A method for identifying abnormal parking on highways based on vehicle-to-everything (V2X) big data

Abstract

This invention provides a high-speed abnormal parking identification method based on vehicle network big data, including the following steps: S1, collecting vehicle status data, behavior data, and GPS data from the vehicle terminal and transmitting them to the cloud; preprocessing the vehicle status data, behavior data, and GPS data to obtain vehicle speed, rapid acceleration, rapid deceleration, sharp turns, tire pressure, engine coolant temperature, engine status changes, hazard light activation status, whether the driver has exited the vehicle, and GPS information; S2, constructing a high-speed parking model trigger based on the vehicle speed and hazard light activation status; This method can effectively and quickly monitor whether vehicles are abnormally parking on highways, and by combining vehicle signals and driver behavior recognition, it eliminates the need for aftermarket sensors, featuring low cost and high accuracy, thus providing OEMs with after-sales care for accidents, ensuring driver safety, and laying the foundation for after-sales service.

Description

Technical Field

[0001] This invention relates to the field of big data processing technology, and more specifically, to a method for identifying abnormal parking on high speeds based on vehicle network big data. Background Technology

[0002] Abnormal vehicle parking is becoming increasingly common. This not only disrupts the normal driving of other vehicles but also negatively impacts road safety and traffic efficiency. Meanwhile, vehicle accidents and malfunctions remain a key research focus for all automakers and suppliers. Highways, being a special type of road, require automakers to provide rapid assistance and roadside aid to drivers after incidents, ensuring their safety. Furthermore, highway accidents and breakdowns often necessitate towing, leading to high repair costs. Therefore, developing a method and device for identifying abnormal parking on highways has become a crucial technological need. Existing technologies employ methods such as onboard sensors and video analytics to identify abnormal parking behavior. However, these methods suffer from drawbacks, including high installation costs and relatively low accuracy. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a method for identifying abnormal parking on highways based on vehicle network big data, which can quickly and accurately identify abnormal parking situations on highways, in order to address the shortcomings of the above-mentioned technical solutions.

[0004] This invention provides a method for identifying abnormal parking on highways based on vehicle-to-everything (V2X) big data, the method comprising the following steps:

[0005] S1 collects vehicle status data, behavior data, and GPS data from the vehicle's infotainment system and transmits them to the cloud. It then preprocesses the vehicle status data, behavior data, and GPS data to obtain vehicle speed, rapid acceleration, rapid deceleration, sharp turns, tire pressure, engine coolant temperature, engine status changes, hazard lights status, whether the vehicle has been exited, and GPS information.

[0006] S2, construct a high-speed parking model trigger based on the vehicle speed and hazard lights being on; and determine whether the vehicle needs to trigger the highway abnormal parking judgment based on the triggering conditions of the high-speed parking model trigger.

[0007] S3. Based on GPS point positioning data, a high-speed positioning judgment model is constructed based on the two-point-one-line method. For vehicles that meet the triggering conditions of the high-speed parking model trigger, it is further determined whether the current vehicle stops due to vehicle abnormality and is not in a service area while driving on the highway.

[0008] S4. Based on the high-speed positioning judgment model, a high-speed traffic jam behavior recognition model is constructed. When a vehicle has multiple similar stopping behaviors in its historical driving trajectory data on the highway and there is no data on returning to the factory for maintenance, and the current vehicle speed is less than the threshold a1, it is considered that the vehicle may have stopped due to high-speed traffic jam behavior, and such vehicle owner information is marked.

[0009] S5. Based on the highway traffic jam behavior recognition model, construct a behavior discrimination model to further determine whether vehicles that meet both step S2 and step S3 have stopped due to an abnormality on the highway.

[0010] S6. If the conditions of step S5 are met but not step S4 are met, determine the time when the vehicle experienced abnormal high-speed parking behavior based on the behavior discrimination model. If abnormal parking behavior occurs, store the vehicle information that experienced abnormal high-speed parking behavior in the database.

[0011] S7: Obtain all vehicle information from the database that has been communicated and confirmed with the vehicle owner for all vehicles that have experienced abnormal parking behavior on the highway, and build a machine learning intelligent classification model based on all vehicle information that has been communicated and confirmed with the vehicle owner; classify the vehicles that have gone through steps S2, S3 and S4 through the machine learning intelligent classification model, and identify the vehicles that have experienced abnormal parking on the highway and the corresponding fault and accident scenarios.

