Driving risk identification early warning method, system, equipment and medium

Abstract

The invention relates to a driving risk identification and early warning method, system and device and a medium. The method comprises the following steps: collecting a driving environment and driving data of a vehicle; fitting real-time parameters of a high-risk driving model through a genetic algorithm based on a vehicle driving environment and driving data to obtain model parameters representing driver operation characteristics and vehicle motion characteristics; based on the model parameters obtained through fitting, high-risk driving behaviors are classified and recognized through an SVM classification algorithm, and early warning is conducted on driving risks. According to the method, traffic safety is used as guidance, driving risk comprehensive evaluation and early warning under coupling influence factors are carried out by using driving data, namely, the coupling influence relationship of a driver-vehicle system, a road environment, individual characteristics of a driver and a driving state is comprehensively considered, and the driving risk is identified and early warned. The method can be widely applied to the field of driving safety.

Description

Technical Field

[0001] The present invention belongs to the field of driving safety, and in particular relates to a driving risk identification and early warning method, system, equipment and medium based on driving data. Background Art

[0002] In order to ensure road traffic driving safety, it is crucial to provide early warning of possible risks in driving. At present, a key method for identifying and warning of driving risks is to monitor the driver's status and safety assessment, and to provide early warning of driving risks. Since traffic accidents are largely related to drivers, the impact of drivers on driving risks is closely related to the driver's individual characteristics, perception and decision-making characteristics, and driving status. Therefore, the focus of driving safety assurance is to assess the safety of drivers and provide early warning of driving risks. Therefore, the core challenge of driving safety assurance and risk warning lies in being able to comprehensively explore and systematically quantify driver safety characteristics and identify indicators of driving risks. This is not only a key technical challenge facing the current research field, but also a hot issue of common concern to academia and industry.

[0003] With the technological advancement of sensors such as radar and cameras, the assisted driving perception capabilities of driving vehicles and driver monitoring systems have also been further developed. Existing studies have begun to conduct driver status safety assessments from the perspectives of driver facial video, driver operating characteristics, driver physiological characteristics, and voice in the cockpit. For example, by monitoring the driver's facial feature points and extracting ROI features, indicators such as the percentage of eye closure time, continuous eye closure duration, blinking frequency, and mouth aspect ratio are obtained, and the driver's fatigue state is monitored in combination with heart rate, electromyography and other indicators; the driver's emotional state is monitored through the phonetic features of the driver's voice, such as pitch, tone envelope, and spectrum; the driver's distraction state is monitored by the driver's line of sight focus position combined with indicators such as the driver's actions in the cockpit.

[0004] In recent years, with the development of artificial intelligence technology, algorithms such as machine vision and affective computing have made significant progress in monitoring driver fatigue, emotions, distraction and other risk states. However, these methods often monitor, evaluate and warn drivers based on a single state, rarely considering the impact of the driver's personal characteristics and the traffic environment. Therefore, it is difficult to unify the warning thresholds for each risk state, and there is also a lack of evaluation methods for coupled risks under the influence of multiple factors. Summary of the invention

[0005] In view of the above problems, the purpose of the present invention is to provide a driving risk identification and warning method, system, equipment and medium based on driving data, which is guided by traffic safety and uses driving data to carry out comprehensive evaluation and warning of driving risks under coupled influencing factors, that is, comprehensively considering the coupled influence relationship between the "driver-vehicle" system affected by the road environment and the driver's individual characteristics and driving status, and identifying and warning of driving risks.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] In a first aspect, the present invention provides a driving risk identification and warning method, comprising the following steps:

[0008] Collect the vehicle's driving environment and driving data;

[0009] Based on the vehicle driving environment and driving data, the real-time parameters of the high-risk driving model are fitted through genetic algorithms to obtain model parameters that characterize the driver's operating characteristics and vehicle motion characteristics;

[0010] Based on the fitted model parameters, the SVM classification algorithm is used to classify and identify high-risk driving behaviors, and to issue early warnings for driving risks.

