at intersections and are obscured by adjacent vehicles (a phenomenon known as "ghost peek"), drivers and even automatic braking systems cannot react in time, leading to accidents. In some special scenarios, even when pedestrians or two-wheeled vehicles are far from the autonomous vehicle, the vehicle's autonomous driving system may still brake suddenly, causing false warnings of sudden braking and affecting the user's driving experience. Furthermore, traditional ghost peek detection methods rely on additional road infrastructure, resulting in high deployment costs.

[0023] Based on this, embodiments of this application provide a method, device, and electronic device for identifying "ghost pedestrians" during driving. The embodiments provided by this application solve the technical problems in the prior art where, in the scenario where a pedestrian or two-wheeled vehicle suddenly crosses an intersection and is obscured by an adjacent vehicle (referred to as the "ghost pedestrian phenomenon"), the driver and even the automatic braking system cannot react in time, leading to accidents, affecting the user's driving experience, causing false warnings for emergency braking, and resulting in high deployment costs. The embodiments provided by this application, through data collected by multi-source sensing devices, can determine different risk levels of ghost pedestrians and different levels of control strategies, thereby reducing the probability of false warnings for emergency braking and the probability of accidents, reducing deployment costs, and improving the user's driving experience.

[0024] Figure 1 A flowchart illustrating a method for identifying "ghost pedestrians" during driving, as provided in an embodiment of this application, is shown. Figure 1 As shown, the method for identifying "ghost pedestrians" (protruding vehicles) during driving includes the following steps: S101. Obtain driving behavior data of adjacent vehicles in the adjacent lane of the target vehicle, driving status data of the target vehicle, and environmental data of the current road, wherein the current road is used to characterize the road where the target vehicle is located at the current position.

[0025] In this step, in the embodiments provided in this application, the vehicle-mounted multi-source sensing device on the target vehicle first collects driving behavior data of adjacent vehicles in the adjacent lanes of the target vehicle, driving status data of the target vehicle, and environmental data of the current road.

[0026] Understandably, the driving behavior data of adjacent vehicles in the adjacent lane of the target vehicle includes: the position, speed, acceleration, steering angle, and braking status of the adjacent vehicles; the driving status data of the target vehicle includes: the vehicle speed, steering angle, and lane; and the environmental data of the current road includes: the number of lanes, the type of intersection, and the distribution of obstructions.

[0027] It should be noted that the selection and installation deployment of the vehicle-mounted multi-source sensing devices in the embodiments provided in this application can be customized according to different application scenarios and usage conditions. The vehicle-mounted multi-source sensing devices in the embodiments provided in this application can be specifically, but are not limited to, at least two of the following: a front-side camera, a rear-side camera, a millimeter-wave radar, and a lidar of the target vehicle. In addition, the front / rear-side cameras in the embodiments provided in this application are all high-definition cameras with a resolution of 8 megapixels or higher. The millimeter-wave radar and lidar in the embodiments provided in this application are also included.

[0028] In the embodiments provided in this application, the millimeter-wave radar and lidar accurately collect information such as distance and speed; the high-definition camera is responsible for identifying the brake light status and steering intention of adjacent vehicles in adjacent lanes.

[0029] The embodiments provided in this application can improve the accuracy of data perception and the accuracy of locating the current road environment by collecting multi-source data through vehicle-mounted multi-source sensing devices. Specifically, the positioning accuracy can reach 0.1m, the speed accuracy can reach 2km / h, and the acceleration accuracy can reach 0.1m / s².

[0030] S102. Based on driving behavior data and preset abnormal behavior rules, determine the behavior type of adjacent vehicles, whereby the behavior type includes normal driving behavior or abnormal driving behavior.

[0031] In this step, in the embodiment provided in this application, after determining the driving behavior data of adjacent vehicles, it is necessary to preprocess the driving behavior data to obtain standard data. Then, based on the preprocessed driving behavior data and the preset abnormal behavior rules, the behavior type of adjacent vehicles is determined, specifically, whether the driving behavior of adjacent vehicles in adjacent lanes is normal driving behavior or abnormal driving behavior.

[0032] It is understood that the preprocessing methods in the embodiments provided in this application can be customized and used according to different application scenarios and usage conditions. The preprocessing methods in the embodiments provided in this application are specific but not limited to: preprocessing the original driving behavior data by cleaning, noise reduction and normalization, etc., to remove interference data or abnormal data such as data mistakenly collected by sensors.

[0033] It should be noted that the preset abnormal behavior rules in the embodiments provided in this application include at least one of the following: First abnormal rule: If, within a preset first time window, the speed decrease of adjacent vehicles in adjacent lanes is greater than or equal to a first speed threshold, or the brake lights remain on for a duration greater than or equal to a first duration threshold, then the adjacent vehicles are determined to be in an emergency braking state.

[0034] Here, in the embodiments provided in this application, the size of the first time window, the first speed threshold, and the first duration threshold can be customized and used according to different application scenarios and usage conditions. In the embodiments provided in this application, the first time window can be specifically set to 0.5s; the first speed threshold can be specifically set to 10km / h; and the first duration threshold can be specifically set to 0.3s.

[0035] If, in the absence of any obstacle ahead (within the target vehicle's perception range), an adjacent vehicle in an adjacent lane experiences a speed reduction of ≥10km / h, or its brake lights remain illuminated for ≥0.3s, it is determined that the adjacent vehicle has detected a "ghostly" target appearing ahead (in the target vehicle's blind spot) and needs to take emergency evasive action, thus resulting in a sudden deceleration / emergency braking.

[0036] Second abnormal rule: When the longitudinal distance between an adjacent vehicle and the target vehicle is less than or equal to the first distance threshold and the lateral distance is less than or equal to the second distance threshold, the adjacent vehicle suddenly changes lanes to the current lane where the target vehicle is located, and the turn signal is not turned on before the lane change, or the turn signal is turned on for less than the second duration threshold.