[0012] In the highway abnormal parking identification method based on vehicle network big data described in this invention; in step S2: the triggering conditions of the highway parking model trigger include triggering conditions constructed based on the vehicle speed and triggering conditions constructed based on the hazard lights being on. When both the triggering conditions constructed based on the vehicle speed and the triggering conditions constructed based on the hazard lights being on are met simultaneously within a certain time period, the highway abnormal parking judgment is triggered; otherwise, the highway abnormal parking judgment is not triggered.

[0013] In the high-speed abnormal parking identification method based on vehicle network big data described in this invention, the triggering condition based on vehicle speed is that the vehicle is stationary within a continuous time interval (t0-t(N)), i.e., the vehicle speed is 0 at the current moment. The vehicle speed is expressed as... Where ceil represents rounding up, the... Indicates t i The velocity data is collected at the current moment. 0 indicates that the velocity is 0 at the current moment, and t0 indicates the start time of the model trigger.

[0014] In the high-speed abnormal parking identification method based on vehicle network big data described in this invention, the triggering condition constructed based on the hazard light activation status is that the hazard lights are activated within a continuous time interval (t0-t(N)), and the hazard light switching status is expressed as follows: Among them, the Indicates t i The data collected shows the status of the hazard lights switch at the time of acquisition. 1 indicates that the hazard lights are on, and t0 indicates the start time of the model trigger.

[0015] In the high-speed abnormal parking identification method based on vehicle network big data described in this invention, step S3 further includes the following steps:

[0016] S31, using GPS point positioning data, quickly determine whether the vehicle is located on the highway based on the two-point-one-line method; obtain the GPS point positioning data corresponding to three points A, B, and C, where point B is the GPS point positioning data at time t0 in step S2, and point A is the GPS point positioning data corresponding to the N1 minutes before time t0. 0-N1 GPS location data at time t0, where point C is the location of point N minutes after time t0. 0+N The GPS location data at any given time; the GPS location data of points A, B, and C are used to obtain addresses through reverse geocoding, and then fuzzy matching is used to obtain addresses on the highway. The addresses are used to exclude highway locations that are parked in highway service areas and parking areas. If points A, B, and C are all on the highway at the same time, the vehicle is considered to meet the triggering conditions and is on the highway.

[0017] In the high-speed abnormal parking identification method based on vehicle network big data described in this invention, step S5 includes the following parking behaviors:

[0018] A. Before and after the triggering condition in step S2 is triggered at time t, detect the vehicle exit detection period Δγ after time t to see if the driver has exited the vehicle, i.e., whether the driver has unbuckled the seat belt or the left front door of the vehicle has opened.

[0019] B, before and after the triggering condition of step S2 is triggered at time t, detect whether the driver opened the trunk during the vehicle exit detection period after time t, that is, whether there was an act of opening the trunk after getting out of the vehicle.

[0020] C. Before the trigger condition in step S2 is triggered at time t, detect rapid acceleration, rapid deceleration, sharp turns, acceleration, and the vehicle speed detection period Δδ before time t. During the Δδ period, set the vehicle speed threshold to Q1. When the average vehicle speed > Q1 and a rapid deceleration occurs, it is represented as avg(speed). Δδ If the average vehicle speed is greater than Q1 and a sudden deceleration occurs, and within the Δδ time period, a speed threshold of Q2 is set, then when the average vehicle speed is greater than Q2 and a sudden deceleration and sharp turn occur, it is represented as avg(speed). Δδ If the average speed is greater than Q2 and exhibits sudden deceleration and sharp turns, within the Δδ time period, a speed threshold of Q3 is set. When the average speed is greater than Q3 and exhibits sudden deceleration and sharp acceleration, it is represented as avg(speed). Δδ) > Q3, and rapid decrease and rapid increase behaviors occur; within the time period Δδ, set the vehicle speed threshold as Q4 and the acceleration threshold as Q5, if the average vehicle speed > Q4 and acceleration > Q5, expressed as avg(speed Δδ )>Q4, there exists (acceleration>Q5) within the Δδ time period;

[0021] D, before the time t when the trigger condition of step S2 is triggered, detect the tire pressure before time t, within the Δδ time period, set the tire pressure threshold as A1, when the tire pressure < A1 occurs, and for the difference between (maximum tire pressure - minimum tire pressure), set the tire pressure threshold as A2, which satisfies that the tire pressure value of at least one wheel is less than the threshold and greater than 0, expressed as existing in the Δδ time period Meanwhile, within the Δδ time period, the difference of (maximum tire pressure - minimum tire pressure) of at least one wheel > A2, which is expressed as existing in the Δδ time period where i represents 4 tires, i=1 or 2 or 3 or 4;

[0022] E, before the time t when the trigger condition of step S2 is triggered, detect the engine coolant temperature before time t, within the Δδ time period, set the engine coolant temperature threshold as B1, when it is satisfied that the engine coolant temperature exceeds the threshold and continuously exceeds the threshold B1, it is expressed as Coolant_temperature > B1, and each engine coolant temperature in the Δm time interval within the Δδ time period where i represents the number of data entries;

[0023] F, before the time t when the trigger condition of step S2 is triggered, detect the engine state change before time t, within the Δδ time period, M1 starting behaviors occur in the engine state change.