[0011] Furthermore, the collection of the driving environment and driving data of the vehicle includes:

[0012] The driving data of the main vehicle in road traffic is collected through GPS and IMU sensors, and the vehicle motion is modeled in the Frenet coordinate system;

[0013] Collect road environment data around the vehicle and the location and speed information of other traffic participants through joint perception of radar and camera;

[0014] The main vehicle's driving data, road environment data, and the position and speed data of other traffic participants within a preset time period are integrated to form the main vehicle's continuous driving time series data.

[0015] Furthermore, the real-time parameters of the high-risk driving model are fitted by genetic algorithm based on the vehicle driving environment and driving data to obtain model parameters that characterize the driver's operating characteristics and the vehicle's motion characteristics, including:

[0016] According to the vehicle motion characteristics and risk characteristics, the risk perception and expected trade-off decision-making process of the driving subject are comprehensively considered, the vehicle motion process is modeled, and a high-risk driving model is obtained;

[0017] Based on the vehicle driving environment, driving data and high-risk driving model, a genetic algorithm is used to fit the parameters of the high-risk driving model to obtain model parameters that characterize the driver's operating characteristics and vehicle motion characteristics.

[0018] Furthermore, the vehicle motion characteristics and risk characteristics are comprehensively considered, the risk perception and expected trade-off decision-making process of the driving subject are comprehensively considered, and the vehicle motion process is modeled to obtain a high-risk driving model, including:

[0019] The risk field theory is used to model the driver's perceived risk distribution of the current road space, and the comprehensive risk of the road space is quantified based on the risk distribution results;

[0020] The driving subject's expectation of the current driving task is quantified by the ratio of the target displacement to the expected maximum displacement;

[0021] Based on the driving risk perceived by the driving subject and his expectations for the current driving task, the driving decision-making process is modeled, and the risk preference and decision-making characteristics of the driving subject are quantified using model parameters. The decision target point of the driving subject at each time step is obtained to achieve driving simulation.

[0022] Furthermore, the risk field theory is used to model the driver's perceived risk distribution of the current road space, which is expressed as:

[0023]

[0024] In the formula, R represents the risk, (d, s) represent the position coordinates of the main vehicle in the Frenet reference system, and v j,s With v j,d denote the longitudinal and transverse velocities of object j in the road space, T 0 Indicates the driver's reaction time; k s , k d represents a constant parameter that is compatible with the driver's risk perception characteristics; β is a risk factor parameter that considers the severity of the collision consequences; Δs j ,Δd j are the longitudinal and lateral distances from object j to the host vehicle in the road space, respectively.

[0025] Furthermore, the driving subject's expectation for the current driving task is quantified by the ratio of the target displacement to the expected maximum displacement, which is expressed as:

[0026]

[0027] Where Δs represents the longitudinal displacement of the target point relative to the current position; P exp is the driver's maximum expectation of his own movement distance; v s Indicates the current longitudinal speed of the vehicle, a exp represents the driver's expected acceleration, T pre Indicates the driver's preview time.

[0028] Furthermore, based on the driving risk R perceived by the driving subject and its expectation p for the current driving task, the driving decision process is modeled, and the risk preference and decision characteristics of the driving subject are quantified with model parameters, the decision target point of the driving subject at each time step is obtained, and driving simulation is realized, including:

[0029] The driving decision process is modeled as a trade-off between the driver's perceived driving risk R and the driver's expectation of the driving task p, and the relationship between the two is described by the decision curve;

[0030] According to the utility theory, the decision characteristics of drivers are modeled as decision curves, which are divided into four types: neutral, aggressive, conservative, and mixed;

[0031] The feasible domain of decision making for the expected trade-off of driving risk is determined, and the risk preference and decision-making characteristics of the driving subject are quantified using a high-risk driving model parameter set.

[0032] In a second aspect, the present invention provides a driving risk identification and warning system, comprising:

[0033] Real-time data acquisition module, used to collect the vehicle's driving environment and driving data;

[0034] The high-risk driving model real-time parameter fitting module is used to fit the real-time parameters of the high-risk driving model based on the vehicle driving environment and driving data through a genetic algorithm to obtain model parameters that characterize the driver's operating characteristics and vehicle motion characteristics;

[0035] The driving risk identification and warning module is used to classify and identify high-risk driving behaviors based on the high-risk driving model parameters obtained by fitting driving data through the SVM classification algorithm, and to issue warnings for driving risks.