[0037] Here, in the embodiments provided in this application, the values ​​of the first distance threshold, the second distance threshold, and the second duration threshold can be customized and used according to different application scenarios and usage conditions. In the embodiments provided in this application, the first distance threshold can be specifically set to 50m; the second distance threshold can be specifically set to 2m; and the second duration threshold can be specifically set to 0.5s.

[0038] This means that if an adjacent vehicle in an adjacent lane suddenly changes lanes to the target vehicle's current lane or the opposite lane within a longitudinal distance of ≤50m and a lateral distance of ≤2m, and the turn signal is not activated before the lane change, or the turn signal activation time is <0.5s, this behavior may be due to an adjacent vehicle in an adjacent lane making an emergency lane change to avoid a target that suddenly appears from behind.

[0039] The third abnormal rule: At intersections or near pedestrian crossings, the speed of vehicles in adjacent lanes is consistently less than or equal to the second speed threshold, and the speed fluctuation is greater than or equal to the third speed threshold.

[0040] Here, in the embodiments provided in this application, the magnitudes of the second speed threshold and the third speed threshold can be customized and used according to different application scenarios and usage conditions. In the embodiments provided in this application, the second speed threshold can be specifically set to 10km / h; the third speed threshold can be specifically set to m; and the second duration threshold can be specifically set to 5km / h.

[0041] This refers to adjacent vehicles in adjacent lanes lingering at low speeds near intersections or pedestrian crossings, specifically with a continuous speed of ≤10km / h and a speed fluctuation of ≥5km / h (in the absence of congestion and red lights). This behavior may be due to adjacent vehicles noticing that pedestrians or non-motorized vehicles are preparing to cross the road behind an obstruction, and therefore, they are observing and giving way.

[0042] In summary, the preset abnormal behavior rules provided in the embodiments of this application include, but are not limited to, the above-described methods, and normal driving behavior refers to all driving behaviors other than abnormal driving behavior.

[0043] S103. Based on behavior type, driving status data, environmental data, and preset ghost peep risk weight ratio, determine the ghost peep risk level of the target vehicle during driving.

[0044] In this step, in the embodiments provided in this application, after determining the behavior type of adjacent vehicles, the embodiments provided in this application will determine the ghost peep risk level of the target vehicle during driving based on the preset ghost peep risk weight ratio, driving status data, environmental data, and the preset ghost peep risk weight ratio.

[0045] It is understood that the preset ghost peek risk weight ratio in the embodiments provided in this application can be customized according to different application scenarios and usage conditions.

[0046] It should be noted that the risk levels of ghost peeks in the embodiments provided in this application include high risk level, medium risk level and low risk level.

[0047] In the embodiments provided in this application, the risk level of the ghost peeping can be characterized by R.

[0048] S104. Based on the risk level of the "ghost vehicle" (a vehicle that suddenly appears in front of another vehicle), determine the corresponding graded control strategy for the target vehicle.

[0049] In this step, in the embodiments provided in this application, after determining the hierarchical control strategy corresponding to the target vehicle based on the ghost peep risk level, the embodiments provided in this application also need to continuously acquire updated driving behavior data of adjacent vehicles in adjacent lanes, updated driving status data of the target vehicle, and updated environmental data of the current road, and redetermine the ghost peep risk level based on the updated data, and when the redetermined ghost peep risk level is lower than the threshold corresponding to the current ghost peep risk level, stop executing the current hierarchical control strategy, and control the target vehicle in the current lane to gradually return to normal driving state.

[0050] The ghost peep detection method provided in this application, compared with the prior art, obtains the driving behavior data of adjacent vehicles in the adjacent lane of the target vehicle, and determines the behavior type of the adjacent vehicles based on the driving behavior data and preset abnormal behavior rules. This enables real-time monitoring and analysis of the dynamic driving behavior of adjacent vehicles in adjacent lanes. Based on the behavior type, driving status data, environmental data, and preset ghost peep risk weight ratio, the method determines the ghost peep risk level of the target vehicle during driving, and predicts the ghost peep risk level in the area blocked by adjacent vehicles. This allows for early triggering of graded control strategies and intervention operations to reduce the probability of false emergency braking warnings and accidents, thereby reducing deployment costs and improving the user's driving experience.

[0051] This application does not require additional road infrastructure modifications. It can achieve data collection and policy execution by relying solely on existing onboard multi-source sensing equipment. It can be integrated into L1 and above level intelligent driving vehicles and is adaptable to multiple vehicle types and road scenarios, thereby improving the reliability of the system.

[0052] For example, based on driving behavior data and preset abnormal behavior rules, the behavior types of adjacent vehicles are determined, including: The system extracts change features from driving behavior data to determine the driving behavior change features of adjacent vehicles. When the driving behavior change features meet the preset abnormal behavior rules, the behavior type of the adjacent vehicle is determined to be abnormal behavior, and the behavior confidence score is output. When the driving behavior change features do not meet the preset abnormal behavior rules, the behavior type of the adjacent vehicle is determined to be normal behavior.

[0053] In the embodiments provided in this application, data showing characteristic changes in driving behavior data are extracted, such as the rate of change of speed, braking duration, steering angle change, and relative distance / relative speed to the target vehicle. After determining the extraction of the above-mentioned driving behavior change features, a denoising filtering algorithm is used to remove image noise from the driving behavior change features, and an interpolation algorithm is used to fill in the missing data in the driving behavior change features to ensure the validity and consistency of the data.

[0054] It is understood that after determining the characteristics of the change in driving behavior, the embodiments provided in this application will determine whether the characteristics of the change in driving behavior after noise reduction and interpolation calculation meet the preset abnormal behavior rules. When it is determined that the preset abnormal behavior rules are met, the behavior type of the adjacent vehicle is determined to be abnormal behavior, and the behavior confidence is output. When it is determined that the characteristics of the change in driving behavior do not meet the preset abnormal behavior rules, the behavior type of the adjacent vehicle is determined to be normal behavior.

[0055] In this application, by comparing the characteristics of changes in driving behavior with preset abnormal behavior rules, ghost cameras can be accurately identified during the driving process. Compared with the traditional method of ghost camera identification that relies on external road infrastructure, the deployment cost is lower.

