An intelligent vehicle lamp adaptive illumination interaction control method based on vehicle-road cooperation
By combining the K-nearest neighbor algorithm and light pheromones, a unified expression of the state of all traffic participants and adaptive lighting control are achieved, solving the problems of insufficient perception blind spots and overall coordination in existing vehicle lighting systems, and improving the safety and interactivity of nighttime driving lighting.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- LANCE VEHICLE TECH CO LTD
- Filing Date
- 2026-06-29
- Publication Date
- 2026-07-24
AI Technical Summary
Existing adaptive headlight systems suffer from limited perception range, blind spots, lack of integration with all traffic participants, incomplete dimming judgment for different zones, and inability to achieve coordinated lighting across the entire area, resulting in insufficient safety and interactivity for nighttime driving.
The K-nearest neighbor algorithm is used to uniformly represent the state of traffic lights across the entire region. Similarity state clusters are divided through vehicle-road cooperative interaction data, and light pheromones are released for adaptive lighting control. A two-layer lighting control architecture is constructed to realize the cooperative interaction of human and vehicle lights across the entire region.
It effectively solves the problems of traditional vehicle headlight dimming logic not adjusting in time and abrupt switching between light and dark in complex traffic and pedestrian environments, improving the safety and interactivity of nighttime driving lighting.
Smart Images
Figure CN122458285A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of electric lighting source control circuit technology, and more specifically to an intelligent vehicle lighting adaptive interactive control method based on vehicle-road cooperation. Background Technology
[0002] Early vehicle headlights only had basic functions such as manual high / low beam switching, turn signals, and side marker lights. Headlight control relied entirely on manual operation by the driver, which could only meet basic road lighting needs and could not adapt to the lighting requirements of complex road conditions at night. With the iteration of vehicle electronic control technology, single-vehicle adaptive headlight systems gradually appeared on the market. These systems relied on onboard cameras and light sensors to achieve basic automatic switching between high and low beams and fixed-angle dimming. At this stage, the headlight control logic only served the vehicle's own driving lighting and could only recognize static information such as the brightness of vehicles ahead and the road surface. The light control was simple and lagging.
[0003] With the rapid implementation of intelligent transportation, vehicle-to-vehicle, vehicle-to-infrastructure, and pedestrian-vehicle interconnection has become the mainstream development direction for adaptive headlight control. Existing intelligent headlights can perform small-scale adaptive adjustments to brightness and angle based on vehicle speed, road curvature, and ambient light, significantly reducing the burden of manual dimming. Current industry research focuses on visual perception dimming at the vehicle level and unified dimming of roadside streetlights. The mainstream research direction is to rely on the coordinated control of headlights across the entire road network to achieve interactive and coordinated traffic flow lighting, aiming to solve the problems of light interference from high-density traffic flow at night and lighting conflicts between pedestrians and vehicles.
[0004] Most existing adaptive headlights rely on onboard sensing devices to collect data, which has a limited sensing range and blind spots in curved and obstructed road sections. They do not combine the characteristics of traffic participants across the entire area to divide the lighting zones, and the correlation between road traffic flow and pedestrian and vehicle lighting is not utilized. The judgment of zone dimming is one-sided, and the ability to adapt to dynamic traffic flow is poor. At the same time, the existing headlight control lacks layered adaptation logic, does not distinguish between different lighting scenarios for main roads, auxiliary roads, and sidewalks, and provides insufficient protection for pedestrians and non-motorized traffic participants. In addition, the control of individual vehicle headlights cannot be uploaded to the roadside platform for closed-loop management, resulting in weak overall lighting collaborative control capabilities. The safety and interactivity of nighttime driving lighting cannot meet the standards of smart road driving. Summary of the Invention
[0005] In order to overcome the above-mentioned defects of the prior art, the present invention provides an intelligent vehicle headlight adaptive lighting interactive control method based on vehicle-road cooperation to solve the problems existing in the background art.
[0006] This invention provides the following technical solution: an intelligent vehicle headlight adaptive lighting interaction control method based on vehicle-road cooperation, comprising the following steps: Step S01: Real-time collection of vehicle-road cooperative interaction data: Based on the vehicle-road cooperative interaction data, construct a multi-dimensional raw feature dataset containing traffic participant status, vehicle lighting parameters, road geometric features, ambient lighting, and meteorological parameters; Step S02: Use the K-nearest neighbor algorithm to uniformly represent the traffic light status of the entire area: Use the K-nearest neighbor algorithm to perform spatial clustering on the multi-dimensional original feature dataset, divide the traffic units of the entire road area into various similarity state clusters, and output the traffic light environment status of the entire road area; Step S03: Release light pheromones based on the overall road lighting environment: Based on the overall road lighting environment, the vehicle releases light pheromones according to its own driving conditions and interaction needs; Step S04: Vehicle analyzes light pheromones for adaptive lighting control: Based on the light pheromones, the vehicle releases lights by self-analysis and outputs the analyzed headlight execution command for adaptive lighting control; Step S05: Output control data to the interactive control center: The interactive control center stores the control data through the vehicle-road cooperative network to complete the adaptive lighting interactive control of the intelligent vehicle lights.