[0036] In a third aspect, the present invention provides a computer-readable storage medium storing one or more programs, wherein the one or more programs include instructions, which, when executed by a computing device, cause the computing device to perform any method.

[0037] In a fourth aspect, the present invention provides a computing device, comprising: one or more processors and a memory, wherein the memory stores one or more programs and is configured to be executed by the one or more processors, and the one or more programs include instructions for executing any method.

[0038] The present invention adopts the above technical solution, which has the following advantages:

[0039] 1. The present invention combines driving environment data with driving data, digitally fits a high-risk driving model, and parameterizes the driving risk characteristics, so as to perform comprehensive driving risk identification and early warning under the influence of the "driver-vehicle" system coupling state. This method makes full use of the vehicle's motion characteristics, which is a direct influencing factor of driving risk, and quantitatively characterizes the influence of coupling factors such as the driver's individual characteristics, decision-making characteristics, operating characteristics, and driving status on driving risk. It also extracts the parameters of driving risk by fitting the parameters of the high-risk driving model, and can regard the driver and the vehicle as a whole. Compared with the monitoring and early warning method for a single driver state, it can more systematically and fully reflect the coupling influence of different factors on driving risk, thereby giving a more accurate evaluation of the overall driving risk.

[0040] 2. The risk identification and early warning based on model fitting parameters proposed in the present invention can improve the calculation speed, perform real-time parameterized characterization and early warning of driving risks, thereby reducing human-induced traffic accidents and improving the level of road traffic safety.

[0041] Therefore, the present invention can be widely applied to the field of driving safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Throughout the accompanying drawings, the same reference numerals are used to represent the same components. In the accompanying drawings:

[0043] Figure 1 is a flow chart of a driving risk identification and warning method provided in an embodiment of the present invention;

[0044] Figure 2 It is a quantitative schematic diagram of road space comprehensive risk modeling provided in an embodiment of the present invention;

[0045] Figure 3 Schematic diagrams of different decision curves provided in embodiments of the present invention;

[0046] Figure 4 It is a schematic diagram of the feasible domain of decision making for the expected trade-off of driving risk provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solution and advantages of the embodiment of the present invention clearer, the technical solution of the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings of the embodiment of the present invention. Obviously, the described embodiment is a part of the embodiment of the present invention, not all of the embodiments. Based on the described embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the present invention.

[0048] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0049] Driving risk monitoring and early warning systems are in a period of rapid development. Current research is mostly focused on the development of monitoring indicators and early warning algorithms for risk states. Few studies have proposed comprehensive early warning methods for coupled states, that is, how to conduct risk assessment and early warning for the complex "driver-vehicle" system under the influence of multiple factors such as driver characteristics, namely different driving states, which has not yet received sufficient research and technical support.

[0050] Therefore, in some embodiments of the present invention, a driving risk identification and warning method is provided, starting from the movement and interaction characteristics of the "driver-vehicle" system and other traffic participants, which are the direct influencing factors of driving risk, and parameterizing the driver's risk perception and decision-making characteristics to quantify driving risks, conduct coupling state safety assessments, and issue warnings for high-risk driving behaviors. At the same time, the simulation results of natural driving data sets are used to verify that new ideas are provided for the above-mentioned problems, and technical support is provided for driver safety monitoring and traffic safety assurance.

[0051] Correspondingly, in other embodiments of the present invention, a driving risk identification and warning system, device and medium are provided.