[0056] For example, when the characteristics of the change in driving behavior meet the preset abnormal behavior rules, the behavior type of the adjacent vehicles is determined to be abnormal behavior, and the behavior confidence score is output, including: When the characteristics of changes in driving behavior meet the preset abnormal behavior rules, the behavior type of adjacent vehicles is determined as a candidate abnormal behavior; the characteristics of changes in driving behavior are input into the trained abnormal behavior learning model, and the behavior confidence score output by the trained abnormal behavior learning model is determined, wherein the behavior confidence score is used to characterize the probability that the behavior type of adjacent vehicles is a real abnormal behavior; the candidate abnormal behavior is corrected according to the behavior confidence score, and the behavior type of adjacent vehicles is determined.

[0057] In the embodiments provided in this application, after determining that the driving behavior change features meet the preset abnormal behavior rules and defining the behavior type corresponding to the driving behavior change features as a candidate abnormal behavior, the embodiments provided in this application will input the driving behavior change features into the trained abnormal behavior learning model, determine the behavior type of adjacent vehicles and its confidence level, and compare and correct the output behavior type and the behavior confidence level corresponding to the behavior type with the candidate abnormal behavior type to determine the behavior type of adjacent vehicles with high precision.

[0058] It is understood that in the embodiments provided in this application, the actual behavior type determined by the driving behavior change features after being input into the trained abnormal behavior learning model may be the same as or different from the candidate abnormal behavior. That is, there may be a situation where the trained abnormal behavior learning model outputs a normal behavior type. Therefore, the result output by the trained abnormal behavior learning model should be taken as the main one.

[0059] It should be noted that the type of the trained abnormal behavior learning model in the embodiments provided in this application can be customized and used according to different application scenarios and usage conditions. The trained abnormal behavior learning model in the embodiments provided in this application can be specifically, but is not limited to, a Long Short-Term Memory (LSTM) neural network model.

[0060] In the embodiments provided in this application, abnormal behavior is defined as behavior that is strongly associated with the risk of ghost peeking out.

[0061] In this application, the behavior type of adjacent vehicles is reconfirmed by the type of the trained abnormal behavior learning model, so as to improve the accuracy and real-time performance of abnormal behavior recognition, avoid misjudgment and missed judgment of the risk level of ghost peeks, and avoid warning deviation caused by misjudgment due to a single rule.

[0062] For example, the target vehicle's driving status data includes the target vehicle's speed, and the current road environment data includes the current road environment type. Based on the behavior type, driving status data, environmental data, and a preset "ghost peek" risk weight ratio, the target vehicle's "ghost peek" risk level during driving is determined, including: Based on the behavioral confidence level corresponding to the behavior type, the confidence coefficient is determined; based on the target vehicle speed, the speed coefficient corresponding to the target vehicle is determined; based on the current road environment type, the environmental coefficient corresponding to the target vehicle is determined; based on the confidence coefficient, speed coefficient, environmental coefficient, and the preset ghost peep risk weight ratio, the ghost peep risk level of the target vehicle during the driving process is determined.

[0063] In the embodiments provided in this application, different behavior types correspond to different behavior confidence levels, and different behavior confidence levels correspond to different confidence coefficients; different target vehicle speeds also correspond to different speed coefficients; and different current road environment types correspond to different environment coefficients.

[0064] It is understood that the behavior confidence level in the embodiments provided in this application is represented by C1, which is the probability of determining that the behavior is a real abnormal behavior (0≤C1≤1); the vehicle speed coefficient is represented by Vf, and the faster the target vehicle speed, the larger Vf is; the environmental coefficient is represented by Ef.

[0065] It should be noted that in the embodiments provided in this application, Vf=0.2 when the target vehicle speed is ≤30km / h; Vf=0.5 when 0km / h<target vehicle speed≤60km / h; and Vf=0.8 when the target vehicle speed>60km / h.

[0066] At intersections, pedestrian crossings, and near bus stops (high-risk environment), Ef=0.8; on ordinary straight road sections (low-risk environment), Ef=0.3; and on suburban roads (medium-risk environment), Ef=0.5.

[0067] In the above, the preset risk weight ratio for "ghost peek" includes the behavior weight of vehicles in adjacent lanes, denoted by W1; the status weight of the vehicle itself, denoted by W2, with W2 fixed at 0.3; and the road environment weight, denoted by W3, with W3 fixed at 0.2.

[0068] Here, the weights of adjacent lane vehicle behavior are set according to the behavior type. In the embodiment provided in this application, abnormal behavior 1 (such as adjacent vehicle suddenly decelerating / emerging), W1=0.5; abnormal behavior 2 (such as adjacent vehicle changing lanes abnormally), W1=0.4; abnormal behavior 3 (such as adjacent vehicle lingering at low speed), W1=0.3; and normal behavior W1=0.1.

[0069] In this application, the method for determining the risk level of a target vehicle suddenly appearing out of the road during driving can be specifically as follows: R = W1 × C1 + W2 × Vf + W3 × Ef; R is used to characterize the risk level of ghost peeks.

[0070] In this application, the method for determining the risk level of a target vehicle lurking in the road during driving does not rely on the forward perception of a single target vehicle. Instead, it utilizes the strong correlation between the behavior of adjacent vehicles in adjacent lanes and the risk of lurking in the road. By identifying abnormal behavior of adjacent vehicles in adjacent lanes, the risk of lurking in the blind spot can be predicted in advance. Compared with existing technologies, potential risks can be detected a period of time in advance, significantly allowing for reaction time, thereby improving the avoidance rate and prediction accuracy of lurking accidents.

