Intelligent vehicle lamp system based on DLP projection, control method and vehicle

Abstract

The invention relates to the technical field of vehicle lamp systems, in particular to an intelligent vehicle lamp system based on DLP projection, a control method and a vehicle. The intelligent vehicle lamp system based on DLP projection comprises an optical system, an intelligent control system, an execution module and a power management system, the system collects environment data in real time through a sensor, a central processing unit fuses multi-source information and generates a control instruction through an AI algorithm, and the light source brightness, the irradiation angle and the road surface projection content are dynamically adjusted; meanwhile, voice interaction and remote control are supported, and the self-adaptive lighting and safety warning functions are achieved. Through the intelligent control and DLP projection technology, the vehicle lamp performance is improved, the lighting effect is optimized, and the night driving safety is enhanced; self-adaptive adjustment is realized, and the driving experience is improved; the energy consumption is reduced by more than 30%; the battery life is prolonged; v2X communication is supported, and intelligent early warning is realized; a voice control function is provided, and operation is more convenient and safer; a road surface projection function is added, navigation arrows and warning marks can be displayed, and visual driving assistance is provided.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle lighting systems, and in particular to a DLP projection-based intelligent vehicle lighting system, a control method, and a vehicle. Background Art

[0002] Existing headlight systems primarily rely on manual operation or simple light sensors to control the on / off and brightness adjustment of the lights. Some high-end vehicles are equipped with adaptive high beam and cornering lighting functions, but the implementation of these functions is relatively simple and has the following problems: Inconvenient manual operation: The driver needs to frequently adjust the lights manually, which distracts the driver and increases safety risks; Lack of intelligence: Existing light sensors cannot distinguish complex environments (such as tunnels and rainy weather), which can easily lead to false triggering; Low energy efficiency: Fixed brightness adjustment mode cannot adapt to dynamic road conditions, resulting in energy waste; Lack of remote control: Lighting parameters cannot be adjusted remotely in real time, making it difficult to meet the needs of fleet management or shared vehicles. Single function: The existing headlight system has a relatively single function and cannot meet the diverse lighting needs under complex road conditions; Lack of voice interaction: Existing lighting systems do not support voice control, making it difficult for drivers to easily adjust lighting settings while driving. Lack of projection function: Existing lighting systems are unable to project navigation information or warning signs on the road surface, and cannot provide more intuitive driving assistance. Summary of the Invention

[0003] The technical problem to be solved by the present invention is: in order to solve the problems of the existing technology in the above-mentioned background technology, such as reliance on manual operation, low intelligence, single function, high energy consumption, lack of interactive function, lack of voice interaction and lack of projection function, an intelligent car lighting system based on DLP projection is provided.

[0004] The technical solution adopted by the present invention to solve the technical problem is: an intelligent vehicle lighting system based on DLP projection, comprising: The optical system includes a light source module, an optical lens group, and a DLP projection module. The light source module uses a high-brightness LED or laser light source, and the DLP projection module integrates a high-resolution DMD chip. Intelligent control system, including sensor module, central processing unit, voice recognition module and communication module; An execution module, including a headlight drive circuit, a mechanical adjustment mechanism, and a projection drive circuit; Power management system, including power management chip and backup power supply; Among them, the sensor module collects environmental data in real time, the voice recognition module enters the listening state synchronously, the communication module receives remote control signals, and the central processing unit generates control instructions through multi-sensor data fusion and AI algorithm analysis. The execution module controls the car light drive circuit to adjust the brightness, controls the mechanical adjustment mechanism to adjust the illumination angle, and controls the DLP projection module to generate projection content according to the control instructions.

[0005] The optical system uses high-brightness LED / laser light sources and DLP projection modules to achieve basic lighting and road information projection; the intelligent control system integrates multimodal perception (sensors), decision-making (AI algorithms), interaction (voice recognition) and communication functions; the execution module accurately executes various control instructions; the power management system ensures stable power supply; the lighting system is intelligent, adaptive and multifunctional, and can automatically adjust lighting strategies according to complex environments, significantly improving driving safety.