[0052] Example 1

[0053] like Figure 1 As shown, the present invention proposes a driving risk identification and early warning method, which collects road traffic and driving data, fits high-risk driving model parameters, and carries out driving risk identification and early warning based on characteristic parameters, which can effectively carry out comprehensive driving risk assessment and safety early warning based on multiple factors such as individual characteristics of drivers, decision-making characteristics, and driving status. Specifically, it includes the following steps:

[0054] Step 1: Collect the vehicle's driving environment and driving data;

[0055] Step 2: Based on the vehicle driving environment and driving data, the real-time parameters of the high-risk driving model are fitted by genetic algorithm to obtain the model parameters that characterize the driver's operating characteristics and the vehicle's motion characteristics;

[0056] Step 3: Based on the fitted model parameters, the SVM classification algorithm is used to classify and identify high-risk driving behaviors and issue warnings for driving risks.

[0057] Furthermore, in the above step 1, when collecting the driving environment and driving data of the vehicle, the following steps are included:

[0058] Step 1.1: Collect driving data such as the position, speed, and acceleration of the main vehicle (i.e., the vehicle driven by the driver) in road traffic through sensors such as GPS and IMU, and model the vehicle motion in the Frenet coordinate system;

[0059] Step 1.2: Collect road environment data around the vehicle, such as lanes, obstacles, and traffic signals, as well as the location and speed information of other traffic participants, such as motor vehicles, non-motor vehicles, and pedestrians, through the joint perception of radar and camera;

[0060] Step 1.3: The driving data of the main vehicle, the road environment data, and the position and speed data of other traffic participants within a preset time period are integrated to form the continuous driving time series data of the main vehicle.

[0061] Furthermore, in the above step 2, based on the vehicle driving environment and driving data, the real-time parameters of the high-risk driving model are fitted by a genetic algorithm to obtain model parameters that characterize the driver's operating characteristics and the vehicle's motion characteristics, including the following steps:

[0062] Step 2.1: Based on the vehicle motion characteristics and risk characteristics, the risk perception and expected trade-off decision-making process of the driving subject are comprehensively considered to model the vehicle motion process, obtain a high-risk driving model, and realize the simulation of high-risk vehicle motion driving behavior;

[0063] Step 2.2: Based on the vehicle driving environment, driving data and high-risk driving model, a genetic algorithm is used to fit the parameters of the high-risk driving model to obtain model parameters that characterize the driver's operating characteristics and the vehicle's motion characteristics.

[0064] Furthermore, in the above step 2.1, this embodiment proposes a high-risk driving model in order to parameterize and characterize the risk of the driving process. The high-risk driving model consists of two parts: risk perception and driving decision-making.

[0065] Specifically, the steps include:

[0066] Step 2.1.1: Use risk field theory to model the driver's perceived risk distribution R of the current road space, and quantify the comprehensive risk of the road space based on the perceived risk distribution.

[0067] In this embodiment, the driving risk perception process uses the risk field theory to model the risk distribution of the current road space:

[0068]

[0069] In the formula, R represents the risk, (d, s) represent the position coordinates of the main vehicle in the Frenet reference system, and v j,s With v j,d They represent the longitudinal and lateral velocities of object j in the road space along the Frenet reference frame, T 0 Indicates the driver's reaction time; k s , k d represents a constant parameter that is compatible with the driver's risk perception characteristics; β is a risk factor parameter that considers the severity of the collision consequences; Δs j ,Δd j are the longitudinal and lateral distances from object j in the road space to the host vehicle respectively. Based on the above formula, the comprehensive risk modeling of road space is quantified as follows: Figure 2 The driver’s perception of driving risk has a certain time delay compared to the actual road conditions, which is determined by the driver’s reaction time T 0 Decide.

[0070] Step 2.1.2: Quantify the driving subject's expectation p for the current driving task by the ratio of the target displacement to the expected maximum displacement.

[0071] Specifically, if the driver drives in a manner that best meets his driving expectations, during the preview time T pre After that, the driver’s maximum expectation of his own moving distance can be expressed by P exp To quantify:

[0072]

[0073] In the formula, v s Indicates the current longitudinal speed of the vehicle, a exp represents the driver's expected acceleration, T pre Indicates the driver's preview time.

[0074] The driving subject has a certain expectation p for the current driving task, which is quantified by the ratio of the target displacement to the expected maximum displacement in this embodiment:

[0075]

[0076] Where Δs represents the longitudinal displacement of the target point relative to the current position.