[0071] For example, based on confidence coefficient, vehicle speed coefficient, environmental coefficient, and preset "ghost peek" risk weight ratio, the "ghost peek" risk level of the target vehicle during driving is determined, including: Based on the confidence coefficient, vehicle speed coefficient, environmental coefficient, and preset risk weight ratio for "ghost peepholes," the risk value of the target vehicle during driving is determined. When the risk value is less than a first preset risk threshold, the risk level of the target vehicle during driving is determined to be low risk. When the risk value is greater than or equal to the first preset risk threshold and less than a second preset risk threshold, the risk level of the target vehicle during driving is determined to be medium risk. When the risk value is greater than or equal to the second preset risk threshold, the risk level of the target vehicle during driving is determined to be high risk.

[0072] In the embodiments provided in this application, the range of the ghost camera risk value for the target vehicle during driving is 0-1. Based on the value of R, a level 3 ghost camera risk level is determined. If 0 ≤ R < 0.3, then the risk level of the target vehicle during the driving process is determined to be low risk level, indicating that there is no obvious risk of ghosting, and the behavior of adjacent vehicles is normal, or the confidence level of the abnormal behavior of adjacent vehicles is extremely low.

[0073] If 0.3≤R<0.6, then the risk level of the target vehicle's "ghosting" during driving is determined to be medium risk level, indicating that the adjacent vehicle is exhibiting abnormal behavior, but the target vehicle's speed is low or the distance between it and the adjacent vehicle in the adjacent lane is far.

[0074] If R≥0.6, the risk level of the target vehicle appearing out of nowhere during driving is determined to be high risk level. This indicates that the adjacent vehicle in the adjacent lane has clearly exhibited abnormal behavior, and the target vehicle is traveling at a high speed or is close to the adjacent vehicle, and there is a high probability that the target vehicle has entered the blind spot.

[0075] It is understood that the first preset risk threshold and the second preset risk threshold in the embodiments provided in this application can be selected according to different application scenarios and usage conditions. The first preset risk threshold in the embodiments provided in this application can be specifically set to 0.3; the second preset risk threshold can be specifically set to 0.6.

[0076] In this application, by setting different preset risk thresholds, the risk value of "ghost peek" is set to different levels, so that warnings and interventions are triggered only when there is a real risk. This avoids the problem of reduced traffic efficiency caused by indiscriminate deceleration of target vehicles in the prior art, and improves road traffic efficiency and driving experience while ensuring safety.

[0077] For example, based on the risk level of a "ghost peek" (a sudden, unexpected movement of pedestrians), a tiered control strategy is determined for the target vehicle, including: When the risk level of a "ghost peek" is low, the target vehicle is controlled to maintain its current driving state, and the driving behavior data of adjacent vehicles in adjacent lanes are monitored in real time. When the risk level of a "ghost peek" is medium, the corresponding graded control strategy for the target vehicle is determined to be a level one warning strategy, and the target vehicle is controlled to enter a pre-deceleration preparation state. When the risk level of a "ghost peek" is high, the corresponding graded control strategy for the target vehicle is determined to be a level two warning strategy, and graded braking operations are triggered based on the target vehicle's speed.

[0078] In the above-described embodiments provided in this application, determining the hierarchical control strategy corresponding to the target vehicle as a level one warning strategy and controlling the target vehicle to enter the pre-deceleration preparation state includes: generating a first warning signal, wherein the first warning signal includes a first color light prompt, a first frequency prompt sound, and a first text display information, and controlling the braking system to build up pressure in advance for a preset pressure building time to eliminate the free travel of the brake pedal.

[0079] The target vehicle's corresponding graded control strategy is determined to be a level-two warning strategy. Based on the target vehicle's speed, a graded braking operation is triggered, including: generating a second warning signal, which includes a second color light indicator, a second frequency warning tone, and a second text display message; when the vehicle speed is less than or equal to a first speed threshold, a first-level deceleration operation is performed to reduce the vehicle speed to a first deceleration range; when the vehicle speed is greater than the first speed threshold, a second-level deceleration operation is performed to reduce the vehicle speed to a second deceleration range, wherein the upper limit of the second deceleration range is greater than the upper limit of the first deceleration range; after the first or second-level deceleration operation, if the risk level of a "ghost peek" (a sudden, unexpected movement of a vehicle appearing in front of another vehicle) remains high and no driver intervention is detected, an emergency braking operation is triggered to reduce the vehicle speed to below a safe speed threshold.

[0080] It is understood that in the embodiments provided in this application, the emergency braking operation adopts a three-stage control method, including: the first stage: pre-pressurizing the braking system to eliminate the braking gap; the second stage: linearly increasing the braking pressure to make the braking acceleration reach the target acceleration; and the third stage: adjusting the braking pressure by using intermittent braking to avoid wheel lock-up.

[0081] It should be noted that the colors of the first and second color light indicators in the embodiments provided in this application can be customized and used according to different application scenarios and usage conditions. In the embodiments provided in this application, the color of the first color light indicator is determined to be a yellow warning light; the color of the second color light indicator is a red warning light.

[0082] Specifically, when the risk level of a "ghost peek" is low, no warning is triggered, the target vehicle continues to drive normally, and only the behavior of adjacent vehicles and the surrounding environment of the target vehicle are continuously monitored to ensure that changes in risk can be captured in a timely manner.

[0083] When the risk level of a "ghost peek" is medium, the first warning signal is triggered, and the yellow warning light on the dashboard illuminates. At the same time, a gentle warning sound (frequency 1Hz) is emitted, and the vehicle's infotainment system displays "Abnormal adjacent lane, beware of blind spot risk," reminding the driver to be vigilant, keep both hands on the steering wheel, and focus on the direction of the obstructing blind spot. Simultaneously, the target vehicle enters the "pre-deceleration preparation" state, and the braking system builds pressure 0.3 seconds in advance to eliminate pedal free travel, reserving reaction time for possible subsequent braking operations. In this application, the target vehicle's speed remains unchanged (if the speed is >40km / h, the speed can be slightly reduced by 5-10km / h).

