Automobile driving scene recognition control system, method and storage medium

Abstract

The application discloses a vehicle driving scene recognition control system and method and a storage medium. The system and method calculate a near-field model of a current vehicle through a vehicle-mounted map, positioning, radar, a camera and vehicle communication information, recognize a current driving scene of the vehicle, perform relevant safety calculation, and thus issue corresponding warning prompts and vehicle intervention. The application constructs a vehicle near-field driving model through vehicle near-field communication technology, so as to achieve the purposes of predicting the driving behavior of a neighboring vehicle, suggesting the behavior of the vehicle, and warning and intervening, thereby reducing the number of traffic accidents in a highway scene.

Description

Technical Field

[0001] This application pertains to a method for recognizing and controlling driving scenes in a highway driving environment within the field of automotive assisted driving, specifically involving a vehicle driving scene recognition control system, method, and storage medium. Background Technology

[0002] With the rapid development of my country's economy, the road traffic safety situation has become increasingly severe. How to effectively reduce the number of road traffic accidents has become a problem that city managers and major automakers must face. Therefore, it is necessary to propose a vehicle driving scene recognition control system and method from the perspective of highway driving safety, which is the most dangerous part of traffic accidents. This system uses vehicle near-field communication technology to sense the driving information of other vehicles and provides suggestions, warnings, and interventions for drivers' behaviors that may lead to traffic accidents, thereby reducing the number of traffic accidents. Summary of the Invention

[0003] To address the problems existing in the prior art, this application discloses a vehicle driving scene recognition control system and method. By constructing a vehicle near-field driving model through vehicle near-field communication technology, it aims to predict the driving behavior of adjacent vehicles, provide suggestions, warnings and interventions for its own vehicle's behavior, thereby reducing the number of traffic accidents in highway scenarios.

[0004] This application also discloses a storage medium that runs the above-described method and program to construct a vehicle near-field driving model through vehicle near-field communication technology, so as to achieve the purpose of predicting the driving behavior of adjacent vehicles, providing suggestions, warnings and interventions on the behavior of its own vehicle, thereby reducing the number of traffic accidents in highway scenarios.

[0005] The vehicle driving scene recognition and control system disclosed in this application includes a distance calculation module, an image acquisition and processing module, an on-board positioning module, a vehicle communication module, a near-field module, and an execution module; the near-field module includes a scene recognition module and a behavior prediction module.

[0006] The near-field module collects lane information in real time during driving through the image acquisition and processing module. When other vehicles enter the near-field range, it takes pictures of the environment around the current vehicle, identifies the feature information of the surrounding vehicles, and generates a simplified two-dimensional near-field model of the vehicle through image fusion calculation.

[0007] The near-field module uses lane signals obtained from the vehicle positioning module and vehicle radar signals from the distance calculation module to obtain precise distance relationships between vehicles and determine the vehicle near-field model; then, based on lane information and surrounding vehicle speed information, it calculates the driving trajectories of surrounding vehicles within the model, and then dynamically corrects the near-field model.

[0008] Furthermore, the scene recognition module obtains driving scene information through the vehicle map system of the positioning module, determines whether to enter a following scenario by calculating the same lane following scenario, determines whether to enter a lane changing and overtaking scenario by steering operation, and determines whether to enter a constant speed driving scenario by vehicle speed and distance from other vehicles in the same lane.

[0009] Furthermore, the behavior prediction module predicts the possible acceleration, deceleration, and lane-changing behaviors of surrounding vehicles based on their driving trajectory data, and provides corresponding risk warnings to the current vehicle.

[0010] Furthermore, it also includes a lane merging calculation module and / or a rear-end collision module; the lane merging calculation module is triggered by lane merging scenarios, overpass traffic scenarios, ramp driving scenarios, and service area entry / exit scenarios. After activation, it calculates the safety of the vehicle in front of the merging vehicle and the vehicle behind the merging vehicle based on the acceleration of the merging vehicle, and provides corresponding risk warnings; the rear-end collision calculation module triggers rear-end collision calculation and provides corresponding risk warnings when the current vehicle or the vehicle behind is in a following scenario and the vehicle has not entered ACC adaptive cruise control.