[0077] Step 2.1.3: Based on the driving risk R perceived by the driving subject and his expectation p for the current driving task, the driving decision-making process is modeled, and the risk preference and decision-making characteristics of the driving subject are quantified using model parameters to obtain the decision target point of the driving subject at each time step, thereby realizing driving simulation.

[0078] Specifically, the steps include:

[0079] ① The driving decision process is modeled as a trade-off between the driver's perceived driving risk R and the driver's expectation of the driving task p. The relationship between the two is described by the decision curve:

[0080] p=R γ

[0081] Where γ is the driver's decision-making characteristic parameter.

[0082] ② According to the utility theory, the driver's decision characteristics are modeled as decision curves, which are divided into four types: neutral, radical, conservative and mixed.

[0083] The driver's acceptance of driving risk and driving expectation level can be modeled through the decision curve. Corresponding to different driving expectation levels, the driver's acceptable driving risk varies. Different drivers have different trade-off characteristics between driving risk and driving expectation realization. Referring to the utility theory, the driver's decision-making style can be divided into four types: neutral, radical, conservative and mixed. The decision curves of different decision-making styles have different characteristics. The decision curves are divided into four types: neutral, radical, conservative and mixed. For example Figure 3 The driver's risk preference and decision-making characteristics are quantified by the decision curve parameters, and the driver's decision target point at each time step is obtained, thereby realizing driving simulation.

[0084] ③ Determine the feasible domain of decision-making for the expected trade-off of driving risk, and quantify the risk preference and decision-making characteristics of the driving subject using a high-risk driving model parameter set.

[0085] Considering that the driver has a certain acceptable risk level R under zero expectation a , so the lower bound of the driver's decision feasible region can be expressed as follows:

[0086]

[0087] Based on the risk balance theory, it can be seen that the driver will not blindly pursue the maximum expected level when driving the vehicle. Comfort is also one of the factors affecting the driver's behavior. Therefore, in the decision curve pL In the upper region, the driver often chooses to operate the vehicle so that its driving state is within an acceptable range. Therefore, the upper bound of the decision feasible region can be quantified by the following formula:

[0088]

[0089] where ω 1 represents the weight of the conservative component in the hybrid decision curve, 0 ≤ ω 1 ≤ 1, and the larger ω 1 is, the greater the proportion of the risk-averse component in the driver's decision preference, and the more conservative the driver's driving style;

[0090] ω 2 represents the weight of the aggressive component in the hybrid decision curve, and ω 1 + ω 2 = 1; the larger ω 2 is, the greater the proportion of the risk-preference component in the driver's decision preference, and the more aggressive the driver's driving style;

[0091] γ 1 is a characteristic parameter of the conservative decision curve, representing the degree of risk aversion of the conservative component in the decision curve, with γ 1 ≥ 1; and the larger γ 1 is, the farther the conservative decision curve is from the neutral decision curve, indicating that in the conservative decision curve, the degree of risk aversion is higher and the driver's style is more conservative;

[0092] γ 2 is a characteristic parameter of the aggressive decision curve, representing the degree of risk aversion of the aggressive component in the decision curve, with 0 < γ 2 ≤ 1; and the larger γ 2 is, the closer the aggressive decision curve is to the neutral decision curve, indicating that in the aggressive decision curve, the degree of risk aversion is higher and the driver's style is more conservative;

[0093] R a represents the maximum risk level that the driver can accept when the expected achievement level is 0, 0 < R a < 1; and the larger R a is, the greater the risk that the driver can accept under zero expectation, and the more aggressive the driver's style;

[0094] m represents the maximum expected achievement level that the driver is willing to pursue under zero risk, 0 < m < 1; and the larger m is, the higher the driver's pursuit of achieving the expectation, showing a more aggressive driving style;

[0095] R m represents the maximum risk level that the driver can accept under any circumstances, with Ra <R m <1; and R m The larger it is, the higher the driver's tolerance for risk and the more aggressive the driver's style;

[0096] T delay It represents the delay time of the driver from perceiving the current state to making a decision and responding to the change of vehicle acceleration; T delay The larger it is, the longer the driver's reaction time is.