[0084] When the risk level of a "ghost peek" is high, a second warning signal is triggered, and a red warning light illuminates on the dashboard. The warning tone changes to a high-frequency, rapid sound (frequency 2Hz), and the vehicle's infotainment system flashes the message "Blind spot risk, slow down immediately." Simultaneously, tiered intervention operations are triggered: If the vehicle's speed is ≤30km / h, a slight deceleration is triggered (speed decreases by 10-15km / h) to maintain the target vehicle's controllability; if the target vehicle's speed is >30km / h, a moderate deceleration is triggered (speed decreases by 15-25km / h). If the risk value remains ≥0.6 after deceleration and there is no driver intervention, emergency braking is triggered (braking acceleration controlled within 0.3g), bringing the target vehicle to a stop or reducing its speed to a safe level (≤10km / h) to avoid a collision with the intruding target.

[0085] When the adjacent vehicle in the adjacent lane resumes normal behavior, or the vehicle leaves the blind spot and the risk value R < 0.3, the warning will stop and the target vehicle will be gradually restored to normal driving status (if deceleration has been triggered, the vehicle speed will be increased uniformly to the original driving speed).

[0086] In this application, a graded warning is issued to the driver based on the risk level of a "ghost peek" (a sudden, unexpected pedestrian movement). When the risk level of a ghost peek reaches a preset risk threshold, a graded vehicle control strategy (deceleration or emergency braking) is triggered. This improves driving comfort while reducing the risk of rear-end collisions caused by sudden braking. It maximizes reaction time while avoiding false warnings and false braking, thus balancing driving safety and traffic efficiency.

[0087] For example, the method also includes: Acquire the working status of the vehicle-mounted multi-source sensing device; when a fault is detected in the vehicle-mounted multi-source sensing device, generate a fault prompt message and output a manual driving prompt message.

[0088] In the embodiments provided in this application, after generating fault warning information, the target vehicle on the current road will switch to basic defense mode. In basic defense mode, monitoring is carried out solely by the vehicle's forward sensing equipment, and the weight parameters in the risk assessment model are reduced to avoid false warnings.

[0089] This application maximizes reaction time while avoiding false warnings and false braking, thus balancing driving safety and traffic efficiency.

[0090] The following is a general implementation process to illustrate the process of determining the hierarchical control strategy corresponding to the target vehicle.

[0091] General Implementation Examples: In the embodiments provided in this application, the overall response delay for monitoring ghost pedestrian events during driving is determined to be ≤100ms. First: After the vehicle system is started, it uses multi-source sensing sensors to collect data on the target vehicle's driving status (such as target speed, steering angle, and driving lane) and road environment (such as number of lanes and intersection type) to complete the initialization of system parameters.

[0092] Then, the multi-source sensing sensors begin to collect driving behavior data (such as position, speed, acceleration, steering angle, and braking status) of adjacent vehicles in adjacent lanes, with a collection frequency of ≥10Hz to ensure the real-time nature of the collected data.

[0093] Next, the collected data is cleaned, denoised, and normalized to remove outliers and extract the driving behavior characteristics of adjacent vehicles in adjacent lanes (such as speed change rate, braking duration, steering angle change, and relative distance / relative speed with the vehicle).

[0094] Then, the behavior type of adjacent vehicles is identified to determine whether the adjacent vehicles in the adjacent lanes are driving normally or abnormally, and the behavior confidence C1 is output. Then, based on the behavior type, driving status data, environmental data and preset ghost peep risk weight ratio, the ghost peep risk value R of the target vehicle during the driving process is determined, and the ghost peep risk level is determined according to the value of R.

[0095] Then, based on the corresponding risk levels of the "ghost pedestrian" (a vehicle suddenly appearing in front of another vehicle), the corresponding graded warning and intervention operations are determined, and a warning is issued to the driver, and if necessary, the target vehicle in the current lane is slowed down or braked urgently.

[0096] In this application, if the multi-source sensing device malfunctions (such as sensor failure), a fault warning is immediately issued, and the system automatically switches to manual control driving mode to reduce the weight of risk assessment and avoid false warnings.

[0097] The following three examples illustrate the graded control strategies for target vehicles under different risk levels of "ghost peeks".

[0098] Example 1 (Target vehicle is at a high risk level): Sub-step 1, Data Acquisition: The onboard multi-source sensing device collects real-time driving behavior data of the car in the adjacent lane on the left. It detects that the car's speed drops from 50km / h to 25km / h (i.e., a decrease of 25km / h within 0.5s), the brake lights remain on (duration ≥0.3s), the longitudinal distance to the target vehicle is 30m, and the lateral distance is 1.5m. At the same time, it collects the target vehicle's speed as 50km / h (i.e., >30km / h), and the road environment is near a pedestrian crossing at an intersection (i.e., environmental coefficient Ef=0.8).

[0099] Sub-step 2, data preprocessing: The collected driving behavior data is cleaned, denoised, and normalized, and driving behavior change features are extracted—speed change rate is -50km / h / s, braking time is 0.4s, steering angle is 0° (no steering), and the relative speed with the target vehicle is -25km / h (i.e., the vehicle in the adjacent lane decelerates faster than the target vehicle).

[0100] Sub-step 3, behavior type identification: Determine the behavior type of the car in the adjacent lane as "abnormal behavior 1 (sudden deceleration / emergency braking)", and determine the behavior confidence level C1=0.95.

[0101] Sub-step 4, risk level identification: Substitute into the risk assessment formula, W1=0.5, C1=0.95, W2=0.3, Vf=0.5 (vehicle speed 50km / h, 30km / h<vehicle speed≤60km / h), W3=0.2, Ef=0.8, calculate R=0.5×0.95+0.3×0.5+0.2×0.8=0.475+0.15+0.16=0.785≥0.6, and determine it as a high-risk level.