[0011] Furthermore, the execution module includes the vehicle audio-visual and lighting control module, braking system, ACC adaptive cruise control module, and powertrain system;

[0012] If the driver fails to follow the risk warnings and forces a lane change during lane change calculations, the powertrain will make minor torque increases or decreases; or in the case of following a vehicle, if it is safe ahead before a rear-end collision occurs, the powertrain will make minor torque increases or decreases for the current vehicle.

[0013] In scenarios such as forced merging by the vehicle in front, following another vehicle, or slow-moving traffic in tunnels and curves ahead, the braking system performs minor braking corrections to the vehicle before a collision occurs, with the powertrain's output torque gradually dropping to zero upon braking. Alternatively, if emergency braking is triggered, the braking system applies the brakes to the vehicle.

[0014] The vehicle infotainment and lighting control module combines audio, lighting, and alert sounds to promptly broadcast the risks detected by the vehicle within the cabin.

[0015] The ACC adaptive cruise control module allows the driver to choose whether to activate the ACC adaptive cruise control function in both following and constant speed driving scenarios.

[0016] This application also discloses a method for recognizing and controlling vehicle driving scenes, including the following steps.

[0017] A simplified two-dimensional near-field model of the vehicle is generated. The vehicle collects lane information in real time during the driving process through the image acquisition and processing module. When other vehicles enter the near-field range, the vehicle takes pictures of the surrounding environment, identifies the feature information of the surrounding vehicles, and generates a simplified two-dimensional near-field model of the vehicle through image fusion calculation.

[0018] The near-field model is generated and corrected. The relative positions between vehicles are calculated by image processing algorithms after the camera is taken, and the precise distance relationship between vehicles is obtained by using lane signals and vehicle radar signals. After the vehicle near-field model is generated, the driving trajectory of the surrounding vehicles in the model is calculated based on lane information and vehicle speed information, and then the near-field model is dynamically corrected.

[0019] Scene recognition: The scene recognition module identifies the current scene, including driving scene information, following other vehicles, lane changing and overtaking, and constant speed driving;

[0020] Behavior prediction: The behavior prediction module performs behavior prediction calculations, predicts the possible acceleration, deceleration, and lane changing behaviors of surrounding vehicles based on their driving trajectories, and provides corresponding risk warnings to the current vehicle.

[0021] Lane merging calculation is triggered by lane merging scenarios, overpass traffic scenarios, ramp driving scenarios, and service area entry and exit scenarios. Once activated, it calculates the safety of vehicles in front of the merging vehicle and vehicles behind the merging vehicle based on the acceleration of the merging vehicle, and provides corresponding risk warnings.

[0022] Rear-end collision calculation is triggered if the current vehicle or the vehicle behind it is in a following scenario and the vehicle has not entered ACC adaptive cruise control, and a corresponding risk warning is given.

[0023] Furthermore, in the lane change calculation, if the driver does not follow the risk warning and forces a lane change, the power system of the execution module will make a small-scale torque increase or decrease correction; or in the case of following a vehicle, if it is safe ahead before a rear-end collision occurs, it will make a small-scale torque increase correction for the current vehicle.

[0024] In scenarios such as forced merging by the vehicle in front, following another vehicle, or slow-moving traffic in tunnels and curves ahead, the braking system of the execution module performs minor braking corrections to the vehicle before a collision occurs, with the powertrain output torque gradually dropping to zero upon braking. Alternatively, if emergency braking is triggered, the braking system applies the brakes to the vehicle.

[0025] The vehicle infotainment and lighting control module of the execution module combines audio, lighting, and alert sounds to promptly broadcast the risks detected by the vehicle in the cabin.

[0026] The ACC adaptive cruise control module of the execution module allows the driver to choose whether to activate the ACC adaptive cruise control function in both following and constant speed driving scenarios.