[0007] Preferably, in step S01, the specific content of real-time collection of vehicle-road cooperative interaction data is as follows: Based on the C-V2X vehicle-road cooperative communication architecture, vehicle-road cooperative interaction data is collected in real time and synchronously. The vehicle-road cooperative interaction data includes: traffic participant status, vehicle lighting parameters, and road geometric feature parameters. The traffic participant status data includes: real-time location, instantaneous speed, and heading angle of surrounding vehicles, non-motorized vehicles, and pedestrians; vehicle lighting parameters include: light brightness parameters and light illumination angle parameters; and road geometric feature parameters include: road curvature and local ambient light intensity characteristics.
[0008] Preferably, in step S02, the specific content of using the K-nearest neighbor algorithm to uniformly represent the state of traffic lights across the entire region is as follows: The entire traffic area is divided into independent traffic units, and each independent traffic participant is divided into an independent dynamic traffic unit. The independent dynamic traffic units divided within the current road area constitute the integrated traffic unit set of the road at the current moment. The number of these units is dynamically updated according to the real-time distribution status of road traffic participants. A vehicle-road cooperative feature vector expression is constructed based on the multi-dimensional original feature dataset of each traffic unit. The feature vector expression is as follows: ,in Represents the lateral spatial coordinates of the transportation unit. Represents the longitudinal spatial coordinates of a transportation unit. Indicates instantaneous speed, Indicates the heading angle of the motion. Indicates the brightness parameter of the light. Indicates the parameter of the light illumination angle. Indicates the curvature characteristics of a local road. This represents the local ambient light intensity characteristics, and i represents the number of traffic units. Using the vehicle as the central reference for global state perception, it is marked as the vehicle reference unit. Based on the high-dimensional feature vectors of each traffic unit, the similarity of the state association between each traffic unit and the vehicle reference unit in the global domain is determined. Based on the state association similarity, the traffic units of the entire road area are divided into different similarity state clusters, and the lighting environment status of the entire road area is output.
[0009] Preferably, the method for determining the similarity of the state association between each traffic unit and the vehicle reference unit within the entire domain is as follows: A multi-dimensional feature hierarchical matching method is adopted, which divides the original feature data of each dimension into feature matching levels: Level 1 matching dimension: spatial location dimension, which includes horizontal spatial coordinates and vertical spatial coordinates, with a weight of 0.5; Level 2 matching dimension: motion state dimension, which includes instantaneous driving speed and motion heading angle, with a weight of 0.3; Level 3 matching dimension: lighting and environment dimension, which includes light brightness, light illumination angle, local road curvature and local ambient light intensity, with a weight of 0.2. The single-dimensional feature difference values between each traffic unit and the vehicle reference unit in the entire domain are calculated based on the multi-dimensional original feature dataset. Based on the hierarchical division of spatial location dimension, motion state dimension, and lighting and environment dimension, the comprehensive matching degree of each level dimension is calculated: the average difference value of all sub-features within the same level is taken to obtain the overall difference degree of each level dimension, and the level matching degree is obtained by subtracting the overall difference degree of each level dimension from 1. The level matching degree includes: spatial location dimension matching degree, motion state dimension matching degree, and lighting and environment dimension matching degree. The matching degree of spatial location dimension, motion state dimension, and lighting environment dimension is weighted and fused with the configuration weights of spatial location dimension, motion state dimension, and lighting and environment dimension to output the state correlation similarity between each traffic unit and the vehicle baseline unit in the entire domain.