[0102] Sub-step 5, early warning and intervention of the graded control strategy: Trigger the secondary early warning strategy (such as flashing red warning lights, high-frequency rapid prompts, and flashing HUD display "Blind spot risk, slow down immediately"). Since the target vehicle's speed is 50km / h > 30km / h, moderate deceleration is triggered, reducing the speed to 25-35km / h, while monitoring the risk value change. If, after 1 second, the adjacent vehicle in the adjacent lane is still in an emergency braking state, the risk value of the "ghost pedestrian" is determined to be R = 0.72 ≥ 0.6, and there is no driver intervention. Then, emergency braking is triggered, and the speed of the target vehicle is manually reduced to ≤ 10km / h. If it is subsequently discovered that a pedestrian is crossing the road in the blind spot of a car in the left lane, since the target vehicle has already slowed down and braked in advance, a collision can be avoided to the greatest extent.

[0103] Sub-step 6, risk resolution: When the pedestrian crosses the road, the car in the left lane resumes normal driving (speed increased to 50km / h). At this time, the risk value of the "ghost pedestrian" is R=0.25<0.3. At this time, the warning is stopped, the speed of the target vehicle is increased to 50km / h at a constant speed, and normal driving resumes.

[0104] Example 2 (Target vehicle is at a medium risk level): Sub-step 1, data collection: The target vehicle is traveling on a straight section of urban road at a speed of 40 km / h. The adjacent lane is on the right, and there is an SUV in the adjacent lane. The onboard multi-source sensing device collects data showing that the SUV's speed is consistently 8 km / h with a speed fluctuation range of 6 km / h (no congestion and no red lights). The longitudinal distance between the SUV and the target vehicle is 40m, and the lateral distance is 1.8m. At the same time, the environmental data of the current road is collected as being near a pedestrian crossing (i.e., Ef=0.8).

[0105] Sub-step 2, data preprocessing: The collected driving behavior data is cleaned, denoised, and normalized, and driving behavior change characteristics are extracted—vehicle speed ≤ 10km / h, speed fluctuation range ≥ 5km / h, no braking, no abnormal steering, and relative speed with the target vehicle is -32km / h (i.e., vehicles in adjacent lanes decelerate faster than the target vehicle).

[0106] Sub-step 3, behavior type identification: Determine the behavior of the SUV in the adjacent lane as "abnormal behavior 3 (low-speed wandering)", and determine the behavior confidence level C1=0.9.

[0107] Sub-step 4, risk level identification: The calculated R = 0.3 × 0.9 + 0.3 × 0.5 + 0.2 × 0.8 = 0.27 + 0.15 + 0.16 = 0.58, where 0.3 ≤ 0.58 < 0.6, thus classifying it as a medium-risk level.

[0108] Sub-step 5, warning and intervention of graded control strategy: trigger the first-level warning (yellow warning light, gentle prompt sound, HUD display "abnormal adjacent lane, beware of blind spot risk"), the vehicle enters the pre-deceleration preparation state, the braking system builds pressure in advance, and the vehicle speed is slightly reduced to 35km / h.

[0109] Sub-step 6, risk resolution: When the target vehicle leaves the pedestrian crossing area, the SUV in the adjacent lane will increase its speed to 40km / h and resume normal driving. If the ghost vehicle R=0.28<0.3 is confirmed, the warning will be stopped and the vehicle will resume normal driving at 40km / h.

[0110] Example 3 (Target vehicle is at a low risk level): Sub-step 1, data collection: The target vehicle is traveling on a suburban road at a speed of 60 km / h. The adjacent lane is to the left, and the adjacent vehicle in the adjacent lane is traveling at a constant speed of 62 km / h. It is not braking and is not turning abnormally. The longitudinal distance between the target vehicle and the adjacent vehicle is 60 m and the lateral distance is 2.2 m. The current road environment data is a normal straight road section (i.e., Ef=0.3).

[0111] Sub-step 2, behavior type identification: determine that the driving behavior of the adjacent vehicle is normal behavior, and determine the behavior confidence level C1=0.98.

[0112] Sub-step 3, risk level identification: Calculate R = 0.1 × 0.98 + 0.3 × 0.5 + 0.2 × 0.3 = 0.098 + 0.15 + 0.06 = 0.308, which is close to the threshold of medium risk level, but still belongs to low risk level (0 ≤ R < 0.3, here the behavioral confidence is adjusted, normal behavior C1 = 0.8, R = 0.1 × 0.8 + 0.3 × 0.5 + 0.2 × 0.3 = 0.08 + 0.15 + 0.06 = 0.29 < 0.3), and is judged as low risk level.

[0113] Sub-step 4, early warning and intervention of the hierarchical control strategy: If no early warning is triggered, control the target vehicle in the current lane to maintain a normal driving speed of 60km / h, and continuously monitor the driving behavior of vehicles in adjacent lanes.

[0114] Figure 2 This illustration shows a structural schematic diagram of a ghost camera detection device during vehicle operation, provided in an embodiment of this application. Figure 2 As shown, the ghost peek detection device 200 during vehicle operation includes: The acquisition module 210 is used to acquire driving behavior data of adjacent vehicles in the adjacent lane of the target vehicle, driving status data of the target vehicle, and environmental data of the current road, wherein the current road is used to represent the road where the target vehicle is located at the current position.

[0115] The first determining module 220 is used to determine the behavior type of adjacent vehicles based on driving behavior data and preset abnormal behavior rules, wherein the behavior type includes normal driving behavior or abnormal driving behavior.

[0116] The second determining module 230 is used to determine the risk level of a target vehicle during driving based on behavior type, driving status data, environmental data, and a preset risk weight ratio for ghost peepholes.

[0117] The third determination module 240 is used to determine the corresponding hierarchical control strategy for the target vehicle based on the risk level of the ghost probe.

[0118] The acquisition module 250 is used to acquire the working status of the vehicle-mounted multi-source sensing device.

[0119] The generation module 260 is used to generate fault prompt information and output manual driving prompt information when a fault is detected in the vehicle-mounted multi-source sensing device.

[0120] For example, the first determining module 220 is specifically used for: Change features are extracted from driving behavior data to determine the driving behavior change features of adjacent vehicles.

[0121] When the characteristics of the driving behavior change meet the preset abnormal behavior rules, the behavior type of the adjacent vehicles is determined to be abnormal behavior, and the behavior confidence score is output.