[0027] Furthermore, before generating a simplified two-dimensional near-field model of a vehicle, the process also includes determining the vehicle IDs within the near-field range. Specifically, this involves assigning vehicle IDs to vehicles entering the near-field range, or interacting with the near-field models of surrounding vehicles, until all vehicles within the near-field range are assigned vehicle IDs. Vehicle IDs with different driving roads or driving directions are masked, and for vehicle IDs that are not masked, vehicle IDs within a set distance from the vehicle are included in the near-field calculation range.

[0028] Furthermore, in generating and correcting the near-field model, the relative positions between vehicles are calculated through image processing algorithms after the camera is taken, and the precise distance relationship between vehicles is obtained using lane signals and onboard radar signals. For the positions of vehicles that are outside the radar test range or are obscured, this vehicle refers to the near-field model of the middle vehicle to confirm the relevant vehicle position data.

[0029] Within a certain period after the near-field model is generated and corrected, the vehicle's near-field module will periodically update the radar signal from the distance calculation module, the signal from the graphics acquisition and processing module, the near-field model obtained from the vehicle communication module, and the lane signal obtained from the vehicle positioning module to accurately correct the current vehicle's near-field model.

[0030] This application also discloses a computer-readable storage medium storing a computer program thereon, which is executed by a processor to implement the above-described vehicle driving scene recognition and control method.

[0031] The beneficial technical effects of this application are as follows:

[0032] This application calculates the near-field model of the current vehicle using in-vehicle maps, positioning, radar, cameras, and vehicle communication information. It identifies the current driving scenario, performs relevant safety calculations, and then issues corresponding warnings and intervenes in the vehicle's operation. By employing the above technologies and utilizing near-field communication technology to construct a near-field driving model, this application aims to predict the driving behavior of adjacent vehicles, provide suggestions, warnings, and interventions for its own vehicle's behavior, thereby reducing the number of traffic accidents on highways. Attached Figure Description

[0033] Fig. 1 This is a schematic diagram of the automotive sensor installation in this application;

[0034] Fig. 2 This is a block diagram of the vehicle driving scene recognition and control system of this application;

[0035] Fig. 3 This is a schematic diagram of the near-field model of the vehicle during high-speed driving. Detailed Implementation

[0036] Embodiments of this application will now be described in more detail with reference to the accompanying drawings. While some embodiments of this application are shown in the drawings, it should be understood that this application can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this application. It should be understood that the drawings and embodiments of this application are for illustrative purposes only and are not intended to limit the scope of protection of this application.

[0037] It should be understood that the steps described in the method embodiments of this application may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of this application is not limited in this respect.

[0038] The names of the messages or information exchanged between multiple devices in the embodiments of this application are for illustrative purposes only and are not intended to limit the scope of these messages or information.

[0039] like Figs. 1 to 3 As shown, in one embodiment of this application, the vehicle driving scene recognition control system includes a distance calculation module 3, an image acquisition and processing module 2, an on-board positioning module 10, a vehicle communication module 5, a near-field module 4, and an execution module; the near-field module includes a scene recognition module 41 and a behavior prediction module 42.

[0040] The vehicle receives communication signals from surrounding vehicles in real time while in motion and processes signals from vehicles with different IDs. Signals from vehicles with IDs traveling on different roads or in different directions are directly blocked. For IDs that are not blocked, distance calculations are performed to determine if the vehicle with that ID has entered the near-field range. Real-time near-field calculations are only performed on vehicles within a distance of less than 100 meters. During the calculation process, vehicle IDs in the image are identified in two ways: Scenario 1: Vehicle entry recognition, i.e., identifying vehicles that have entered the near-field range. In this case, IDs are directly assigned to vehicles entering the near-field range. Scenario 2: Near-field interactive recognition. Due to the limitations of special scenarios such as low-speed driving with small distances between vehicles and abnormal road conditions like service areas and toll stations, near-field model calculations are stopped at low speeds. Near-field model calculations will restart as the vehicle speed increases. If there are many surrounding vehicles, the model calculation will continuously exchange and compare with simplified near-field models of surrounding vehicles until all IDs within the near-field range are correctly identified.