[0010] Preferably, the specific content of the road-wide traffic unit is output as follows: The road-wide lighting environment status is divided into different similarity state clusters based on state association similarity. Using state association similarity as the sole clustering feature, a threshold-based hierarchical clustering method is employed to segment the entire traffic area into multiple independent state clusters based on similarity. Strong similarity state cluster: State association similarity ≥ high similarity threshold, Road-wide lighting environment status: Spatial range: Output all traffic unit numbers, corresponding road sections, and lane ranges contained in this cluster, Lighting interference characteristics: Strong light concentration points, road surface reflectivity, and light interference level on the vehicle's on-board photosensitive equipment; Medium similarity state cluster: Medium similarity threshold ≤ state association similarity < high similarity threshold, road-wide lighting environment status: spatial range: the road zone to which the traffic unit in this cluster belongs, and the spatial distance range from the vehicle reference unit, illumination offset characteristics: the duration of the peak illumination within the cluster being earlier or later than the reference unit, and the proportion of synchronous overlap of illumination. Weak similarity state cluster: State association similarity < medium similarity threshold, road-wide lighting environment status: spatial range: the remote road segment, isolated intersection, auxiliary road or pedestrian independent area to which the cluster unit belongs, illumination desynchronization characteristics: illumination change cycle, illuminance peak time period does not overlap with the reference unit.
[0011] Preferably, in step S03, the specific content of releasing light pheromones based on the overall road lighting environment status is as follows: Real-time reading of the road segment to which the vehicle belongs, its corresponding spatial range, and lighting parameters; Acquire vehicle driving condition parameters: real-time vehicle speed, lane position, following distance, acceleration and deceleration status, and vehicle dynamic driving behavior. The dynamic driving behavior includes lane changing, turning, starting and stopping, and overtaking. The vehicle driving condition parameters are basic parameters, equipped with preset corresponding basic lights. The rules for releasing light pheromones are as follows: Strong similarity state cluster light pheromone release logic: The release is triggered when a vehicle enters the cluster domain boundary, and the vehicle light signals of the traffic units in the cluster are output from the same source. The logic for releasing light pheromones in clusters with similarity states is as follows: the light pheromone release is differentiated according to the gradient of the cluster's distance from the reference unit, carrying fixed light delay parameters, and staggering the peak illuminance of vehicles across the entire domain. The logic for releasing light pheromones in weakly similar state clusters is as follows: release is triggered locally and independently, serving only non-motorized vehicles, pedestrians, and scattered motorized vehicles in the area, and does not link with the light signals of the main road reference unit.
[0012] Preferably, in step S04, the specific content of the vehicle analyzing light pheromones for adaptive lighting control is as follows: Preset basic lights are loaded based on the vehicle's driving condition parameters, and the lights are direct output lights; Based on light pheromones, the vehicle's emitted light is self-analyzed, and the directly output light is adaptively adjusted, specifically as follows: Strong similarity state cluster: Automatically turn off the high beam direct mode, unify the direct output light output to the vehicle's low beam reference illuminance, and adjust the headlight irradiance angle according to the road surface reflection intensity; Medium similarity state cluster: Taking the vehicle reference unit as the origin of coordinates, the four closed-loop distance gradients are divided along the vehicle's driving direction. Each gradient is matched with a corresponding brightness adjustment coefficient to correct the direct output light brightness by a multiplier. Weak similarity state cluster: The auxiliary road and pedestrian areas automatically switch to low brightness and soft low beam, turn off high frequency strong light warning and link flashing lights; the vehicle headlight brightness cycle operates autonomously, and when pedestrians and non-motorized vehicles approach, the vehicle headlight brightness is automatically reduced and the light illumination coverage is expanded.
[0013] Preferably, in step S05, the specific content of outputting control data to the interactive control center is as follows: The vehicle headlight adaptive lighting control data, associated cluster numbers, and light pheromone parameters are packaged and uploaded to the vehicle-road cooperative interactive control center. The interactive control center classifies and stores the zoned lighting control data through the full-domain vehicle-road cooperative communication network to complete the intelligent vehicle headlight adaptive lighting interactive closed-loop control.