[0024] In the high-speed abnormal parking recognition method based on internet of vehicles big data according to the present invention; in the step S6, when A parking behavior and B parking behavior are satisfied simultaneously, or A parking behavior and one of C parking behavior, D parking behavior, E parking behavior and F parking behavior are satisfied, and the parking behavior corresponding to step S4 is not satisfied, then it is determined that the vehicle has an abnormal parking behavior on the expressway at time t, otherwise no abnormal parking behavior occurs.

[0025] In the high-speed abnormal parking recognition method based on internet of vehicles big data according to the present invention; said step S7 further comprises the following steps:

[0026] S71, record and mark historical communication information, when the number of high-speed abnormal parking vehicles in historical communication reaches ten-thousand level, various abnormal conditions recognized are used as multi-classification labels; collect historical driving information, alarm information, factory return information and GPS information of vehicles.

[0027] In the high-speed abnormal parking identification method based on vehicle network big data described in this invention, step S7 further includes the following steps:

[0028] S72 trains a machine learning classification model based on multi-class labels and collected historical driving information, alarm information, return-to-factory information and GPS information of vehicles. By classifying vehicles that have passed through steps S2 and S3, it distinguishes whether the vehicles are abnormally parked on the highway and the corresponding fault and accident scenarios, and promptly cares for the owners of vehicles that are abnormally parked on the highway.

[0029] The high-speed abnormal parking identification method based on vehicle network big data of the present invention can effectively and quickly monitor whether vehicles have abnormal parking behavior on highways. By combining vehicle signals and driver behavior identification, it does not require aftermarket sensors and has the characteristics of low cost and high accuracy. It lays the foundation for OEMs to realize after-sales care for accidents, ensure the safety of car owners, and provide after-sales services. Attached Figure Description

[0030] Figure 1 This is a flowchart illustrating an embodiment of the high-speed abnormal parking identification method based on vehicle network big data of the present invention;

[0031] Figure 2 This is a time range diagram for identifying abnormal parking at high speeds in the high-speed abnormal parking identification method based on vehicle network big data of the present invention. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0033] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0034] like Figure 1-2 As shown, Figure 1This is a flowchart illustrating an embodiment of the high-speed abnormal parking identification method based on vehicle-to-everything (V2X) big data according to the present invention. A high-speed abnormal parking identification method based on V2X big data is provided, comprising the following steps:

[0035] In step S1, vehicle status data, behavior data and GPS data are collected from the vehicle terminal and transmitted to the cloud. The vehicle status data, behavior data and GPS data are preprocessed to obtain vehicle speed, rapid acceleration, rapid deceleration, sharp turn, tire pressure, engine coolant temperature, engine status change, hazard lights on, whether the vehicle has been exited and GPS information.

[0036] In step S2, a high-speed parking model trigger is constructed based on the vehicle speed and hazard lights being on; and the triggering conditions of the high-speed parking model trigger are used to determine whether the vehicle needs to trigger the highway abnormal parking judgment.

[0037] In step S3, based on GPS point positioning data, a high-speed positioning judgment model is constructed based on the two-point-one-line method, and for vehicles that meet the triggering conditions of the high-speed parking model trigger, it is further determined whether the current vehicle stops due to vehicle abnormality and is not in a service area during its journey on the highway.

[0038] In step S4, a highway traffic jam behavior recognition model is constructed based on the high-speed positioning judgment model. When a vehicle exhibits multiple similar stopping behaviors in its historical driving trajectory data on the highway, without any return-to-work repair data, and the current vehicle speed is less than a threshold a1, it is considered that the vehicle may have stopped due to highway traffic jam behavior, and the owner information for this type of vehicle is marked. The threshold a1 = (a11, a12, a13…a14), and the speed threshold a1 is obtained based on a multi-layered median definition of historical data (layers are based on speed, holidays, weekdays, climate, etc.). The highway traffic jam behavior recognition model excludes stopping due to traffic jam behavior.