[0041] The near-field module collects lane information in real time during driving through the image acquisition and processing module. When other vehicles enter the near-field range, it takes pictures of the environment around the current vehicle, identifies the feature information of the surrounding vehicles, and generates a simplified two-dimensional near-field model of the vehicle through image fusion calculation.

[0042] The near-field module uses lane signals obtained from the vehicle positioning module and vehicle radar signals from the distance calculation module to obtain precise distance relationships between vehicles, thus determining the vehicle near-field model. Regarding distance calculation, considering the accuracy deviation of the vehicle positioning system, the precise vehicle distance information in the near-field model is calculated using the vehicle radar system. For example... Fig. 3 As shown, taking the calculation of the distance between the current vehicle and the first vehicle as an example, the distance AB between the two vehicles can be obtained through the vehicle-mounted radar. Using the lane width AC, the distance BC between the current vehicle and the first vehicle can be obtained through the Pythagorean theorem. Similarly, the distance between the current vehicle and surrounding vehicles can be calculated.

[0043] Secondly, for confirming the location of vehicles that are outside the radar test range or are obscured, the current vehicle will refer to the near-field model of the intermediate vehicle to confirm the relevant location data.

[0044] Then, based on lane information and the speed information of surrounding vehicles, the driving trajectories of surrounding vehicles within the model are calculated, and the near-field model is dynamically corrected.

[0045] Meanwhile, at regular intervals, the near-field model of the current vehicle is precisely corrected using near-field models and lane signals obtained from radar signals, cameras, and vehicle communication. Vehicle communication involves the vehicle broadcasting signals to other vehicles within its near-field range. These broadcast signals include information such as the road, vehicle ID, location, direction of travel, lane, speed, accelerator, brake, vehicle dimensions, turn signals, steering angle, driving scenario, and near-field model.

[0046] The scene recognition module acquires driving scene information through the vehicle's onboard map system in the positioning module. It determines whether to enter a following scenario by calculating the same-lane following situation, whether to enter a lane-changing / overtaking scenario by steering input, and whether to enter a constant-speed driving scenario by vehicle speed and distance from other vehicles in the same lane. Driving scene information includes overpass driving, traffic jams, tunnel driving, curve driving, bridge driving, ramp driving, and entering / exiting toll stations and service areas. The following scenario is triggered when following another vehicle in the same lane for 30 seconds at a distance of 1 to 1.5 times the safe distance, or when the following distance is less than 1 times the safe distance. The lane-changing / overtaking scenario is triggered by turn signal and steering angle signals. Turning on the turn signal indicates a lane-changing plan, activating the lane-changing scenario; changes in steering angle and steering rate indicate the start of the lane change, activating the lane-changing scenario. If there are no other vehicles within a safe distance of twice the vehicle speed in the current lane, and the range of speed variation is small, then it is judged as a constant speed driving scenario.

[0047] The behavior prediction module predicts the possible acceleration, deceleration, and lane-changing behaviors of surrounding vehicles based on their driving trajectory data, and provides corresponding risk warnings to the current vehicle. Specifically, it can be divided into two scenarios: Scenario 1: If a vehicle in an adjacent lane is following closely and the distance between the vehicle in front of it and the current vehicle is greater than 10 meters, it is determined that the vehicle may change lanes; Scenario 2: If vehicles in the current lane are following closely in front and behind, and there is no lane-changing behavior, the current vehicle is warned of the risk of rear-end collision and needs to increase the safe distance from the vehicle in front.