[0014] The technical effects and advantages of this invention are as follows: This invention uses the K-nearest neighbor algorithm to dynamically divide motor vehicles, non-motor vehicles, and pedestrians into independent dynamic traffic units. It calculates the similarity of state associations by weighting and divides the similarity state clusters into three categories: strong, medium, and weak by hierarchical threshold. The clustering fits the dynamic change pattern of vehicle flow and pedestrian flow, providing stable data support for subsequent lighting interaction and vehicle headlight dimming. A dual-layer lighting control architecture is constructed, which relies on clustered light pheromones to achieve full-domain human-vehicle lighting collaborative interaction. By setting a dual-layer output mode of native working condition light + pheromone-corrected light, the vehicle outputs stable basic direct light according to the driving conditions. Then, combined with the differentiated release of light pheromones based on the characteristics of three types of related clusters, the brightness of the direct output light is corrected by a multiplier. This effectively avoids the problems of untimely adjustment and abrupt switching between light and dark when dealing with complex traffic and pedestrian flow using traditional fixed dimming logic. Attached Figure Description
[0015] Figure 1 This is a flowchart illustrating an intelligent vehicle headlight adaptive lighting interactive control method based on vehicle-road cooperation. Detailed Implementation
[0016] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. In addition, the forms of the various structures described in the following embodiments are merely illustrative. The intelligent vehicle headlight adaptive lighting interactive control method based on vehicle-road cooperation involved in the present invention is not limited to the structures described in the following embodiments. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0017] like Figure 1 As shown, this invention provides an intelligent vehicle headlight adaptive lighting interaction control method based on vehicle-road cooperation, comprising the following steps: Step S01: Real-time collection of vehicle-road cooperative interaction data: Based on the vehicle-road cooperative interaction data, construct a multi-dimensional raw feature dataset containing traffic participant status, vehicle lighting parameters, road geometric features, ambient lighting, and meteorological parameters; Step S02: Use the K-nearest neighbor algorithm to uniformly represent the traffic light status of the entire area: Use the K-nearest neighbor algorithm to perform spatial clustering on the multi-dimensional original feature dataset, divide the traffic units of the entire road area into various similarity state clusters, and output the traffic light environment status of the entire road area; Step S03: Release light pheromones based on the overall road lighting environment: Based on the overall road lighting environment, the vehicle releases light pheromones according to its own driving conditions and interaction needs; Step S04: Vehicle analyzes light pheromones for adaptive lighting control: Based on the light pheromones, the vehicle releases lights by self-analysis and outputs the analyzed headlight execution command for adaptive lighting control; Step S05: Output control data to the interactive control center: The interactive control center stores the control data through the vehicle-road cooperative network to complete the adaptive lighting interactive control of the intelligent vehicle lights.
[0018] In this embodiment, it should be specifically explained that the specific content of real-time collection of vehicle-road cooperative interaction data in step S01 is as follows: Based on the C-V2X vehicle-road cooperative communication architecture, vehicle-road cooperative interaction data is collected in real time and synchronously. The vehicle-road cooperative interaction data includes: traffic participant status, vehicle lighting parameters, and road geometric feature parameters. The traffic participant status data includes: real-time location, instantaneous speed, and heading angle of surrounding vehicles, non-motorized vehicles, and pedestrians; vehicle lighting parameters include: light brightness parameters and light illumination angle parameters; and road geometric feature parameters include: road curvature and local ambient light intensity characteristics.
[0019] In this embodiment, it should be specifically explained that the specific content of using the K-nearest neighbor algorithm to uniformly express the state of traffic lights across the entire region in step S02 is as follows: The entire traffic area is divided into independent traffic units, and each independent traffic participant is divided into an independent dynamic traffic unit. A single vehicle, a single non-motorized vehicle, and a single pedestrian each correspond to an independent dynamic traffic unit. All dynamic traffic units rely on the vehicle-road cooperative perception network to collect status data in real time. The independent dynamic traffic units divided within the current road area constitute the integrated traffic unit set of the road at the current moment. The number of these units is dynamically updated according to the real-time distribution status of road traffic participants. A vehicle-road cooperative feature vector expression is constructed based on the multi-dimensional original feature dataset of each traffic unit. The feature vector expression is as follows: The original feature data for each dimension are all normalized to... Standardized values of the interval Represents the lateral spatial coordinates of the transportation unit. Represents the longitudinal spatial coordinates of a transportation unit. Indicates instantaneous speed, Indicates the heading angle of the motion. Indicates the brightness parameter of the light. Indicates the parameter of the light illumination angle. Indicates the curvature characteristics of a local road. The local ambient light intensity features are represented by i, which represents the number of traffic units. For motor vehicles, the original eight-dimensional data is directly used. For non-motor vehicles and pedestrians, the empty dimensions of vehicle lights are filled with the default lighting benchmark value of the regional environment, which does not affect the similarity matching and ensures that the vector dimensions are fully aligned. Using the vehicle as the central reference for global state perception, it is marked as the vehicle reference unit. Based on the high-dimensional feature vectors of each traffic unit, the similarity of the state association between each traffic unit and the vehicle reference unit in the global domain is determined. Based on the state association similarity, the traffic units of the entire road area are divided into different similarity state clusters, and the lighting environment status of the entire road area is output.