[0097] Based on this, the feasible domain of decision-making for the trade-off between driving risk and expectation can be expressed as Figure 4 In the middle gray area, the high-risk driving model parameter set is set to {ω 1 ,ω 2 ,γ 1 ,γ 2 ,R a ,m,R m ,R delay}.

[0098] Furthermore, in the above step 2.2, during the actual driving process, the road traffic timing information and the driving data of the main vehicle during the driving process can be collected by the sensor in step 1, and the risk characteristics and parameters of the driving segment can be extracted by combining the construction of the high-risk driving behavior model and parameter fitting.

[0099] Based on the high-risk driving model constructed in step 2.1 and the road environment and driving data of the main vehicle collected in step 1, the parameters of the driving model are fitted using a genetic algorithm to obtain the real-time parameter set {ω 1 ,ω 2 ,γ 1 ,γ 2 ,R a ,m,R m ,T delay}.

[0100] Furthermore, in the above step 3, the parameter set {ω 1 ,ω 2 ,γ 1 ,γ 2 ,R a ,m,R m ,T delay}, the SVM classification method can be used to classify driving risks by fitting the model parameter values ​​obtained by fitting the high-risk driving model in the actual driving process as input. Different driving model parameter inputs will be classified as "medium-low risk" or "high risk", thereby warning high-risk driving. Considering that the number of high-risk samples is much smaller than the number of medium- and low-risk samples, a higher weight is set for high-risk samples to balance the impact of the imbalanced sample size. The driving risk identification and warning method was verified, and the accuracy of high-risk driving behavior identification reached 94.6%.

[0101] Example 2

[0102] This embodiment further introduces the driving risk identification and warning method based on driving data provided by the present invention.

[0103] The driving data used in this embodiment comes from CitySim aerial vehicle trajectory data.

[0104] CitySim is a large-scale, high-precision drone aerial vehicle trajectory open source dataset launched by researchers from Tongji University and the University of Central Florida in 2022. The data covers driving scenarios in multiple countries, multiple weather conditions, and multiple locations, and collects driving data on various road types, including basic sections of highways, highway entrance and exit ramps, urban expressways, signalized intersections, roundabouts, etc. The sampling frame rate is 30fps, and the total recording time is more than 1200 minutes. The CitySim dataset provides 7 key point positioning of the vehicle, which can help researchers know the position and boundary information of the vehicle on the road, and reconstruct the driving scene in the simulated driving scene, providing very valuable data support for the simulation research of natural driving vehicle motion.

[0105] The high-risk driving model is simulated and verified by using the natural driving time series data in the CitySim dataset. The accuracy of the model in describing the driving subject and vehicle motion characteristics can be obtained. The quantitative comparison of the accuracy between different models is shown in Table 1.

[0106] Table 1 Errors of various driving models on the CitySim dataset

[0107] Driving Model Model Error Driving Model Model Error BL-OV 8.966 BLVD 10.255 FVD 4.810 GHR 1.776 Gipps 0.346 IBDM 3.914 IDM 3.647 OVM 5.250 RRDM 0.065 This model 0.0199

[0108] It can be seen from the table that the high-risk driving model used in the present invention can well simulate the movement of a naturally driven vehicle, and the corresponding parameters such as reaction time can well characterize the individual characteristics, decision-making characteristics, and operating characteristics of the driver.

[0109] Based on the above high-risk driving model, the collected road environment and driving data of the main vehicle, the genetic algorithm is used to fit the parameters of the driving model, so as to obtain the real-time parameter set {ω 1 ,ω 2 ,γ 1 ,γ 2 ,R a ,m,R m ,T delay}.

[0110] Based on the parameter set {ω 1 ,ω 2 ,γ 1 ,γ 2 ,R a ,m,R m ,T delay}, the SVM classification method can be used to classify driving risks by fitting the model parameter values ​​obtained by fitting the high-risk driving model in the actual driving process as input. Different driving model parameter inputs will be classified as "medium-low risk" or "high risk", thereby warning high-risk driving. Considering that the number of high-risk samples is much smaller than the number of medium- and low-risk samples, a higher weight is set for high-risk samples to balance the impact of the imbalanced sample size. The driving risk identification and warning method was verified, and the accuracy of high-risk driving behavior identification reached 94.6%.