[0122] When the characteristics of changes in driving behavior do not meet the preset abnormal behavior rules, the behavior type of adjacent vehicles is determined to be normal behavior.

[0123] For example, when the characteristics of the change in driving behavior meet the preset abnormal behavior rules, the behavior type of the adjacent vehicles is determined to be abnormal behavior, and the behavior confidence score is output, including: When the characteristics of changes in driving behavior meet the preset abnormal behavior rules, the behavior type of adjacent vehicles is determined as a candidate abnormal behavior.

[0124] The driving behavior change features are input into the trained abnormal behavior learning model, and the behavior confidence of the trained abnormal behavior learning model is determined. The behavior confidence is used to characterize the probability that the behavior type of the adjacent vehicles is a real abnormal behavior.

[0125] Candidate abnormal behaviors are corrected based on behavioral confidence to determine the behavior types of adjacent vehicles.

[0126] For example, the target vehicle's driving status data includes the target vehicle's speed, and the current road environment data includes the current road environment type. The second determining module 230 is specifically used for: The confidence coefficient is determined based on the behavioral confidence corresponding to the behavioral type.

[0127] Based on the target vehicle's speed, determine the corresponding speed coefficient for the target vehicle.

[0128] Based on the current road environment type, determine the environmental coefficient corresponding to the target vehicle.

[0129] Based on confidence coefficient, vehicle speed coefficient, environmental coefficient, and preset ghost peep risk weight ratio, the ghost peep risk level of the target vehicle during driving is determined.

[0130] For example, based on confidence coefficient, vehicle speed coefficient, environmental coefficient, and preset "ghost peek" risk weight ratio, the "ghost peek" risk level of the target vehicle during driving is determined, including: Based on the confidence coefficient, vehicle speed coefficient, environmental coefficient, and preset ghost peep risk weight ratio, the ghost peep risk value of the target vehicle during the driving process is determined.

[0131] When the risk value of a "ghost peek" is less than the first preset risk threshold, the risk level of the target vehicle during the driving process is determined to be low risk.

[0132] When the risk value of a "ghost peek" is greater than or equal to the first preset risk threshold and less than the second preset risk threshold, the risk level of the target vehicle during driving is determined to be medium risk.

[0133] When the risk value of a "ghost peek" is greater than or equal to the second preset risk threshold, the risk level of the target vehicle during the driving process is determined to be high risk.

[0134] For example, the third determining module 240 is specifically used to: when the risk level of the ghost peek is low, control the target vehicle to maintain its current driving state and monitor the driving behavior data of adjacent vehicles in adjacent lanes in real time.

[0135] When the risk level of a "ghost peek" is medium risk, the corresponding graded control strategy for the target vehicle is determined to be a level one warning strategy, and the target vehicle is controlled to enter a pre-deceleration preparation state.

[0136] When the risk level of a "ghost peek" is high, the corresponding graded control strategy for the target vehicle is determined to be a level two warning strategy, and graded braking operation is triggered based on the target vehicle's speed.

[0137] The ghost peep detection device 200 provided in this application embodiment, compared with the prior art, obtains the driving behavior data of adjacent vehicles in the adjacent lane of the target vehicle, and determines the behavior type of adjacent vehicles based on the driving behavior data and preset abnormal behavior rules. It realizes real-time monitoring and analysis of the dynamic driving behavior of adjacent vehicles in adjacent lanes, and determines the ghost peep risk level of the target vehicle during driving based on the behavior type, driving status data, environmental data and preset ghost peep risk weight ratio. It can predict the ghost peep risk level in the area blocked by adjacent vehicles, and thus trigger graded control strategies and intervention operations in advance to reduce the probability of false emergency braking warnings and accidents. While reducing deployment costs, it improves the user's driving experience.

[0138] This application does not require additional road infrastructure modifications. It can achieve data collection and policy execution by relying solely on existing onboard multi-source sensing equipment. It can be integrated into L1 and above level intelligent driving vehicles and is adaptable to multiple vehicle types and road scenarios, thereby improving the reliability of the system.

[0139] Please see Figure 3 , Figure 3 This application provides a schematic diagram of the structure of an electronic device according to an embodiment of the present application. Figure 3 As shown, the electronic device 300 includes a processor 310, a memory 320, and a bus 330.

[0140] Memory 320 stores machine-readable instructions executable by processor 310. When electronic device 300 is running, processor 310 and memory 320 communicate via bus 330. When the machine-readable instructions are executed by processor 310, they can perform the operations described above. Figure 1 The steps of the ghost camera detection method during driving in the method embodiment shown are described in detail in the method embodiment, and will not be repeated here.

[0141] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, can perform the above-described actions. Figure 1 The steps of the ghost camera detection method during driving in the method embodiment shown are described in detail in the method embodiment, and will not be repeated here.

[0142] 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.

[0143] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0144] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-readable program code.

[0145] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0146] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0147] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0148] This application also provides a computer program product, which includes computer software instructions that, when executed on a processing device, cause the processing device to execute a process for identifying ghost pedestrians during vehicle operation.

[0149] A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this application is generated. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a server or data center that integrates one or more available media. The available medium may be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0150] Those skilled in the art will clearly 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.

[0151] In the several embodiments provided in this application, it should be understood that the disclosed devices, apparatuses, 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, or indirect coupling or communication connection between devices or units, and may be electrical, mechanical, or other forms.

[0152] 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.

[0153] Furthermore, the functional units in the various embodiments of this application 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. The integrated unit can be implemented in hardware or as a software functional unit.

[0154] If the integrated unit 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 application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to 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 application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0155] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

[0156] Although preferred embodiments have been described in this specification, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this specification.

[0157] Obviously, those skilled in the art can make various modifications and variations to this specification without departing from its spirit and scope. Therefore, if such modifications and variations fall within the scope of the claims and their equivalents, this specification is also intended to include such modifications and variations.