[0048] The vehicle driving scene recognition and control system also includes a lane merging calculation module 43 and / or a rear-end collision calculation module 44. The lane merging calculation module is triggered by lane merging scenarios, overpass traffic scenarios, ramp driving scenarios, and service area entry / exit scenarios. After activation, it calculates the safety of the vehicle in front of the merging vehicle and the vehicle behind the merging vehicle based on the acceleration of the merging vehicle, and provides corresponding risk warnings. The rear-end collision calculation module triggers rear-end collision calculation and provides corresponding risk warnings when the current vehicle or the vehicle behind is in a following scenario and the vehicle has not entered ACC adaptive cruise control.

[0049] The execution modules include the vehicle audio-visual and lighting control module 6, the braking system 7, the ACC adaptive cruise control module 8, and the power system 9;

[0050] If the driver fails to follow the risk warnings and forces a lane change during lane change calculations, the powertrain will make minor torque increases or decreases; or in the case of following a vehicle, if it is safe ahead before a rear-end collision occurs, the powertrain will make minor torque increases or decreases for the current vehicle.

[0051] In scenarios such as forced merging by the vehicle in front, following another vehicle, or slow-moving traffic in tunnels and curves ahead, the braking system performs minor braking corrections to the vehicle before a collision occurs, with the powertrain's output torque gradually dropping to zero upon braking. Alternatively, if emergency braking is triggered, the braking system applies the brakes to the vehicle.

[0052] The vehicle's infotainment and lighting control module combines audio, video, lighting, and alerts to promptly announce risks detected by the vehicle within the cabin. If a safety risk is detected in a particular direction, a light and sound alert will be issued for that direction. For general risks, a slow-flashing yellow light and a relatively low volume will be used; for high-risk risks, a fast-flashing red light and a louder sound will be used, along with a voice prompt to prepare for a collision. Different scenarios will trigger different alerts. For example, in overpass driving scenarios, a voice alert will be issued regarding the risk of merging or exiting; in following other vehicles, an alert will be issued regarding the risk of rear-end collisions or the activation of ACC control; in constant speed driving scenarios, an alert will be issued regarding the activation of ACC control; in lane-changing and overtaking scenarios, an alert will be issued regarding the risk of collisions; in slow-moving traffic scenarios entering or exiting tunnels, an alert will be issued regarding the risk of rear-end collisions; in slow-moving traffic scenarios on curves, an alert will be issued regarding the risk of speeding and rear-end collisions; in ramp driving scenarios, an alert will be issued regarding the risk of collisions when merging; and in service area entry / exit scenarios, an alert will be issued regarding the risk of collisions when merging.

[0053] The ACC adaptive cruise control module allows the driver to choose whether to activate the ACC adaptive cruise control function in both following and constant speed driving scenarios.

[0054] One embodiment of this application discloses a vehicle driving scene recognition and control method, including the following steps.

[0055] Determine the vehicle IDs within the near field range. Specifically, assign a vehicle ID to vehicles entering the near field range, or interact with the near field models of surrounding vehicles until all vehicles within the near field range are assigned a vehicle ID. Mask vehicle IDs that are on different roads or in different directions of travel. For vehicle IDs that are not masked, include vehicle IDs within a set distance from the vehicle in the near field calculation range.

[0056] A simplified two-dimensional near-field model of the vehicle is generated. The vehicle collects lane information in real time during the driving process through the image acquisition and processing module. When other vehicles enter the near-field range, the vehicle takes pictures of the surrounding environment, identifies the feature information of the surrounding vehicles, and generates a simplified two-dimensional near-field model of the vehicle through image fusion calculation.

[0057] The near-field model is generated and corrected by calculating the relative positions of vehicles using image processing algorithms after camera capture, and obtaining precise distance relationships between vehicles using lane signals and onboard radar signals. After the near-field model is generated, the driving trajectories of surrounding vehicles within the model are calculated based on lane information and vehicle speed information, thereby dynamically correcting the near-field model. In the generation and correction of the near-field model, the relative positions of vehicles are calculated using image processing algorithms after camera capture, and precise distance relationships between vehicles are obtained using lane signals and onboard radar signals. For vehicles outside the radar test range or whose positions are obstructed, the vehicle references the near-field model of an intermediate vehicle to confirm the relevant vehicle position data. Within a certain period after the near-field model is generated and corrected, the vehicle's near-field module periodically updates the radar signals from the distance calculation module, the signals from the image acquisition and processing module, the near-field model obtained from the vehicle communication module, and the lane signals obtained from the onboard positioning module to precisely correct the current vehicle's near-field model.