[0039] In step S5, a behavior discrimination model is constructed based on the highway traffic jam behavior recognition model to further determine whether vehicles that meet the requirements of step S2 and step S3 have stopped due to abnormal behavior on the highway.

[0040] Using a behavioral discrimination model, vehicles meeting the trigger conditions are further assessed to determine whether they stopped due to abnormal behavior on the highway. Analysis of pre-trigger behaviors, speed changes, acceleration, and water temperature and tire pressure further determines whether the vehicle exited the vehicle or unfastened its seatbelt. This leads to the conclusion that the vehicle experienced an anomaly on the highway.

[0041] In step S6, for vehicles that meet the requirements of step S5 but not step S4, the behavior discrimination model is used to determine when the vehicle experienced abnormal high-speed parking behavior. If abnormal parking behavior occurs, the vehicle information that experienced abnormal high-speed parking behavior is stored in the database.

[0042] In step S7, all vehicle information that has been communicated and confirmed with the vehicle owner for all vehicles that have experienced abnormal high-speed parking behavior is obtained from the database, and a machine learning intelligent classification model is constructed based on all vehicle information that has been communicated and confirmed with the vehicle owner. The machine learning intelligent classification model is used to classify the vehicles that have gone through steps S2, S3 and S4, and to identify the vehicles that have experienced abnormal high-speed parking and the corresponding fault and accident scenarios.

[0043] In this embodiment, in step S2: the triggering conditions of the high-speed parking model trigger include triggering conditions based on the vehicle speed and triggering conditions based on the hazard lights being on. When both the triggering conditions based on the vehicle speed and the triggering conditions based on the hazard lights being on are met simultaneously within a certain time period, the highway abnormal parking judgment is triggered; otherwise, the highway abnormal parking judgment is not triggered.

[0044] In this embodiment, the triggering condition based on vehicle speed is that the vehicle is stationary for a continuous time interval (t0-t(N)), i.e., the vehicle speed is 0 at the current moment. The vehicle speed is expressed as... Where ceil represents rounding up, the... Indicates t i The velocity data is collected at the current moment. 0 indicates that the velocity is 0 at the current moment, and t0 indicates the start time of the model trigger.

[0045] In this embodiment, the triggering condition based on the double flash on status is that the double flash is on within a continuous time interval (t0-t(N)), and the double flash on / off status is represented as follows: Among them, the Indicates t i The data collected shows the status of the hazard lights switch at the time of acquisition. 1 indicates that the hazard lights are on, and t0 indicates the start time of the model trigger.

[0046] In this embodiment, step S3 further includes the following steps:

[0047] In step S31, GPS point positioning data is used to quickly determine whether the vehicle is located on the highway based on the two-point-one-line method; GPS point positioning data corresponding to three points A, B, and C are obtained, where point B is the GPS point positioning data at time t0 in step S2, and point A is the GPS point positioning data corresponding to the N1 minutes before time t0. 0-N1GPS point positioning data at time, point C is t corresponding to N minutes after time t0 0+N GPS point positioning data at time; after obtaining addresses through inverse geocoding from the GPS point positioning data of points A, B and C, addresses on the expressway are obtained through fuzzy matching, and the addresses are used to exclude expressway positions where the vehicle parks in expressway service areas and parking areas. If points A, B and C are all on the expressway at the same time, it is determined that the vehicle meeting the trigger condition is on the expressway.

[0048] In this embodiment, said step S5 includes the following parking behaviors:

[0049] A, before and after time t when the trigger condition in step S2 is triggered, detecting whether the driver has a alighting behavior within an alighting detection period Δγ after time t, that is, whether there is a seatbelt unbuckling signal for the main driver or an opening signal for the left front door of the vehicle; wherein Δγ is determined according to the specific specifications of the vehicle.

[0050] B, due to the special road conditions on the expressway, most accident or fault scenarios require opening the trunk to get the tripod. Therefore, before and after time t when the trigger condition in step S2 is triggered, detecting whether the driver opens the trunk within the alighting detection period Δγ after time t, that is, whether there is a trunk opening behavior after alighting;