[0020] In this embodiment, it should be specifically explained that the method for determining the similarity of the state association between each traffic unit and the vehicle reference unit within the entire domain is as follows: A multi-dimensional feature hierarchical matching method is adopted, which divides the original feature data of each dimension into feature matching levels: Level 1 matching dimension: spatial location dimension, which includes horizontal spatial coordinates and vertical spatial coordinates, with a weight of 0.5; Level 2 matching dimension: motion state dimension, which includes instantaneous driving speed and motion heading angle, with a weight of 0.3; Level 3 matching dimension: lighting and environment dimension, which includes light brightness, light illumination angle, local road curvature and local ambient light intensity, with a weight of 0.2. Based on the multi-dimensional original feature dataset, the one-dimensional feature difference value between each traffic unit and the vehicle reference unit within the entire domain is calculated. The formula for calculating the one-dimensional feature difference value is as follows: ,in Let represent the single-dimensional feature difference value, and represent the k-th feature value of the i-th traffic unit to be matched. This represents the k-th eigenvalue of the vehicle reference unit; Based on the hierarchical division of spatial location dimension, motion state dimension, and lighting and environment dimension, the comprehensive matching degree of each level dimension is calculated: the average difference value of all sub-features within the same level is taken to obtain the overall difference degree of each level dimension, and the level matching degree is obtained by subtracting the overall difference degree of each level dimension from 1. The level matching degree includes: spatial location dimension matching degree, motion state dimension matching degree, and lighting and environment dimension matching degree. The matching degree of spatial location dimension, motion state dimension, and lighting environment dimension is weighted and fused with the configuration weights of spatial location dimension, motion state dimension, and lighting and environment dimension to output the state correlation similarity between each traffic unit and the vehicle baseline unit in the entire domain.
[0021] In this embodiment, it should be specifically explained that the specific content of the road-wide traffic unit being divided into different similarity state clusters based on state association similarity, and the output of the road-wide lighting environment status, is as follows: Using state association similarity as the sole clustering feature, a threshold-based hierarchical clustering method is employed to segment the entire traffic area into multiple independent similarity state clusters. The threshold is pre-stored within the vehicle lighting controller. Strong similarity state cluster: State association similarity ≥ high similarity threshold, Road-wide lighting environment status: Spatial range: Output all traffic unit numbers, corresponding road sections, and lane ranges contained in this cluster, Lighting interference characteristics: Strong light concentration points, road surface reflectivity, and light interference level on the vehicle's on-board photosensitive equipment; Medium similarity state cluster: Medium similarity threshold ≤ state association similarity < high similarity threshold, road-wide lighting environment status: spatial range: the road zone to which the traffic unit in this cluster belongs, and the spatial distance range from the vehicle reference unit, illumination offset characteristics: the duration of the peak illumination within the cluster being earlier or later than the reference unit, and the proportion of synchronous overlap of illumination. Weak similarity state cluster: State association similarity < medium similarity threshold, road-wide lighting environment status: spatial range: the remote road segment, isolated intersection, auxiliary road or pedestrian independent area to which the cluster unit belongs, illumination desynchronization characteristics: illumination change cycle, illuminance peak time period does not overlap with the reference unit.
[0022] In this embodiment, it should be specifically explained that the specific content of releasing light pheromones based on the overall road lighting environment status in step S03 is as follows: Real-time reading of the road segment to which the vehicle belongs, its corresponding spatial range, and lighting parameters; Acquire vehicle driving condition parameters: real-time vehicle speed, lane position, following distance, acceleration and deceleration status, and vehicle dynamic driving behavior. The dynamic driving behavior includes lane changing, turning, starting and stopping, and overtaking. The vehicle driving condition parameters are basic parameters, equipped with preset corresponding basic lights. The rules for releasing light pheromones are as follows: Strong similarity state cluster light pheromone release logic: The release is triggered when a vehicle enters the cluster domain boundary, and the vehicle light signals of the traffic units in the cluster are output from the same source. The logic for releasing light pheromones in clusters with similarity states is as follows: the light pheromone release is differentiated according to the gradient of the cluster's distance from the reference unit, carrying fixed light delay parameters, and staggering the peak illuminance of vehicles across the entire domain. The logic for releasing light pheromones in weakly similar state clusters is as follows: release is triggered locally and independently, serving only non-motorized vehicles, pedestrians, and scattered motorized vehicles in the area, and does not link with the light signals of the main road reference unit.