[0058] Scene recognition: The scene recognition module identifies the current scene, including driving scene information, following other vehicles, lane changing and overtaking, and constant speed driving;

[0059] Behavior prediction: The behavior prediction module performs behavior prediction calculations, predicts the possible acceleration, deceleration, and lane changing behaviors of surrounding vehicles based on their driving trajectories, and provides corresponding risk warnings to the current vehicle.

[0060] Lane merging calculation is triggered by lane merging scenarios, overpass traffic scenarios, ramp driving scenarios, and service area entry and exit scenarios. Once activated, it calculates the safety of vehicles in front of the merging vehicle and vehicles behind the merging vehicle based on the acceleration of the merging vehicle, and provides corresponding risk warnings.

[0061] Rear-end collision calculation is triggered if the current vehicle or the vehicle behind it is in a following scenario and the vehicle has not entered ACC adaptive cruise control, and a corresponding risk warning is given.

[0062] It also includes the following steps,

[0063] If the driver fails to follow the risk warning and forces a lane change during lane change calculations, the power system of the execution module will make minor torque increase or decrease corrections; or in the case of following a vehicle, if it is safe ahead before a rear-end collision occurs, it will make minor torque increase corrections to the current vehicle.

[0064] In scenarios such as forced merging by the vehicle in front, following another vehicle, or slow-moving traffic in tunnels and curves ahead, the braking system of the execution module performs minor braking corrections to the vehicle before a collision occurs, with the powertrain output torque gradually dropping to zero upon braking. Alternatively, if emergency braking is triggered, the braking system applies the brakes to the vehicle.

[0065] The vehicle infotainment and lighting control module of the execution module combines audio, lighting, and alert sounds to promptly broadcast the risks detected by the vehicle in the cabin.

[0066] The ACC adaptive cruise control module of the execution module allows the driver to choose whether to activate the ACC adaptive cruise control function in both following and constant speed driving scenarios.

[0067] Another embodiment of this application discloses a computer-readable storage medium storing a computer program thereon, which is executed by a processor to implement the above-described vehicle driving scene recognition and control method.

[0068] The vehicle motion model is established as follows:

[0069] 1) Due to safety considerations for high-speed driving, the actual lateral speed during lane merging is relatively small, therefore it is not considered in the lateral speed and acceleration model.

[0070] 2) The speed of vehicles merging into the lane is faster than the speed of vehicles exiting the lane.

[0071] 3) The vehicle's acceleration remains unchanged during acceleration.

[0072] 4) The condition for no collision with a vehicle merging into the lane is that when the two vehicles accelerate to the same speed, the distance between them is greater than zero meters.

[0073] 5) The condition for no collision with the vehicle in front in the merging lane is that the distance between the two vehicles is greater than zero meters 1 second after merging.

[0074] The collision equations with vehicles merging into the lane are:

[0075] V0+a0t1=V1+a1t1

[0076]

[0077] Therefore, Where V1 > V0.

[0078] The collision equation with the vehicle in front of the merging lane is:

[0079]

[0080] When the parallel output time t2 is 1s, the formula simplifies to:

[0081]

[0082] Therefore, a0-a2<2(L2+V2-V0).

[0083] L1 is Fig. 3 The distance between the first vehicle and the current vehicle in the driving direction;

[0084] L2 is Fig. 3 In the middle, the distance between the second vehicle and the current vehicle in the direction of travel;

[0085] V0 is the current vehicle speed, V1 is the speed of the first vehicle, and V2 is the speed of the second vehicle.

[0086] a0 is the current vehicle's acceleration, a1 is the first vehicle's acceleration, and a2 is the second vehicle's acceleration.