[0006] According to one embodiment of the present invention, the brightness of the light source module can be adjusted within a range of 5000-15000 lumens, and the steering angle accuracy of the mechanical adjustment mechanism is 0.5°.

[0007] A wide brightness adjustment range of 5,000-15,000 lumens can adapt to various lighting environments, from tunnels to rainy and foggy weather; 0.5° high-precision mechanical adjustment ensures accurate deflection of the light beam when lighting corners, ensuring optimized performance of the lighting system under various operating conditions, meeting regulatory requirements and enhancing the driving experience.

[0008] According to one embodiment of the present invention, the sensor module includes a camera, a radar, a photosensor and a rain sensor; the communication module includes a 4G / 5G mobile communication unit, a Wi-Fi wireless unit and a Bluetooth unit.

[0009] The combination of multiple types of sensors enables all-weather, all-scenario environmental perception; the multi-mode communication solution of 4G / 5G, Wi-F, and Bluetooth supports both remote control and V2X vehicle-to-everything (V2X) interaction; and the construction of a complete environmental perception network and communication link provides a data foundation for intelligent decision-making.

[0010] According to one embodiment of the present invention, the resolution of the DMD chip is 1024×768, and the refresh frequency is 60 Hz.

[0011] The 1024×768 resolution ensures clear projection content, while the 60Hz high refresh rate prevents projection flicker. The technical effect of this claim is to ensure the visibility and stability of road projection information, enabling the optimal display of warning signs, navigation arrows, and other content.

[0012] A control method for the DLP projection-based intelligent vehicle lighting system described in the above solution is also provided, characterized in that it includes the following steps: S1. Data acquisition: The camera collects road images and identifies lane lines, pedestrians, or traffic signs; the radar monitors the distance, speed, and azimuth of obstacles ahead in real time; the light sensor detects ambient light intensity, and the rain sensor identifies rainfall intensity; the voice recognition module collects voice commands; and the communication module receives remote control signals. S2. Preprocessing: Use median filtering and Gaussian filtering to eliminate image noise; synchronize the timestamps of each sensor data through the hardware clock; set the 3σ principle to eliminate sudden abnormal sensor data; S3. Feature processing: Use the YOLOv8 neural network to extract target contour, color, and motion vector features; analyze the radial velocity of obstacles through Fourier transform and predict the motion trajectory through Kalman filtering; fit the ambient brightness curve based on the light intensity data and generate weather labels based on the rainfall data; assign weights based on the sensor confidence level. The specific formula is as follows:

[0013] in, is the sensor weight, It is sensor data; when the camera and radar detect the target position conflict, the radar data is used first and the secondary verification mechanism is triggered; S4, Decision Analysis: Train a CNN-LSTM hybrid model based on over 100,000 kilometers of actual road data; calculate and mark the possible dangerous areas that may be entered within the next 3 seconds based on the target coordinates and velocity vector; and establish a danger level matrix; S5. Control strategy generation: Over 200 control rules are preset, and the optimal control sequence is generated using model predictive control. The lighting area of ​​the LED light source array is dynamically adjusted, arrow guidance is generated based on navigation data, and warning symbols are generated based on dangerous areas, with the projection distance controlled to be 5-15 meters in front of the vehicle. S6, execution feedback: transmit control instructions through the CAN bus, execute status feedback and projection effect feedback; S7, Algorithm optimization: Compare the instruction parameters with the actual execution results, use the reinforcement learning algorithm, set the reward function, and continuously optimize the control strategy.

[0014] The complete closed-loop process from data collection to algorithm optimization reflects the system's intelligence level; through multi-sensor fusion, AI decision-making and real-time feedback optimization, the headlight control response time is ≤80ms and the hazard recognition accuracy is ≥95%, significantly better than traditional headlight systems.

[0015] According to one embodiment of the present invention, in step S1, the frame rate of the camera collecting road conditions is ≥30fps, the light intensity range of the ambient light detected by the photosensor is 0-100klx, and the sampling rate of the voice recognition module collecting voice commands is 16kHz.