[0051] C, before time t when the trigger condition in step S2 is triggered, detecting rapid acceleration, rapid deceleration, sharp turn, acceleration and vehicle speed within a detection period Δδ before time t, setting a vehicle speed threshold Q1 within the period Δδ. When the average vehicle speed is > Q1 and a rapid deceleration behavior occurs, which is represented as avg(speed Δδ ) > Q1, and a rapid deceleration behavior occurs; setting a vehicle speed threshold Q2 within the period Δδ. When the average vehicle speed is > Q2 and rapid deceleration and sharp turning behaviors occur, which is represented as avg(speed Δδ ) > Q2, and rapid deceleration and sharp turning behaviors occur; setting a vehicle speed threshold Q3 within the period Δδ. When the average vehicle speed is > Q3 and rapid deceleration and rapid acceleration behaviors occur, which is represented as avg(speed Δδ ) > Q3, and rapid deceleration and rapid acceleration behaviors occur; setting a vehicle speed threshold Q4 and an acceleration threshold Q5 within the period Δδ, satisfying that the average vehicle speed is > Q4 and acceleration is > Q5, which is represented as avg(speed Δδ ) > Q4, there exists (acceleration > Q5) within the period Δδ;

[0052] D, before time t when the trigger condition in step S2 is triggered, detecting the tire pressure within the period Δδ before time t, setting a tire pressure threshold A1. When tire pressure < A1 occurs, and for the difference between (maximum tire pressure - minimum tire pressure) a tire pressure threshold A2 is set, satisfying that the tire pressure of at least one wheel is less than the threshold and greater than 0, which is represented as existing in the period Δδ Meanwhile, within the time period Δδ, the difference between (maximum tire pressure - minimum tire pressure) of at least one wheel is > A2, which indicates that there exists wherein i represents 4 tires, i = 1 or 2 or 3 or 4;

[0053] E, before the time t when the trigger condition of step S2 is triggered, detecting the engine coolant temperature before time t; within the time period Δδ, setting the engine coolant temperature threshold as B1, when it is satisfied that the engine coolant temperature exceeds the threshold and continuously exceeds the threshold B1, it is expressed as Coolant_temperature > B1, and each engine coolant temperature within the Δm time interval existing in the time period Δδ wherein i represents the number of data pieces;

[0054] F, before the time t when the trigger condition of step S2 is triggered, detecting the engine state change before time t; within the time period Δδ, the engine state change has M1 starting behaviors.

[0055] The parking behavior further includes, before the time t when the trigger condition of step S2 is triggered, detecting fault-related behaviors before time t, such as signal values of abnormal tire pressure behavior, abnormal water temperature behavior, difficult starting behavior and the like, to further judge abnormal parking. Therefore, the detection period Δδ for water temperature, tire pressure and engine start-stop before time t is detected; within the time period Δδ, setting a tire pressure threshold A1, corresponding to that when tire pressure < A1 occurs, and setting a threshold A2 for the difference of (maximum tire pressure - minimum tire pressure), if it is satisfied that the tire pressure value of at least one wheel is less than the threshold and greater than 0, and meanwhile within the time period Δδ, the difference between (maximum tire pressure - minimum tire pressure) of at least one wheel is > A2, it indicates that there exists within the time period Δδ and wherein i represents 4 tires, i = 1 or 2 or 3 or 4.

[0056] within the time period Δδ, setting the engine coolant temperature threshold as B1, and it is satisfied that the engine coolant temperature exceeds the threshold and continuously exceeds the threshold, which is expressed as Coolant_temperature > B1, and each engine coolant temperature within the Δm time interval existing in the time period Δδ wherein i represents the number of data pieces. Within the time period Δδ, the engine state change has M1 starting behaviors.

[0057] In this embodiment, in step S6, when parking behavior A and parking behavior B are satisfied at the same time, or parking behavior A and one of parking behavior C, parking behavior D, parking behavior E and parking behavior F are satisfied, and the parking behavior corresponding to step S4 is not satisfied, then it is determined that the vehicle has an abnormal parking behavior on the expressway at time t; otherwise, no abnormal parking behavior occurs.

[0058] In this embodiment, step S7 further includes the following steps:

[0059] In step S71, historical communication information is recorded and marked. When the number of vehicles with abnormal parking at high speeds in historical communication reaches tens of thousands, multiple abnormal situations will be identified as multi-classification labels. Historical driving information, alarm information, return-to-factory information and GPS information of the vehicles are collected.

[0060] In this embodiment, step S7 further includes the following steps:

[0061] In step S72, a machine learning classification model is trained based on multi-class labels and the collected historical driving information, alarm information, return-to-factory information and GPS information of the vehicle. By classifying the vehicles that have passed through steps S2 and S3, it is determined whether the vehicle is abnormally parked on the highway and the corresponding fault and accident scenarios, and timely care is provided to the owners of vehicles that are abnormally parked on the highway.

[0062] Specifically, abnormal parking at high speeds and the corresponding fault and accident scenarios include abnormal water temperature, tire blowout, accident, difficulty starting, and others.