[0111] Example 3

[0112] The above-mentioned embodiment 1 provides a driving risk identification and warning method based on driving data. Correspondingly, this embodiment provides a driving risk identification and warning system based on driving data. The system provided in this embodiment can implement the driving risk identification and warning method based on driving data of embodiment 1, and the system can be implemented by software, hardware, or a combination of software and hardware. For example, the system may include integrated or separate functional modules or functional units to execute the corresponding steps in each method of embodiment 1. Since the system of this embodiment is basically similar to the method embodiment, the process described in this embodiment is relatively simple, and the relevant parts can refer to the partial description of embodiment 1. The embodiment of the system provided in this embodiment is only illustrative.

[0113] The driving risk identification and warning system provided in this embodiment includes:

[0114] Real-time data acquisition module, used to collect the vehicle's driving environment and driving data;

[0115] The high-risk driving model real-time parameter fitting module is used to fit the real-time parameters of the high-risk driving model based on the vehicle driving environment and driving data through a genetic algorithm to obtain model parameters that characterize the driver's operating characteristics and vehicle motion characteristics;

[0116] The driving risk identification and warning module is used to classify and identify high-risk driving behaviors based on the high-risk driving model parameters obtained by fitting driving data through the SVM classification algorithm, and to issue warnings for driving risks.

[0117] Example 4

[0118] This embodiment provides a processing device corresponding to the driving risk identification and warning method based on driving data provided in this embodiment 1. The processing device can be a processing device for a client, such as a mobile phone, a laptop computer, a tablet computer, a desktop computer, etc., to execute the method of embodiment 1.

[0119] The processing device includes a processor, a memory, a communication interface and a bus, and the processor, the memory and the communication interface are connected through the bus to complete mutual communication. The memory stores a computer program that can be run on the processor, and the processor executes the driving risk identification and warning method based on driving data provided in this embodiment 1 when running the computer program.

[0120] Preferably, the memory may be a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory.

[0121] Preferably, the processor may be a central processing unit (CPU), a digital signal processor (DSP) or other general-purpose processors of various types, which are not limited here.

[0122] Example 5

[0123] The driving risk identification and warning method based on driving data of this embodiment 1 can be specifically implemented as a computer program product, which may include a computer-readable storage medium on which computer-readable program instructions for executing the driving risk identification and warning method based on driving data described in this embodiment 1 are loaded.

[0124] Computer readable storage media can be tangible devices that hold and store instructions used by instruction execution devices. Computer readable storage media can be, for example, but not limited to, electronic storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any combination thereof.

[0125] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0126] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0127] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0128] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0129] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A driving risk identification and warning method, characterized in that: The following steps are involved: Collect the vehicle's driving environment and driving data; Based on the vehicle driving environment and driving data, the real-time parameters of the high-risk driving model are fitted through genetic algorithms to obtain model parameters that characterize the driver's operating characteristics and vehicle motion characteristics; Based on the fitted model parameters, the SVM classification algorithm is used to classify and identify high-risk driving behaviors, and to issue early warnings for driving risks.

2. A driving risk identification and warning method as claimed in claim 1, characterized in that: The collecting of the driving environment and driving data of the vehicle includes: Through GPS and IMU sensors, the driving data of the main vehicle in road traffic is collected and the Frenet coordinate system is constructed; Collect road environment data around the vehicle and the location and speed information of other traffic participants through joint perception of radar and camera; The main vehicle's driving data, road environment data, and the position and speed data of other traffic participants within a preset time period are integrated to form the main vehicle's continuous driving time series data.