[0016] The 30fps high-frame rate camera ensures dynamic target capture; the 0-100klx light intensity detection range covers various environments from darkness to strong light; the 16kHz voice sampling ensures command recognition accuracy; and the system's real-time performance and environmental adaptability indicators are quantified.

[0017] According to one embodiment of the present invention, the weights of the camera, radar, and light sensor in step S3 are 0.4, 0.3, and 0.2, respectively.

[0018] The weight distribution of cameras, radars, and light-sensitive sensors has been verified by a large amount of measured data, which can optimize computing resource usage while ensuring detection accuracy; it achieves optimal fusion of sensor data and reduces the false alarm rate by more than 40%.

[0019] According to an embodiment of the present invention, the danger levels in step S4 are specifically pedestrians>obstacles>curves>normal road conditions, and the priority coefficients are 1.0, 0.8, 0.6 and 0.3 respectively.

[0020] The priority setting of pedestrians (1.0) > obstacles (0.8) > curves (0.6) complies with traffic safety principles, allowing system resources to prioritize high-risk targets and making the hazard response strategy more scientific and reasonable.

[0021] According to an embodiment of the present invention, generating the optimal control sequence in step S5 needs to consider the vehicle lamp response delay, specifically: LED driving delay ≤ 50ms, DLP projection delay ≤ 30ms.

[0022] Delay control of LED driving and DLP projection ensures that the response time of the entire process from perception to execution is less than 200ms, meeting real-time requirements and avoiding safety hazards caused by delays.

[0023] A vehicle is also provided, equipped with the DLP projection-based intelligent vehicle lighting system described in the above solution. A vehicle equipped with this system can significantly improve nighttime driving safety and intelligence.

[0024] Beneficial effects of the present invention: Using high-brightness LED or laser as light source, it provides high-brightness and high-contrast lighting effects, significantly improving nighttime driving safety; Through multi-sensor fusion technology, complex light pattern control is achieved, such as adaptive high beam, cornering lighting, dynamic high beam shielding, intelligent zone lighting and other functions, and the light pattern is automatically adjusted according to road conditions; Integrate multiple sensors (such as cameras, radars, and light sensors) and combine them with AI algorithms to automatically switch lighting modes based on road conditions, weather, and vehicle status; Adopting efficient power management and adaptive brightness adjustment technology to reduce the energy consumption of the lighting system and extend the vehicle battery life; Support interaction with other vehicles or traffic infrastructure, realize intelligent warning and communication functions, and improve traffic safety; Support users to remotely control the lighting parameters through mobile phone APP or cloud server, and monitor the lighting status in real time, meeting the needs of fleet management and shared vehicles; Through the voice recognition module, the driver can safely and quickly adjust the lighting settings through voice commands while driving, improving driving convenience and safety; Using DLP technology to project navigation information or warning signs onto the road surface, it provides more intuitive driving assistance and enhances driving safety. BRIEF DESCRIPTION OF THE DRAWINGS The present invention will be further described below with reference to the accompanying drawings and examples.

[0025] Figure 1 This is a structural block diagram of an intelligent vehicle lighting system according to an embodiment of the present invention.

[0026] Figure 2 This is a flow chart of a method for controlling an intelligent vehicle lighting system according to a second embodiment of the present invention.

[0027] Figure 3 yes Figure 2 Specific flow chart for decision analysis and control strategy generation.

[0028] In the figure: 1. Optical system; 11. Light source module; 12. Optical lens group; 13. DLP projection module; 2. Intelligent control system; 21. Sensor module; 211. Camera; 212. Radar; 213. Photosensor; 214. Rain sensor; 22. Central processing unit; 23. Voice recognition module; 24. Communication module; 3. Execution module; 31. Headlight drive circuit; 32. Mechanical adjustment mechanism; 33. Projection drive circuit; 4. Power management system; 41. Power management chip; 42. Backup power supply. DETAILED DESCRIPTION

[0029] The present invention will now be described in further detail with reference to the accompanying drawings, which are simplified schematic diagrams that illustrate the basic structure of the present invention in a schematic manner.