[0087] t1 is the time when the current vehicle merges in, and t2 is the time when the current vehicle leaves the grid.

[0088] Rear-end collision calculation is triggered if the current vehicle or the vehicle behind it is in a following scenario and the vehicle has not entered ACC adaptive cruise control, and a corresponding risk warning is given.

[0089] The formula for calculating rear-end collisions is the same as the equation for collisions between vehicles in front of the vehicle merging out of the lane. At this time, t2 starts timing from the moment the vehicle enters the following scenario.

Claims

1. A vehicle driving scene recognition and control system, characterized in that: It includes a distance calculation module (3), an image acquisition and processing module (2), an on-board positioning module (10), a vehicle communication module (5), a near-field module (4), and an execution module; the near-field module includes a scene recognition module (41) and a behavior prediction module (42). The near-field module collects lane information in real time during driving through the image acquisition and processing module. When other vehicles enter the near-field range, it takes pictures of the environment around the current vehicle, identifies the feature information of the surrounding vehicles, assigns vehicle IDs to the vehicles entering the near-field range, or interacts with the near-field models of surrounding vehicles until all vehicles in the near-field range are assigned vehicle IDs. It masks vehicle IDs that are on different roads or in different directions of travel. For vehicle IDs that are not masked, it includes vehicle IDs within a set distance from the current vehicle in the near-field calculation range and generates a simplified two-dimensional near-field model of the vehicle through image fusion calculation. The near-field module uses lane signals obtained from the vehicle positioning module and vehicle radar signals from the distance calculation module to obtain precise distance relationships between vehicles and determine the vehicle near-field model. For the position of vehicles that are outside the radar test range or are obscured, the vehicle refers to the near-field model of the middle vehicle to confirm the relevant vehicle position data; then, based on lane information and the speed information of surrounding vehicles, it calculates the driving trajectory of surrounding vehicles in the model and dynamically corrects the near-field model. The scene recognition module obtains driving scene information through the vehicle map system of the positioning module. It determines whether to enter a following scenario by calculating the same lane following scenario, whether to enter a lane changing and overtaking scenario by steering operation, and whether to enter a constant speed driving scenario by vehicle speed and distance from other vehicles in the same lane.

2. The vehicle driving scene recognition and control system as described in claim 1, characterized in that: The behavior prediction module predicts the possible acceleration, deceleration, and lane changing behaviors of surrounding vehicles based on their driving trajectory data, and provides corresponding risk warnings to the current vehicle.

3. The vehicle driving scene recognition and control system as described in claim 1, characterized in that: It also includes a lane merging calculation module (43) and / or a rear-end collision calculation module (44); the lane merging calculation module is controlled and triggered by lane merging scenarios, overpass traffic scenarios, ramp driving scenarios, and service area entry and exit scenarios. After activation, it calculates the safety of vehicles in front of the merging lane and vehicles behind the merging lane based on the acceleration of the merging vehicle, and gives corresponding risk warnings. The rear-end collision calculation module triggers rear-end collision calculation and provides corresponding risk warnings when the current vehicle or the vehicle behind is in a following scenario and the vehicle has not entered ACC adaptive cruise control.

4. The vehicle driving scene recognition and control system as described in claim 1, characterized in that: The execution module includes the vehicle audio and lighting control module (6), the braking system (7), the ACC adaptive cruise control module (8), and the power system (9); If the driver fails to follow the risk warnings and forces a lane change during lane change calculations, the powertrain will make minor torque increases or decreases; or in the case of following a vehicle, if it is safe ahead before a rear-end collision occurs, the powertrain will make minor torque increases or decreases for the current vehicle. In scenarios such as forced merging by the vehicle in front, following another vehicle, and slow-moving traffic in tunnels and curves ahead, the braking system performs minor braking corrections to the vehicle before a collision occurs. When braking is triggered, the power system output torque slowly drops to zero. Alternatively, if emergency braking is triggered, the braking system will apply the brakes to the vehicle. The vehicle infotainment and lighting control module combines audio, lighting, and alert sounds to promptly broadcast the risks detected by the vehicle within the cabin. The ACC adaptive cruise control module allows the driver to choose whether to activate the ACC adaptive cruise control function in both following and constant speed driving scenarios.