[0063] First, identify vehicles that may be parked abnormally, such as those with their hazard lights on for more than N minutes. Second, use the driving trajectory and the "two points determine a line" method to quickly determine the vehicle's location on the highway, excluding vehicles parked in highway service areas and parking areas. Third, exclude vehicles involved in highway traffic jams. Finally, the remaining vehicles are those with abnormal parking behavior. Send the information about vehicles with abnormal parking behavior to the dealership, and these vehicles will wait for assistance.

[0064] Specifically, the vehicle's infotainment system transmits vehicle status data, behavior data, and GPS data to the cloud via CAN bus, selecting key data such as vehicle speed, rapid acceleration, rapid deceleration, sharp turns, hazard light status, whether the driver has exited the vehicle, and GPS data. Step S1 also includes data filtering to remove abnormal data where the vehicle speed exceeds a threshold. The threshold is the maximum value that the CAN bus transmits.

[0065] The trigger condition for the high-speed parking model is that it will trigger if both of the following conditions are met simultaneously within a certain time period; otherwise, it will not trigger. Specifically, it must meet the condition that the vehicle is continuously parked and its hazard lights are on during the BC time period. When the data collection interval is n minutes, and the BC segment is a time interval (t0-t(N)) arranged sequentially, the specific trigger condition is: 1) The vehicle remains stationary during the consecutive BC time intervals, expressed as... Where ceil represents rounding up. For example, if the data acquisition rate is set to 0 for 5 consecutive minutes, and the data acquisition interval is 10 seconds (1 / 6) minutes, then... 2) Double flash is on during the continuous BC segment time interval, which is indicated as If the dual flash is set to be on continuously for 5 minutes, the data collection interval is 10 seconds = (1 / 6) minutes, where is... Among them, parameters Indicates t i The speed at which data is collected, where 0 indicates that the speed is 0 at the current moment. Indicates t i The data represents the status of the dual flash on / off state at specific times. 1 indicates dual flash is on. t0 represents the start time of the model trigger, i.e., time B. t(N) represents the end time of the model trigger, i.e., time C.

[0066] Specifically, directly using GPS positioning data at each moment to determine if a vehicle is on a highway results in excessive data, slowing down real-time processing. Furthermore, relying on a single GPS address can lead to GPS positioning inaccuracies. Therefore, a method balancing speed and accuracy is proposed: utilizing GPS point positioning data and a two-point-one-line method to quickly determine if a vehicle is on a highway. This involves applying two two-point-one-line logic steps, such as obtaining GPS point positioning data for points A, B, and C. Point B represents time t0 in step S2, and point A represents the location N1 minutes prior to time t0. 0-N1 At time C, point C corresponds to t N minutes after time t0. 0+N The system uses reverse geocoding to obtain addresses from three GPS points (A, B, and C). Fuzzy matching is then used to obtain addresses along the highway, and these addresses are used to exclude vehicles parked at highway service areas or rest areas. If all three points (A, B, and C) are simultaneously on the highway, the vehicle is considered to meet the triggering condition and is on the highway. For example, if point A represents the Changge toll station on the G4 Beijing-Hong Kong-Macau Expressway in Laocheng Town, Changge City, Xuchang City, Henan Province; point B represents the Xuchang service area on the G4 Beijing-Hong Kong-Macau Expressway in Foerhu Town, Changge City, Xuchang City, Henan Province; and point C represents the Xuchang service area on the G4 Beijing-Hong Kong-Macau Expressway in Foerhu Town, Changge City, Xuchang City, Henan Province, then this triggering vehicle is excluded.

[0067] Furthermore, this invention can be widely applied to highway management and traffic safety, and is of great significance for improving the safety and smooth flow of highways. This invention also has high practicality, and can provide drivers with a more convenient driving experience, enhancing their sense of security and confidence.

[0068] The beneficial effects of the high-speed abnormal parking identification method based on vehicle network big data provided in this embodiment of the invention are at least as follows:

[0069] 1. By using the timing of hazard lights to identify vehicles that may be parked abnormally, the likelihood of misjudgment is reduced.

[0070] 2. By using the two-point-one-line method twice to determine that the vehicle is located on the highway, vehicles parked in other places are excluded, thus improving the accuracy of identification.

[0071] 3. By eliminating factors such as highway traffic jams, the possibility of misjudgment has been reduced.

[0072] 4. By promptly sending notifications to authorized dealerships, we can quickly respond to abnormal parking incidents, ensuring smooth traffic flow and safety on highways.