[0030] Example 1 like Figure 1As shown, a DLP projection-based intelligent vehicle lighting system includes an optical system 1, an intelligent control system 2, an execution module 3, and a power management system 4. Optical system 1 includes a light source module 11, an optical lens assembly 12, and a DLP projection module 13. Light source module 11 uses a high-brightness LED or laser light source, and DLP projection module 13 integrates a high-resolution DMD chip. Intelligent control system 2 includes a sensor module 21, a central processing unit 22, a voice recognition module 23, and a communication module 24. Execution module 3 includes a vehicle lighting drive circuit 31, a mechanical adjustment mechanism 32, and a projection drive circuit 33. Power management system 4 includes a power management chip 41 and a backup power supply 42.

[0031] High-brightness LEDs or lasers are used as light sources, providing high-brightness, high-contrast lighting. The optical lens assembly 12 optimizes beam distribution to ensure uniform lighting and compliance with regulatory requirements. The DLP projection module 13 integrates DLP (digital light processing) technology and is equipped with a high-resolution DMD (digital micromirror device) chip for projecting navigation information or warning signs onto the road surface. The sensor module 21 integrates multiple sensors, including a camera 211 (for road condition recognition), a radar 212 (for distance monitoring), a photosensor 213 (for ambient light intensity monitoring), and a rain sensor 214 (for weather monitoring).

[0032] The central processing unit 22 uses AI algorithms to analyze sensor data and generate lighting control commands (such as on / off, brightness, illumination angle, and light pattern switching). The voice recognition module 23 integrates voice recognition technology, allowing the driver to control functions such as the lights on / off, brightness adjustment, and light pattern switching through voice commands. The communication module 24 is equipped with wireless communication modules such as 4G / 5G, Wi-Fi, and Bluetooth, supporting remote control and interaction with other vehicles and transportation infrastructure.

[0033] The headlight driving circuit 31 controls the brightness and light type of the light source according to the instructions of the central processor 22; the mechanical adjustment mechanism 32 is used to adjust the illumination angle and direction of the headlight; the projection driving circuit 33 controls the projection content and brightness of the DLP projection module 13 according to the instructions of the central processor 22.

[0034] The high-efficiency power management chip 41 optimizes power distribution and power consumption control, extending the service life of the vehicle lighting system; the backup power supply 42 provides emergency lighting when the vehicle battery fails.

[0035] Specifically, the sensor module 21 collects environmental data in real time, the voice recognition module 23 enters the listening state synchronously, the communication module 24 receives the remote control signal, the central processing unit 22 generates control instructions through multi-sensor data fusion and AI algorithm analysis, and the execution module 3 controls the headlight drive circuit 31 to adjust the brightness, controls the mechanical adjustment mechanism 32 to adjust the illumination angle, and controls the DLP projection module 13 to generate projection content according to the control instructions.

[0036] Example 2 like Figure 2 As shown, the control method of the intelligent vehicle lighting system based on DLP projection according to the first embodiment includes the following steps: S1. Data acquisition: The camera 211 collects road images and identifies lane lines, pedestrians, or traffic signs; the radar 212 monitors the distance, speed, and azimuth of obstacles ahead in real time; the light sensor 213 detects ambient light intensity, and the rain sensor 214 identifies rainfall intensity; the voice recognition module 23 collects voice commands; and the communication module 24 receives remote control signals. S2. Preprocessing: Use median filtering and Gaussian filtering to eliminate image noise; synchronize the timestamps of each sensor data through the hardware clock; set the 3σ principle to eliminate sudden abnormal sensor data; S3. Feature processing: Use the YOLOv8 neural network to extract target contour, color, and motion vector features; analyze the radial velocity of obstacles through Fourier transform and predict the motion trajectory through Kalman filtering; fit the ambient brightness curve based on the light intensity data and generate weather labels based on the rainfall data; assign weights based on the sensor confidence level. The specific formula is as follows:

[0037] in, is the sensor weight, is sensor data; when the camera 211 and the radar 212 detect a conflict in target position, the radar 212 data is used first and the secondary verification mechanism is triggered; S4, Decision Analysis: Train a CNN-LSTM hybrid model based on over 100,000 kilometers of actual road data; calculate and mark the possible dangerous areas that may be entered within the next 3 seconds based on the target coordinates and velocity vector; and establish a danger level matrix; S5. Control strategy generation: Over 200 control rules are preset, and the optimal control sequence is generated using model predictive control. The lighting area of ​​the LED light source array is dynamically adjusted, arrow guidance is generated based on navigation data, and warning symbols are generated based on dangerous areas, with the projection distance controlled to be 5-15 meters in front of the vehicle. S6, execution feedback: transmit control instructions through the CAN bus, execute status feedback and projection effect feedback; S7, Algorithm optimization: Compare the instruction parameters with the actual execution results, use the reinforcement learning algorithm, set the reward function, and continuously optimize the control strategy.

[0038] In step S1, the frame rate of the camera 211 capturing road conditions is ≥30fps, the light sensor 213 detects ambient light with an intensity range of 0-100klx, and the sampling rate of the voice recognition module 23 capturing voice commands is 16kHz. In step S3, the weights of the camera 211, radar 212, and light sensor 213 are 0.4, 0.3, and 0.2, respectively. In step S4, the hazard levels are specifically pedestrians > obstacles > curves > normal road conditions, with priority coefficients of 1.0, 0.8, 0.6, and 0.3, respectively. Generating the optimal control sequence in step S5 requires considering the vehicle headlight response delay, specifically: LED drive delay ≤50ms, DLP projection delay ≤30ms.

[0039] like Figure 3 As shown in FIG, the detailed operations of step S4 and step S5. Step S4 specifically includes: inputting the target feature data extracted by the camera 211, the obstacle distance, speed and azimuth data detected by the radar 212, the ambient light intensity data of the photosensor 213 and the rainfall level data of the rain sensor 214, processing visual recognition, identifying road structures (such as lane lines, tunnel entrances, curves), traffic signs, pedestrians / vehicles and other targets; analyzing time series data (such as obstacle movement trends, light intensity change curves), predicting the environmental state within the next 3 seconds, and outputting scene labels (such as "tunnel entrance", "curve driving") with a confidence level of ≥92%; inputting the target coordinates (x, y), velocity vector (v x , v y ), acceleration (a x , a y ), the kinematic equation is used to predict the trajectory in the next 3s. The calculation formula is as follows:

[0040] , Among them, x0 and y0 are the starting coordinates, and t represents the time; The area that may enter the vehicle path is marked as ROI (region of interest) and marked with a red frame; then the danger level is set: pedestrians (1.0) > obstacles (0.8) > curves (0.6) > normal road conditions (0.3), and the trigger conditions are: distance <10m, speed >1m / s; distance <15m, located in the lane; curvature radius <50m, vehicle speed >40km / h; no special targets.

[0041] Step S5 specifically includes: presetting more than 200 control rules, for example: Rule #47 (Tunnel Scenario): When a tunnel entrance is detected (light intensity < 500 lux) and the vehicle speed is > 60 km / h, the system automatically switches to tunnel lighting mode (brightness increased by 30%, and light pattern diffusion angle increased by 15°). Rule #89 (Pedestrian Warning): When a pedestrian is detected at a distance less than 10m, the DLP projection module 13 generates a red warning frame and projects it onto the road ahead of the pedestrian. Input the current light status, environmental parameters, and vehicle CAN bus data (such as speed and steering angle), and use the minimization cost function: , The safety error is the proportion of the unilluminated area in the hazardous area. The energy consumption is the weighted sum of the LED drive current and the power consumption of the DLP projection module 13, with weights α = 0.7 and β = 0.3. The control sequence for the next 1 second is output, including the LED current, light type, and projection content. The geometric transformation parameters of the DLP image are adjusted according to the vehicle speed (v) and projection distance (d): , Ensure that the projection is distortion-free within a distance of 5-15m. An example of projection content includes navigation guidance, a green arrow (size 1m × 0.5m), triggered by a steering command from the in-vehicle navigation system. Feedforward control is used to send commands 50ms in advance, ensuring LED drive delay ≤50ms. Preloading the next frame of content through the frame buffer ensures DLP projection delay ≤30ms, ultimately keeping the end-to-end delay from decision-making to execution within 80ms.