[0023] In this embodiment, it should be specifically explained that the specific content of the vehicle's analysis of light pheromones for adaptive lighting control in step S04 is as follows: The preset basic lights are loaded based on the vehicle's driving condition parameters. The lights are direct output lights, which are the original headlight output state preset by the vehicle factory and without environmental coupling intervention. Subsequently, based on the full-domain light pheromone analysis results, the direct output lights are subjected to secondary fine adaptive correction and adjustment without changing the basic headlight on / off mode. Based on light pheromones, the vehicle's emitted light is self-analyzed, and the directly output light is adaptively adjusted, specifically as follows: Strong similarity state cluster: Automatically shuts off the high beam direct beam mode, unifies the direct output light to the vehicle's low beam reference illuminance, and adjusts the headlight irradiance angle according to road surface reflectivity: Level 1: Weak reflectivity (road surface reflectivity ≤ 180 cd / m²), no irradiance angle adjustment; Level 2: Moderate reflectivity (180 cd / m² < road surface reflectivity ≤ 350 cd / m²), the headlights are adjusted downwards by 2° to shorten the direct beam distance and reduce the light's incidence on the road surface reflectivity; Level 3: ... Level 1 Strong Reflectivity: 350 cd / m² < Road surface reflectivity ≤ 600 cd / m², the headlights are tilted downwards by 4°, focusing the light on the road surface near the vehicle, avoiding reflections from the road surface at medium and long distances to the onboard photosensitive module; Level 4 Extremely Strong Reflectivity: Road surface reflectivity > 600 cd / m², the headlights are tilted downwards by 6°, limiting the light illumination range to the road surface 0-15m in front of the vehicle, completely avoiding strong light reflection glare and reducing the overall light interference level of the cluster area; Medium similarity state cluster: Taking the vehicle's reference unit as the origin, four closed-loop distance gradients are divided along the vehicle's driving direction. Each gradient is matched with a corresponding brightness adjustment coefficient to correct the direct output light brightness: First gradient: 0m~50m, Second gradient: 50m~120m, Third gradient: 120m~200m, Fourth gradient: greater than 200m; The basic luminous brightness of the direct output light is 100% of the standard value. First gradient: The illuminance timing offset difference is the smallest, maintaining the direct output light at 100% standard luminous brightness, matching the reference unit's illumination peak; Second gradient: Small timing offset, gradually reducing the brightness to 85% of the original brightness, weakening the superimposed illuminance of traffic lights at medium and close distances; Third gradient: The timing offset increases, gradually reducing the brightness to 70% of the original brightness, coordinating with pheromone delay parameters to stagger peak emission; Fourth gradient: The timing offset difference is the largest, slightly increasing the brightness to 90% of the original brightness, taking into account long-distance road lighting, while avoiding simultaneous glare from multiple vehicles at a distance. Weak similarity state cluster: The auxiliary road and pedestrian areas automatically switch to low brightness and soft low beam, turn off high frequency strong light warning and link flashing lights; the vehicle headlight brightness cycle operates autonomously, and when pedestrians and non-motorized vehicles approach, the vehicle headlight brightness is automatically reduced and the light illumination coverage is expanded.
[0024] In this embodiment, it should be specifically explained that the specific content of outputting control data to the interactive control center in step S05 is as follows: The vehicle headlight adaptive lighting control data, associated cluster numbers, and light pheromone parameters are packaged and uploaded to the vehicle-road cooperative interactive control center. The interactive control center classifies and stores the zoned lighting control data through the full-domain vehicle-road cooperative communication network to complete the intelligent vehicle headlight adaptive lighting interactive closed-loop control.
[0025] The main difference between this implementation and the existing technology is that this embodiment uses the K-nearest neighbor algorithm to dynamically divide motor vehicles, non-motor vehicles and pedestrians into independent dynamic traffic units. It calculates the similarity of state associations by weighting and divides the state clusters into three categories of strong, medium and weak similarity by hierarchical threshold. The clustering fits the dynamic change pattern of traffic flow and pedestrian flow, and provides stable data support for subsequent lighting interaction and vehicle headlight dimming. A dual-layer lighting control architecture is constructed, which relies on clustered light pheromones to achieve full-domain human-vehicle lighting collaborative interaction. By setting a dual-layer output mode of native working condition light + pheromone-corrected light, the vehicle outputs stable basic direct light according to the driving conditions. Then, combined with the differentiated release of light pheromones based on the characteristics of three types of related clusters, the brightness of the direct output light is corrected by a multiplier. This effectively avoids the problems of untimely adjustment and abrupt switching between light and dark when dealing with complex traffic and pedestrian flow using traditional fixed dimming logic.