[0073] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the present invention is not limited to the described order of actions, because according to the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0074] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0075] Therefore, the above description is only a preferred embodiment of the present invention, and the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. The scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for identifying abnormal parking on high-speed roads based on vehicle-to-everything (V2X) big data, characterized in that, The method includes the following steps: S1 collects vehicle status data, behavior data, and GPS data from the vehicle's infotainment system and transmits them to the cloud. It then preprocesses the vehicle status data, behavior data, and GPS data to obtain vehicle speed, rapid acceleration, rapid deceleration, sharp turns, tire pressure, engine coolant temperature, engine status changes, hazard lights status, whether the vehicle has been exited, and GPS information. S2, construct a high-speed parking model trigger based on the vehicle speed and hazard lights being on; and determine whether the vehicle needs to trigger the highway abnormal parking judgment based on the triggering conditions of the high-speed parking model trigger. S3. Based on GPS point positioning data, a high-speed positioning judgment model is constructed based on the two-point-one-line method. For vehicles that meet the triggering conditions of the high-speed parking model trigger, it is further determined whether the current vehicle stops due to vehicle abnormality and is not in a service area while driving on the highway. S4. Based on the high-speed positioning judgment model, a high-speed traffic jam behavior recognition model is constructed. When a vehicle has multiple similar stopping behaviors in its historical driving trajectory data on the highway and there is no data on returning to the factory for maintenance, and the current vehicle speed is less than the threshold a1, it is considered that the vehicle may have stopped due to high-speed traffic jam behavior, and such vehicle owner information is marked. S5. Based on the highway traffic jam behavior recognition model, construct a behavior discrimination model to further determine whether vehicles that meet both step S2 and step S3 have stopped due to an abnormality on the highway. S6. If the conditions of step S5 are met but not step S4 are met, determine the time when the vehicle experienced abnormal high-speed parking behavior based on the behavior discrimination model. If abnormal parking behavior occurs, store the vehicle information that experienced abnormal high-speed parking behavior in the database. S7: Obtain all vehicle information from the database that has been communicated and confirmed with the vehicle owner for all vehicles that have experienced abnormal parking behavior on the highway, and build a machine learning intelligent classification model based on all vehicle information that has been communicated and confirmed with the vehicle owner; classify the vehicles that have gone through steps S2, S3 and S4 through the machine learning intelligent classification model, and identify the vehicles that have experienced abnormal parking on the highway and the corresponding fault and accident scenarios.

2. The method for identifying abnormal parking on high-speed roads based on vehicle network big data according to claim 1, characterized in that, In step S2: the triggering conditions of the high-speed parking model trigger include triggering conditions based on the vehicle speed and triggering conditions based on the hazard lights being on. When both the triggering conditions based on the vehicle speed and the triggering conditions based on the hazard lights being on are met simultaneously within a certain time period, the highway abnormal parking judgment is triggered; otherwise, the highway abnormal parking judgment is not triggered.

3. The method for identifying abnormal parking on high-speed roads based on vehicle network big data according to claim 2, characterized in that, The triggering condition based on vehicle speed is that the vehicle is stationary for a continuous time interval (t0-t(N)), i.e., the vehicle speed is 0 at the current moment. The vehicle speed is expressed as... Where ceil represents rounding up, the... Indicates t i The velocity data is collected at the current moment. 0 indicates that the velocity is 0 at the current moment, and t0 indicates the start time of the model trigger.

4. The method for identifying abnormal parking on high speeds based on vehicle network big data according to claim 3, characterized in that, The triggering condition based on the double flash on status is that the double flash is on within a continuous time interval (t0-t(N)), and the double flash on / off status is expressed as follows: Among them, the Indicates t i The data collected shows the status of the hazard lights switch at the time of acquisition. 1 indicates that the hazard lights are on, and t0 indicates the start time of the model trigger.

5. The method for identifying abnormal parking on high speeds based on vehicle network big data according to claim 4, characterized in that, Step S3 further includes the following steps: S31, using GPS point positioning data, quickly determine whether the vehicle is located on the highway based on the two-point-one-line method; obtain the GPS point positioning data corresponding to three points A, B, and C, where point B is the GPS point positioning data at time t0 in step S2, and point A is the GPS point positioning data corresponding to the N1 minutes before time t0. 0-N1 GPS location data at time t0, where point C is the location of point N minutes after time t0. 0+N The GPS location data at any given time; the GPS location data of points A, B, and C are used to obtain addresses through reverse geocoding, and then fuzzy matching is used to obtain addresses on the highway. The addresses are used to exclude highway locations that are parked in highway service areas and parking areas. If points A, B, and C are all on the highway at the same time, the vehicle is considered to meet the triggering conditions and is on the highway.