Claims

1. A method for identifying "ghost pedestrians" (protruding vehicles) during driving, characterized in that, The method for identifying "ghost pedestrians" during driving includes: Acquire driving behavior data of adjacent vehicles in the adjacent lane of the target vehicle, driving status data of the target vehicle, and environmental data of the current road, wherein the current road is used to characterize the road where the target vehicle is located at the current position; Based on the driving behavior data and preset abnormal behavior rules, the behavior type of the adjacent vehicles is determined, wherein the behavior type includes normal driving behavior or abnormal driving behavior; Based on the behavior type, the driving status data, the environmental data, and the preset ghost peep risk weight ratio, the ghost peep risk level of the target vehicle during the driving process is determined; Based on the risk level of the "ghost peek" (protruding vehicle), a graded control strategy corresponding to the target vehicle is determined.

2. The method for identifying ghost pedestrians during driving as described in claim 1, characterized in that, The step of determining the behavior type of adjacent vehicles based on the driving behavior data and preset abnormal behavior rules includes: The driving behavior data is used to extract change features to determine the driving behavior change features of the adjacent vehicles; When the driving behavior change characteristics meet the preset behavior anomaly rules, the behavior type of the adjacent vehicle is determined to be abnormal behavior, and the behavior confidence score is output. When the driving behavior change characteristics do not meet the preset abnormal behavior rules, the behavior type of the adjacent vehicle is determined to be normal behavior.

3. The method for identifying "ghost pedestrians" during driving as described in claim 2, characterized in that, When the driving behavior change characteristics satisfy the preset abnormal behavior rules, the behavior type of the adjacent vehicle is determined to be abnormal behavior, and the behavior confidence score is output, including: When the driving behavior change characteristics meet the preset behavior anomaly rules, the behavior type of the adjacent vehicle is determined to be a candidate abnormal behavior; The driving behavior change features are input into the trained abnormal behavior learning model, and the behavior confidence score output by the trained abnormal behavior learning model is determined, wherein the behavior confidence score is used to characterize the probability that the behavior type of the adjacent vehicle is a real abnormal behavior. The candidate abnormal behavior is corrected based on the behavior confidence level to determine the behavior type of the adjacent vehicles.

4. The method for identifying ghost pedestrians during driving as described in claim 2, characterized in that, The driving status data of the target vehicle includes the vehicle speed, and the environmental data of the current road includes the current road environment type. Determining the risk level of a pedestrian appearing unexpectedly during the driving process based on the behavior type, the driving status data, the environmental data, and a preset risk weight ratio for such incidents includes: Based on the confidence level of the behavior corresponding to the behavior type, determine the confidence coefficient; Based on the target vehicle speed, determine the corresponding speed coefficient for the target vehicle; Based on the current road environment type, determine the environmental coefficient corresponding to the target vehicle; Based on the confidence coefficient, the vehicle speed coefficient, the environmental coefficient, and the preset ghost peep risk weight ratio, the ghost peep risk level of the target vehicle during driving is determined.

5. The method for identifying "ghost pedestrians" during driving as described in claim 4, characterized in that, The process of determining the risk level of a "ghost peep" during the driving process based on the confidence coefficient, the vehicle speed coefficient, the environmental coefficient, and a preset "ghost peep" risk weight ratio includes: Based on the confidence coefficient, the vehicle speed coefficient, the environmental coefficient, and the preset ghost peep risk weight ratio, the ghost peep risk value of the target vehicle during the driving process is determined; When the risk value of the "ghost peep" is less than a first preset risk threshold, the risk level of the "ghost peep" during the driving process of the target vehicle is determined to be a low risk level. When the risk value of the "ghost peep" is greater than or equal to the first preset risk threshold and less than the second preset risk threshold, the risk level of the "ghost peep" during the driving process of the target vehicle is determined to be a medium risk level. When the risk value of the "ghost peep" is greater than or equal to the second preset risk threshold, the risk level of the "ghost peep" during the driving process of the target vehicle is determined to be high risk.

6. The method for identifying "ghost pedestrians" during driving as described in claim 5, characterized in that, The step of determining the graded control strategy corresponding to the target vehicle based on the risk level of the "ghost peek" includes: When the risk level of the "ghost peek" is low, the target vehicle is controlled to maintain its current driving state, and the driving behavior data of the adjacent vehicles in the adjacent lane are monitored in real time. When the risk level of the "ghost peek" is the medium risk level, the graded control strategy corresponding to the target vehicle is determined to be the first-level warning strategy, and the target vehicle is controlled to enter the pre-deceleration preparation state. When the risk level of the "ghost peek" is high, the corresponding graded control strategy for the target vehicle is determined to be a level two warning strategy, and a graded braking operation is triggered based on the speed of the target vehicle.

7. The method for identifying ghost pedestrians during driving as described in claim 1, characterized in that, The method further includes: Acquire the working status of vehicle-mounted multi-source sensing devices; When a malfunction is detected in the vehicle-mounted multi-source sensing device, a malfunction prompt message is generated, and a manual driving prompt message is output.

8. A device for detecting "ghost pedestrians" (protruding vehicles) during driving, characterized in that, The ghost camera detection device during driving includes: The acquisition module is used to acquire driving behavior data of adjacent vehicles in the adjacent lane of the target vehicle, driving status data of the target vehicle, and environmental data of the current road, wherein the current road is used to characterize the road where the target vehicle is located at the current position; The first determining module is used to determine the behavior type of the adjacent vehicles based on the driving behavior data and preset abnormal behavior rules, wherein the behavior type includes normal driving behavior or abnormal driving behavior. The second determining module is used to determine the risk level of a "ghost peep" during the driving process of the target vehicle based on the behavior type, the driving state data, the environmental data, and a preset "ghost peep" risk weight ratio. The third determining module is used to determine the graded control strategy corresponding to the target vehicle based on the risk level of the "ghost peephole".

9. An electronic device, characterized in that, include: The device includes a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the memory via the bus. The machine-readable instructions are executed by the processor to perform the steps of the ghost-peeping detection method during vehicle operation as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor, performs the steps of the ghost protrusion identification method during driving as described in any one of claims 1-7.

Images