[0042] In step S5, the LED light source array's illuminated area is dynamically adjusted to implement adaptive high beam (e.g., shielding oncoming vehicles) and cornering-focused lighting (light pattern deflects 10° when steering angle ≥ 15°). Arrow guidance (resolution 1280×720) is generated based on navigation data, and warning symbols (e.g., "No Overtaking" icon) are generated for hazardous scenarios. The projection distance is controlled within 5-15 meters in front of the vehicle.

[0043] Step S6 specifically includes adjusting the LED drive current, with a range of 0-1.5A and an accuracy of 10mA; controlling the angle of the DMD chip micromirror of the DLP projection module 13, with a resolution of 1024×768 and a refresh rate of 60Hz; and controlling the stepper motor of the mechanical adjustment mechanism 32, with an angle accuracy of 0.5°.

[0044] The control method of this embodiment utilizes the collaborative work of multiple heterogeneous sensors to build a complete driving environment perception system. Camera 211 captures high-precision road imagery at a frame rate of ≥30 fps and, combined with the YOLOv8 neural network, achieves pedestrian recognition accuracy of ≥95%. Millimeter-wave radar 212 provides distance monitoring with an accuracy of ±0.5 m and, through Fourier transform and Kalman filtering, achieves trajectory prediction with an error of ≤0.3 m. Data fusion from photosensor 213 (0-100 klx) and rain sensor 214 enables accurate identification of over 20 complex scenarios, including tunnels and heavy rain. This multimodal perception solution overcomes the limitations of a single sensor and maintains stable environmental perception even in extreme conditions such as dense fog and strong sunlight.

[0045] A combined denoising algorithm of median filtering and Gaussian filtering is used in the preprocessing stage, which improves the image signal-to-noise ratio by more than 40%. Hardware clock synchronization technology (error ≤ 10ms) is used to solve the problem of asynchronous multi-sensor data. The 3σ outlier filtering mechanism effectively eliminates instantaneous sensor interference. In the feature processing stage, a weighted fusion model is innovatively adopted. When sensor data conflicts, radar 212 data is prioritized and secondary verification is initiated, enabling the system to maintain a decision reliability of ≥ 98% even in special circumstances such as camera 211 glare or radar 212 multipath effects.

[0046] A CNN-LSTM hybrid model, trained on 100,000 kilometers of real-world vehicle data, achieves an impressive 98% recognition rate in tunnel scenarios and 92% in rainy and foggy conditions. By calculating motion trajectories over the next three seconds and marking hazardous areas, combined with a hazard level matrix (pedestrians (1.0) > obstacles (0.8) > curves (0.6), the system possesses true predictive decision-making capabilities. Compared to traditional reactive control, this approach can shorten hazard warning times by 300-500ms, providing critical reaction time for systems like automatic emergency braking.

[0047] The combination of more than 200 preset control rules and model prediction (MPC) enables optimal dynamic adjustment of lighting parameters; precise zoning control of the LED light source array (driving current accuracy of 10mA) combined with a motor steering mechanism with 0.5° accuracy ensures that the adaptive high beam's oncoming vehicle shielding response time is ≤50ms; the DLP projection module 13 uses real-time perspective transformation correction (distortion rate <5%) to ensure that warning signs within a projection distance of 5-15m always maintain optimal visibility.

[0048] The closed-loop feedback mechanism implemented via the CAN bus (500kbps) can monitor parameters such as brightness deviation (>5% triggers calibration) and projection positioning error in real time. The reinforcement learning algorithm used uses the reward function of safety system × 0.7 + energy efficiency × 0.3. After 2000 learning iterations, the system can reduce energy consumption by 22% in rainy and foggy weather while maintaining the lighting effect.