3. A driving risk identification and warning method as claimed in claim 1, characterized in that: Based on the vehicle driving environment and driving data, the real-time parameters of the high-risk driving model are fitted by genetic algorithm to obtain model parameters that characterize the driver's operating characteristics and the vehicle's motion characteristics, including: According to the vehicle motion characteristics and risk characteristics, the risk perception and expected trade-off decision-making process of the driving subject are comprehensively considered, the vehicle motion process is modeled, and a high-risk driving model is obtained; Based on the vehicle driving environment, driving data and high-risk driving model, a genetic algorithm is used to fit the parameters of the high-risk driving model to obtain model parameters that characterize the driver's operating characteristics and vehicle motion characteristics.

4. A driving risk identification and warning method as claimed in claim 3, characterized in that: The vehicle motion characteristics and risk characteristics are comprehensively considered, the risk perception and expected trade-off decision-making process of the driving subject are comprehensively considered, and the vehicle motion process is modeled to obtain a high-risk driving model, including: The risk field theory is used to model the driver's perceived risk distribution of the current road space, and the comprehensive risk of the road space is quantified based on the risk distribution results; The driving subject's expectation of the current driving task is quantified by the ratio of the target displacement to the expected maximum displacement; Based on the driving risk perceived by the driving subject and his expectations for the current driving task, the driving decision-making process is modeled, and the risk preference and decision-making characteristics of the driving subject are quantified using model parameters. The decision target point of the driving subject at each time step is obtained to achieve driving simulation.

5. A driving risk identification and warning method as claimed in claim 3, characterized in that: The risk field theory is used to model the driver's perceived risk distribution of the current road space, and the calculation formula is: In the formula, R represents the risk, (d, s) represent the position coordinates of the main vehicle in the Frenet reference system, and v j,s With v j,d They represent the longitudinal and lateral speeds of object j in the road space along the Frenet reference system, T0 represents the driver's reaction time; k s , k d represents a constant parameter that is compatible with the driver's risk perception characteristics; β is a risk factor parameter that considers the severity of the collision consequences; Δs j ,Δd j are the longitudinal and lateral distances from object j to the host vehicle in the road space, respectively.

6. A driving risk identification and warning method as claimed in claim 3, characterized in that: The driving subject's expectation for the current driving task is quantified by the ratio of the target displacement to the expected maximum displacement, which is expressed as: Where Δs represents the longitudinal displacement of the target point relative to the current position; P exp is the driver's maximum expectation of his own movement distance; v s Indicates the current longitudinal speed of the vehicle, a exp represents the driver's expected acceleration, T pre Indicates the driver's preview time.

7. A driving risk identification and warning method as claimed in claim 3, characterized in that: The driving decision-making process is modeled based on the driving risk R perceived by the driving subject and its expectation p for the current driving task, and the risk preference and decision-making characteristics of the driving subject are quantified with model parameters, the decision target point of the driving subject at each time step is obtained, and the driving simulation is realized, including: The driving decision process is modeled as a trade-off between the driver's perceived driving risk R and the driver's expectation of the driving task p, and the relationship between the two is described by the decision curve; According to the utility theory, the decision characteristics of drivers are modeled as decision curves, which are divided into four types: neutral, aggressive, conservative, and mixed; The feasible domain of decision making for the expected trade-off of driving risk is determined, and the risk preference and decision-making characteristics of the driving subject are quantified using a high-risk driving model parameter set.

8. A driving risk identification and warning system, characterized in that: include: Real-time data acquisition module, used to collect the vehicle's driving environment and driving data; The high-risk driving model real-time parameter fitting module is used to fit the real-time parameters of the high-risk driving model based on the vehicle driving environment and driving data through a genetic algorithm to obtain model parameters that characterize the driver's operating characteristics and vehicle motion characteristics; The driving risk identification and warning module is used to classify and identify high-risk driving behaviors based on the high-risk driving model parameters obtained by fitting driving data through the SVM classification algorithm, and to issue warnings for driving risks.

9. A computer-readable storage medium storing one or more programs, characterized in that: The one or more programs include instructions, which, when executed by a computing device, cause the computing device to perform any one of the methods of claims 1 to 7.

10. A computing device, characterized in that include: One or more processors and a memory, wherein the memory stores one or more programs and is configured to be executed by the one or more processors, wherein the one or more programs include instructions for executing any one of the methods described in claims 1 to 7.

Images