5. A method for recognizing and controlling vehicle driving scenes, characterized in that: Includes the following steps, A simplified two-dimensional near-field model of the vehicle is generated. The vehicle collects lane information in real time during driving through the image acquisition and processing module. When other vehicles enter the near-field range, the vehicle's surrounding environment is photographed to identify the feature information of the surrounding vehicles. Vehicle IDs are assigned to the vehicles entering the near-field range, or the vehicle interacts with the near-field models of surrounding vehicles until all vehicles in the near-field range are assigned vehicle IDs. Vehicle IDs with different driving roads and driving directions are masked. For vehicle IDs that are not masked, vehicle IDs within a set distance of the vehicle are included in the near-field calculation range. A simplified two-dimensional near-field model of the vehicle is generated through image fusion calculation. The near-field model is generated and corrected. The relative positions between vehicles are calculated through image processing algorithms after the camera is taken. The precise distance relationship between vehicles is obtained using lane signals and onboard radar signals. For the positions of vehicles that are outside the radar test range or are obscured, the vehicle refers to the near-field model of the middle vehicle to confirm the position data of the relevant vehicles. After the vehicle near-field model is generated, the driving trajectory of the surrounding vehicles in the model is calculated based on lane information and vehicle speed information, and then the near-field model is dynamically corrected. Scene recognition: The scene recognition module identifies the current scene, including driving scene information, following other vehicles, lane changing and overtaking, and constant speed driving; Behavior prediction: The behavior prediction module performs behavior prediction calculations, predicts the possible behaviors of acceleration, deceleration, and lane changing of surrounding vehicles based on their driving trajectories, and provides corresponding risk warnings to the current vehicle. Lane merging calculation is triggered by lane merging scenarios, overpass traffic scenarios, ramp driving scenarios, and service area entry and exit scenarios. Once activated, it calculates the safety of vehicles in front of the merging vehicle and vehicles behind the merging vehicle based on the acceleration of the merging vehicle, and provides corresponding risk warnings. Rear-end collision calculation is triggered if the current vehicle or the vehicle behind it is in a following scenario and the vehicle has not entered ACC adaptive cruise control, and a corresponding risk warning is given.

6. The vehicle driving scene recognition and control method as described in claim 5, characterized in that: It also includes the following steps; If the driver fails to follow the risk warning and forces a lane change during lane change calculations, the power system of the execution module will make minor torque increase or decrease corrections; or in the case of following a vehicle, if it is safe ahead before a rear-end collision occurs, it will make minor torque increase corrections to the current vehicle. In scenarios such as forced merging by the vehicle in front, following another vehicle, and slow-moving traffic in tunnels and curves ahead, the braking system of the execution module performs small-scale braking corrections on the vehicle before a collision occurs. When braking is triggered, the torque output of the power system slowly drops to zero. Alternatively, if emergency braking is triggered, the braking system will apply the brakes to the vehicle. The vehicle infotainment and lighting control module of the execution module combines audio, lighting, and alert sounds to promptly broadcast the risks detected by the vehicle in the cabin. The ACC adaptive cruise control module of the execution module allows the driver to choose whether to activate the ACC adaptive cruise control function in both following and constant speed driving scenarios.

7. The vehicle driving scene recognition and control method as described in claim 5, characterized in that: Within a certain period after the near-field model is generated and corrected, the vehicle's near-field module will periodically update the radar signal from the distance calculation module, the signal from the graphics acquisition and processing module, the near-field model obtained from the vehicle communication module, and the lane signal obtained from the vehicle positioning module to accurately correct the current vehicle's near-field model.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the vehicle driving scene recognition and control method as described in any one of claims 5 to 7.

Images