[0026] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
[0027] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
Claims
1. A smart vehicle headlight adaptive lighting interactive control method based on vehicle-road cooperation, characterized in that: Includes the following steps: Step S01: Real-time collection of vehicle-road cooperative interaction data: Based on the vehicle-road cooperative interaction data, construct a multi-dimensional raw feature dataset containing traffic participant status, vehicle lighting parameters, road geometric features, ambient lighting, and meteorological parameters; Step S02: Use the K-nearest neighbor algorithm to uniformly represent the traffic light status of the entire area: Use the K-nearest neighbor algorithm to perform spatial clustering on the multi-dimensional original feature dataset, divide the traffic units of the entire road area into various similarity state clusters, and output the traffic light environment status of the entire road area; Step S03: Release light pheromones based on the overall road lighting environment: Based on the overall road lighting environment, the vehicle releases light pheromones according to its own driving conditions and interaction needs; Step S04: Vehicle analyzes light pheromones for adaptive lighting control: Based on the light pheromones, the vehicle releases lights by self-analysis and outputs the analyzed headlight execution command for adaptive lighting control; Step S05: Output control data to the interactive control center: The interactive control center stores the control data through the vehicle-road cooperative network to complete the adaptive lighting interactive control of the intelligent vehicle lights.
2. The intelligent vehicle lighting adaptive interactive control method based on vehicle-road cooperation according to claim 1, characterized in that: In step S01, the specific content of real-time collection of vehicle-road cooperative interaction data is as follows: Based on the C-V2X vehicle-road cooperative communication architecture, vehicle-road cooperative interaction data is collected in real time and synchronously. The vehicle-road cooperative interaction data includes: traffic participant status, vehicle lighting parameters, and road geometric feature parameters. The traffic participant status data includes: real-time location, instantaneous speed, and heading angle of surrounding vehicles, non-motorized vehicles, and pedestrians; vehicle lighting parameters include: light brightness parameters and light illumination angle parameters; road geometric feature parameters include: Road curvature and local ambient light intensity characteristics.
3. The intelligent vehicle lighting adaptive interactive control method based on vehicle-road cooperation according to claim 1, characterized in that: In step S02, the specific details of using the K-nearest neighbor algorithm to uniformly represent the state of traffic lights across the entire region are as follows: The entire traffic area is divided into independent traffic units, and each independent traffic participant is divided into an independent dynamic traffic unit. The independent dynamic traffic units divided within the current road area constitute the integrated traffic unit set of the road at the current moment. The number of these units is dynamically updated according to the real-time distribution status of road traffic participants. A vehicle-road cooperative feature vector expression is constructed based on the multi-dimensional original feature dataset of each traffic unit. The feature vector expression is as follows: ,in Represents the lateral spatial coordinates of the transportation unit. Represents the longitudinal spatial coordinates of a transportation unit. Indicates instantaneous speed, Indicates the heading angle of the motion. Indicates the brightness parameter of the light. Indicates the parameter of the light illumination angle. Indicates the curvature characteristics of a local road. This represents the local ambient light intensity characteristics, and i represents the number of traffic units. Using the autonomous vehicle as the central reference for global state perception, it is marked as the autonomous vehicle reference unit. Based on the high-dimensional feature vectors of each traffic unit, the similarity of the state association between each traffic unit and the autonomous vehicle reference unit in the global domain is determined. Based on the state association similarity, the traffic units of the entire road area are divided into different similarity state clusters, and the lighting environment status of the entire road area is output.
4. The intelligent vehicle lighting adaptive interactive control method based on vehicle-road cooperation according to claim 3, characterized in that: The method for determining the similarity of the state association between each traffic unit and the vehicle reference unit within the entire domain is as follows: A multi-dimensional feature hierarchical matching method is adopted, which divides the original feature data of each dimension into feature matching levels: Level 1 matching dimension: spatial location dimension, which includes horizontal spatial coordinates and vertical spatial coordinates, with a weight of 0.5; Level 2 matching dimension: motion state dimension, which includes instantaneous driving speed and motion heading angle, with a weight of 0.3; Level 3 matching dimension: lighting and environment dimension, which includes light brightness, light illumination angle, local road curvature and local ambient light intensity, with a weight of 0.
2. The single-dimensional feature difference values between each traffic unit and the vehicle reference unit in the entire domain are calculated based on the multi-dimensional original feature dataset. Based on the hierarchical division of spatial location dimension, motion state dimension, and lighting and environment dimension, the comprehensive matching degree of each level dimension is calculated: the average difference value of all sub-features within the same level is taken to obtain the overall difference degree of each level dimension, and the level matching degree is obtained by subtracting the overall difference degree of each level dimension from 1. The level matching degree includes: spatial location dimension matching degree, motion state dimension matching degree, and lighting and environment dimension matching degree. The matching degree of spatial location dimension, motion state dimension, and lighting environment dimension is weighted and fused with the configuration weights of spatial location dimension, motion state dimension, and lighting and environment dimension to output the state correlation similarity between each traffic unit and the vehicle baseline unit in the entire domain.