6. The method for identifying abnormal parking on high-speed roads based on vehicle-to-everything (V2X) big data according to claim 5, characterized in that, Step S5 includes the following parking actions: A. Before and after the triggering condition in step S2 is triggered at time t, detect the vehicle exit detection period Δγ after time t to see if the driver has exited the vehicle, i.e., whether the driver has unbuckled the seat belt or the left front door of the vehicle has opened. B, before and after the triggering condition of step S2 is triggered at time t, detect whether the driver opened the trunk during the vehicle exit detection period after time t, that is, whether there was an act of opening the trunk after getting out of the vehicle. C. Before the trigger condition in step S2 is triggered at time t, detect rapid acceleration, rapid deceleration, sharp turns, acceleration, and the vehicle speed detection period Δδ before time t. During the Δδ period, set the vehicle speed threshold to Q1. When the average vehicle speed > Q1 and a rapid deceleration occurs, it is represented as avg(speed). Δδ Q1 > 1, and a sharp decrease in behavior occurs; Within the Δδ time period, a vehicle speed threshold of Q2 is set. When the average vehicle speed > Q2 and sudden deceleration or sharp turn occurs, it is represented as avg(speed). Δδ The value is greater than Q2, and there are signs of a sharp decrease and abrupt reversal. Within the Δδ time period, a vehicle speed threshold of Q3 is set. When the average vehicle speed > Q3 and sudden deceleration or acceleration occurs, it is represented as avg(speed). Δδ If the average speed is greater than Q3, and there are sudden deceleration and acceleration behaviors; within the Δδ time period, set the vehicle speed threshold to Q4 and the acceleration threshold to Q5, satisfying the condition that the average vehicle speed is greater than Q4 and the acceleration is greater than Q5, denoted as avg(speed). Δδ Q4 > Δδ indicates that (acceleration > 5) exists within the time interval. In D, before the time t when the triggering condition of said step S2 is triggered, detect the tire pressure before time t. Within the Δδ time period, set the tire pressure threshold as A1. When tire pressure < A1 occurs, and set a tire pressure threshold A2 for the difference between the maximum tire pressure and the minimum tire pressure, if at least one wheel satisfies that the tire pressure value is less than the threshold and greater than 0, it indicates that the Δδ time period exists Meanwhile, within the Δδ time period, if the difference between the maximum tire pressure and the minimum tire pressure of at least one wheel is > A2, it indicates that the Δδ time period exists Where i represents 4 tires, i=1 or 2 or 3 or 4; E, before the trigger condition in step S2 is triggered at time t, the engine coolant temperature before time t is detected. During the Δδ time period, the engine coolant temperature threshold is set to B1. When the engine coolant temperature exceeds the threshold and continuously exceeds the threshold B1, it is represented as Coolant_temperature>B1, and each engine coolant temperature within the Δδ time period exists in the Δm time interval. Where i represents the number of data entries; F, before the trigger condition in step S2 is triggered at time t, detect the engine state change before time t. During the Δδ time period, the engine state change occurs M1 times.

7. The method for identifying abnormal parking on high-speed roads based on vehicle-to-everything (V2X) big data according to claim 6, characterized in that, In step S6, if the vehicle simultaneously satisfies parking behavior A and parking behavior B, or satisfies parking behavior A and one of parking behaviors C, D, E, or F, but does not satisfy the parking behavior corresponding to step S4, then it is determined that the vehicle has committed an abnormal parking behavior on the highway at time t; otherwise, no abnormal parking behavior has occurred.

8. The method for identifying abnormal parking on high-speed roads based on vehicle network big data according to claim 7, characterized in that, Step S7 further includes the following steps: S71 records and marks historical communication information. When the number of abnormally stopped vehicles on highways in historical communication reaches tens of thousands, it will identify multiple abnormal situations as multi-classification labels; it collects the vehicle's historical driving information, alarm information, return-to-factory information and GPS information.

9. The method for identifying abnormal parking on high-speed roads based on vehicle-to-everything (V2X) big data according to claim 8, characterized in that, Step S7 further includes the following steps: S72 trains a machine learning classification model based on multi-class labels and collected historical driving information, alarm information, return-to-factory information and GPS information of vehicles. By classifying vehicles that have passed through steps S2 and S3, it distinguishes whether the vehicles are abnormally parked on the highway and the corresponding fault and accident scenarios, and promptly cares for the owners of vehicles that are abnormally parked on the highway.

Images