[0049] The 16kHz sampling voice recognition module 23 supports natural language commands with a recognition accuracy of 97%; combined with 4G / 5G remote control functions, users can preset lighting preferences through a mobile phone app; the navigation arrows and warning symbols generated by the DLP projection module 13 are ergonomically designed, enabling the driver to complete information recognition within 0.3 seconds, significantly reducing the cognitive load of driving.

[0050] Let's take two specific examples to illustrate: (1) Adaptive lighting control in tunnel scenarios When the vehicle approaches the tunnel entrance at a speed of 80 km / h, the system detects the following data: the light sensor 213 detects that the ambient light intensity drops sharply from 2000 lux to 300 lux (threshold 500 lux); the camera 211 recognizes the tunnel outline with a confidence level of 98%; the radar 212 detects a slow-moving vehicle 150 meters ahead; the central processor 22 integrates the light intensity mutation (weight 0.2), visual recognition (0.4) and radar 212 data (0.3) to confirm the tunnel scene and the comprehensive confidence level. 96%; activating tunnel lighting mode, preset rule #47, increasing brightness from 6000 lux to 10000 lux, with a gradual transition time of 200 ms, and adjusting the light pattern diffusion angle from 25° to 40° to cover the tunnel sidewalls; DLP projection module 13 outputs the text "Tunnel Speed ​​Limit 80" at a projection distance of 10 meters; mechanical adjustment mechanism 32 deflects at a speed of 0.5° / ms, and the actual brightness detection value is 9820 lux, with an error of 1.8% < the 5% threshold. A success signal is fed back via the CAN bus.

[0051] Compared with the existing technology, this embodiment improves the tunnel lighting uniformity to 0.85, the driver's recognition distance of obstacles in the tunnel is increased by 35m, and the brightness contrast of the projected text at a distance of 10m reaches 5:1.

[0052] (2) Pedestrian crossing warning scene On a rainy night with a light intensity of 200 lux and heavy rain, radar 212 detects a laterally moving target at a distance of 12 meters, with a speed of 3 meters per second. Camera 211 identifies the pedestrian's outline with a confidence level of 92%. Rain sensor 214 confirms the rainfall intensity, with a weight of 0.2 for the heavy rain level. The CNN-LSTM model predicts that a pedestrian will enter the lane in 3 seconds, with a danger level of 1.0. It calculates the projected warning area, forming a 1.5m×1.5m red frame. The headlight brightness instantly increases by 50% to 9000 lux, with a response time of 80ms. The DLP projector module 13 projects the red warning frame, with a position offset to compensate for raindrop refraction errors. The voice module broadcasts the "pedestrian on the right" warning with a delay of ≤50ms. The power management chip 41 dynamically allocates power, prioritizing the power supply to the DLP projector module 13 and shutting down non-essential LED light source partitions, saving 15% energy.

[0053] Compared with the existing technology, this embodiment shortens the driver's reaction time in pedestrian crossing scenarios by 0.4s, the visibility distance of the projected warning sign on rainy days remains above 8m, and the system peak power consumption is within 180w.

[0054] Example 3 A vehicle is equipped with the DLP projection-based intelligent vehicle lighting system of the first embodiment.

[0055] With the above-described preferred embodiments of the present invention as a guide, and with reference to the above description, relevant personnel are fully capable of making various changes and modifications without departing from the technical scope of this invention. The technical scope of this invention is not limited to the contents of the specification and must be determined according to the scope of the claims.

Claims

1. An intelligent vehicle lighting system based on DLP projection, characterized in that: include: The optical system includes a light source module, an optical lens group, and a DLP projection module. The light source module uses a high-brightness LED or laser light source, and the DLP projection module integrates a high-resolution DMD chip. Intelligent control system, including sensor module, central processing unit, voice recognition module and communication module; An execution module, including a headlight drive circuit, a mechanical adjustment mechanism, and a projection drive circuit; Power management system, including power management chip and backup power supply; Among them, the sensor module collects environmental data in real time, the voice recognition module enters the listening state synchronously, the communication module receives remote control signals, and the central processing unit generates control instructions through multi-sensor data fusion and AI algorithm analysis. The execution module controls the car light drive circuit to adjust the brightness, controls the mechanical adjustment mechanism to adjust the illumination angle, and controls the DLP projection module to generate projection content according to the control instructions.