5. The intelligent vehicle lighting adaptive interactive control method based on vehicle-road cooperation according to claim 3, characterized in that: The road-wide traffic units are divided into different similarity state clusters based on state association similarity, and the specific content of the road-wide lighting environment status is output as follows: Using state association similarity as the sole clustering feature, a threshold-based hierarchical clustering method is employed to segment the entire traffic area into multiple independent state clusters based on similarity. Strong similarity state cluster: State association similarity ≥ high similarity threshold, Road-wide lighting environment status: Spatial range: Output all traffic unit numbers, corresponding road sections, and lane ranges contained in this cluster, Lighting interference characteristics: Strong light concentration points, road surface reflectivity, and light interference level on the vehicle's on-board photosensitive equipment; Medium similarity state cluster: Medium similarity threshold ≤ state association similarity < high similarity threshold, road-wide lighting environment status: Spatial range: the road zone to which this traffic unit belongs, and the spatial distance range from the vehicle reference unit, illumination offset characteristics: The duration by which the peak illumination within the cluster precedes or lags the reference unit, and the percentage of synchronous overlap in illumination; Weak similarity state cluster: State association similarity < medium similarity threshold, road-wide lighting environment status: spatial range: the remote road segment, isolated intersection, auxiliary road or pedestrian independent area to which the cluster unit belongs, illumination desynchronization characteristics: illumination change cycle, illuminance peak time period does not overlap with the reference unit.
6. The intelligent vehicle lighting adaptive interactive control method based on vehicle-road cooperation according to claim 1, characterized in that: In step S03, the specific details of releasing light pheromones based on the overall road lighting environment are as follows: Real-time reading of the road segment to which the vehicle belongs, its corresponding spatial range, and lighting parameters; Acquire vehicle driving condition parameters: real-time vehicle speed, lane position, following distance, acceleration and deceleration status, and vehicle dynamic driving behavior. The dynamic driving behavior includes lane changing, turning, starting and stopping, and overtaking. The vehicle driving condition parameters are basic parameters, equipped with preset corresponding basic lights. The rules for releasing light pheromones are as follows: Strong similarity state cluster light pheromone release logic: The release is triggered when a vehicle enters the cluster domain boundary, and the vehicle light signals of the traffic units in the cluster are output from the same source. The logic for releasing light pheromones in clusters with similarity states is as follows: the light pheromone release is differentiated according to the gradient of the cluster's distance from the reference unit, carrying fixed light delay parameters, and staggering the peak illuminance of vehicles across the entire domain. The logic for releasing light pheromones in weakly similar state clusters is as follows: release is triggered locally and independently, serving only non-motorized vehicles, pedestrians, and scattered motorized vehicles in the area, and does not link with the light signals of the main road reference unit.
7. The intelligent vehicle lighting adaptive lighting interactive control method based on vehicle-road cooperation according to claim 1, characterized in that: In step S04, the specific details of the vehicle's adaptive lighting control based on light pheromones are as follows: Preset basic lights are loaded based on the vehicle's driving condition parameters, and the lights are direct output lights; Based on light pheromones, the vehicle's emitted light is self-analyzed, and the directly output light is adaptively adjusted, specifically as follows: Strong similarity state cluster: Automatically turn off the high beam direct mode, unify the direct output light output to the vehicle's low beam reference illuminance, and adjust the headlight irradiance angle according to the road surface reflection intensity; Medium similarity state cluster: Taking the vehicle reference unit as the origin of coordinates, the four closed-loop distance gradients are divided along the vehicle's driving direction. Each gradient is matched with a corresponding brightness adjustment coefficient to correct the direct output light brightness by a multiplier. Weak similarity state cluster: The auxiliary road and pedestrian areas automatically switch to low brightness and soft low beam, turn off high frequency strong light warning and link flashing lights; the vehicle headlight brightness cycle operates autonomously, and when pedestrians and non-motorized vehicles approach, the vehicle headlight brightness is automatically reduced and the light illumination coverage is expanded.
8. The intelligent vehicle lighting adaptive interactive control method based on vehicle-road cooperation according to claim 1, characterized in that: In step S05, the specific content of outputting control data to the interactive control center is as follows: The vehicle headlight adaptive lighting control data, associated cluster numbers, and light pheromone parameters are packaged and uploaded to the vehicle-road cooperative interactive control center. The interactive control center classifies and stores the zoned lighting control data through the full-domain vehicle-road cooperative communication network to complete the intelligent vehicle headlight adaptive lighting interactive closed-loop control.