2. The DLP projection-based intelligent vehicle lighting system according to claim 1, characterized in that: The brightness of the light source module can be adjusted within a range of 5000-15000 lumens, and the steering angle accuracy of the mechanical adjustment mechanism is 0.5°.

3. The DLP projection-based intelligent vehicle lighting system according to claim 1, characterized in that: The sensor module includes a camera, a radar, a photosensor and a rain sensor; the communication module includes a 4G / 5G mobile communication unit, a Wi-Fi wireless unit and a Bluetooth unit.

4. The DLP projection-based intelligent vehicle lighting system according to claim 1, characterized in that: The resolution of the DMD chip is 1024×768, and the refresh frequency is 60 Hz.

5. A control method for a DLP projection-based intelligent vehicle lighting system according to any one of claims 1 to 4, characterized in that: The steps include: S1. Data Collection: The camera collects road images and identifies lane lines, pedestrians, or traffic signs; the radar monitors the distance, speed, and azimuth of obstacles ahead in real time; the light sensor detects ambient light intensity, and the rain sensor identifies rainfall intensity; The voice recognition module collects voice commands; the communication module receives remote control signals; S2. Preprocessing: Use median filtering and Gaussian filtering to eliminate image noise; synchronize the timestamps of each sensor data through the hardware clock; set the 3σ principle to eliminate sudden abnormal sensor data; S3. Feature processing: Use the YOLOv8 neural network to extract target contour, color, and motion vector features; analyze the radial velocity of obstacles through Fourier transform and predict the motion trajectory through Kalman filtering; fit the ambient brightness curve based on the light intensity data and generate weather labels based on the rainfall data; assign weights based on the sensor confidence level. The specific formula is as follows: in, is the sensor weight, It is sensor data; when the camera and radar detect the target position conflict, the radar data is used first and the secondary verification mechanism is triggered; S4, Decision Analysis: Train a CNN-LSTM hybrid model based on over 100,000 kilometers of actual road data; calculate and mark the possible dangerous areas that may be entered within the next 3 seconds based on the target coordinates and velocity vector; and establish a hazard level matrix; S5. Control strategy generation: Over 200 control rules are preset, and the optimal control sequence is generated using model predictive control. The lighting area of ​​the LED light source array is dynamically adjusted, arrow guidance is generated based on navigation data, and warning symbols are generated based on dangerous areas, with the projection distance controlled to be 5-15 meters in front of the vehicle. S6, execution feedback: transmit control instructions through the CAN bus, execute status feedback and projection effect feedback; S7, Algorithm optimization: Compare the instruction parameters with the actual execution results, use the reinforcement learning algorithm, set the reward function, and continuously optimize the control strategy.

6. The control method of the intelligent vehicle lighting system based on DLP projection according to claim 5, characterized in that: In step S1, the frame rate of the camera collecting road conditions is ≥30fps, the light intensity range of the ambient light detected by the photosensor is 0-100klx, and the sampling rate of the voice recognition module collecting voice commands is 16kHz.

7. The control method of the intelligent vehicle lighting system based on DLP projection according to claim 5, characterized in that: In step S3, the weights of the camera, radar, and light sensor are 0.4, 0.3, and 0.2, respectively.

8. The control method of the intelligent vehicle lighting system based on DLP projection according to claim 5, characterized in that: In step S4, the danger levels are specifically pedestrians>obstacles>curves>normal road conditions, and the priority coefficients are 1.0, 0.8, 0.6, and 0.3, respectively.

9. The control method of the intelligent vehicle lighting system based on DLP projection according to claim 5, characterized in that: The generation of the optimal control sequence in step S5 needs to take into account the vehicle lamp response delay, specifically: LED drive delay ≤ 50ms, DLP projection delay ≤ 30ms.

10. A vehicle, characterized in that: The intelligent vehicle lighting system based on DLP projection as claimed in any one of claims 1 to 4 is installed.

Images