CN-122028263-A - Vehicle light control method, device, equipment and medium

CN122028263ACN 122028263 ACN122028263 ACN 122028263ACN-122028263-A

Abstract

The application discloses a vehicle light control method which comprises the steps of obtaining driving behavior multi-mode data, carrying out state identification through a pre-trained deep learning model based on the driving behavior multi-mode data to obtain a state evaluation result of a driver, generating a light control strategy according to the state evaluation result, and controlling at least one optical parameter of a light system in a vehicle through the light control strategy. According to the application, the driving behavior multi-modal data are processed through fusion, and comprehensive analysis is performed by utilizing a pre-trained deep learning model, so that the state of the driver is judged, the physiological fatigue state and the psychological emotion state of the driver are evaluated, the light control strategy generated based on the evaluation result can be changed from passive environment illumination to active intervention, and the intervention is performed in time through scientifically verified optical parameter adjustment when the driver is in initial fatigue or in emotional tension, so that the fatigue and emotion soothing are effectively relieved, and the driving safety is improved.

Inventors

WANG QIAN
WANG JUNLIN
TAN CHUNYAN
LI NAN

Assignees

重庆赛力斯凤凰智创科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260323

Claims (10)

1. A vehicle light control method, characterized by comprising: Acquiring driving behavior multi-modal data; Based on the driving behavior multi-modal data, carrying out state identification through a pre-trained deep learning model to obtain a state evaluation result of a driver; Generating a lamplight control strategy according to the state evaluation result; and controlling at least one optical parameter of an in-vehicle light system through the light control strategy.
2. The vehicle light control method according to claim 1, characterized in that the state evaluation result includes a driving fatigue index, and the driving fatigue index generating method includes: extracting a plurality of first features characterizing a driver fatigue state from the driving behavior multimodal data; combining the plurality of first features to construct a multi-modal feature matrix; Calculating the multi-modal feature matrix based on an attention mechanism to generate dynamic weight coefficients representing each modal data; Weighting a plurality of features in the multi-mode feature matrix according to the dynamic weight coefficient to obtain a weighted fusion feature; and mapping the weighted fusion characteristics through a preset activation function to obtain a driving fatigue index for representing whether the driver is in a fatigue state.
3. The vehicle light control method according to claim 2, wherein the calculating the multi-modal feature matrix based on the attention mechanism generates dynamic weight coefficients representing each modal data, and the method comprises: according to the multi-mode feature matrix, calculating a query matrix, a key matrix and a value matrix; Calculating an attention score matrix based on the query matrix and the key matrix; Normalizing the attention score matrix to generate an attention weight matrix; multiplying the attention weight matrix with the value matrix to obtain an attention output matrix; The dynamic weighting coefficients are determined based on the attention output matrix.
4. The vehicle light control method according to claim 1, wherein the state evaluation result includes an emotion fluctuation index, and the generation method of the emotion fluctuation index includes: extracting a plurality of second features characterizing the emotional state of the driver from the driving behavior multimodal data; fusing the plurality of second features to obtain a comprehensive feature vector; And mapping the comprehensive feature vector through a preset classifier to obtain an emotion fluctuation index for representing whether the driver is in an emotion fluctuation state.
5. The vehicle light control method of claim 1, wherein the at least one optical parameter comprises a brightness and a color, and wherein the generating a light control strategy based on the state evaluation result comprises: In response to the driving fatigue index exceeding a preset driving fatigue threshold, executing a first control strategy in a light control strategy, wherein the first control strategy comprises the steps of reducing the brightness of an illumination area in a vehicle based on the driving fatigue index and adjusting the tone of the light in the vehicle to a preset warm tone range based on an environmental comfort evaluation value; and responding to the emotion fluctuation index exceeding a preset emotion fluctuation threshold, executing a second control strategy in the lamplight control strategies, wherein the second control strategy comprises the steps of reducing the brightness of an in-vehicle illumination area based on the driving fatigue index and adjusting the in-vehicle lamp light into soft warm-color light with gradual change effect based on the emotion fluctuation index.
6. The vehicle light control method according to claim 5, characterized in that the method for determining the environmental comfort evaluation value includes: Acquiring at least one in-vehicle environment data; Distributing preset weight coefficients for each type of in-vehicle environment data; and carrying out weighting processing on the corresponding at least one type of in-vehicle environment data according to the preset weight coefficient, and combining the weighted in-vehicle environment data to generate an environment comfort evaluation value.
7. The vehicle light control method of claim 1, further comprising iteratively optimizing a light control strategy using a reinforcement learning algorithm, the iterative optimization comprising: acquiring a current state from a state space based on the driving fatigue index, the emotion fluctuation index and the environmental comfort evaluation value, wherein the state in the state space is defined by the driving fatigue index, the emotion fluctuation index and the environmental comfort evaluation value; determining a light control strategy corresponding to the current state from an action space based on the current state, wherein the light control strategy in the action space is defined by a brightness value and a color value of light in a vehicle; generating a reward signal corresponding to the light control strategy according to feedback data of a driver acquired after the light control strategy is executed; Updating a value evaluation function of the light adjustment strategy based on the reward signal, a state transition result observed after the light control strategy is executed, a preset learning rate and a discount factor to optimize a decision mapping relation from a state to an action, wherein the discount factor is used for balancing weights of instant rewards and future rewards in value evaluation, and the learning rate is used for adjusting the influence degree of the update on historical value estimation.
8. A vehicle light control apparatus, characterized in that the vehicle light control includes: The data acquisition module is used for acquiring driving behavior multi-mode data; The state recognition module is used for carrying out state recognition through a pre-trained deep learning model based on the driving behavior multi-modal data to obtain a state evaluation result of a driver; the strategy generation module is used for generating a lamplight control strategy according to the state evaluation result; and the parameter control module is used for controlling at least one optical parameter of the in-vehicle lamp light system through the lamp light control strategy.
9. A vehicle light control apparatus, characterized by comprising: one or more processors; a memory for storing one or more programs that, when executed by the one or more processors, cause the vehicle light control apparatus to implement the steps of the vehicle light control method of any of claims 1 to 7.
10. A computer-readable storage medium, having stored thereon a computer program which, when executed by a processor of a computer, causes the computer to perform the steps of the vehicle light control method of any one of claims 1 to 7.

Description

Vehicle light control method, device, equipment and medium Technical Field The application relates to the technical field of light control, in particular to a vehicle light control method, device, equipment and medium. Background At present, along with the improvement of the intelligent degree of the automobile, the comfort and the safety of the environment in the automobile gradually become research hot spots. In-vehicle lights are an important component of the driving environment, whose optical characteristics directly affect the visual experience and emotional state of the driver and passengers. However, the current mainstream in-vehicle light adjustment schemes rely on a preset fixed mode or manual operation of a user, and are static or passive in nature, and lack intelligence for autonomous adjustment according to real-time physical and psychological states of a driver and dynamic environmental parameters in the vehicle. For example, when the driver is in a fatigue state caused by long-time driving, the brightness and color of the in-vehicle light may not provide effective comfort and safety support. The root cause is that the vehicle light control module and the driver state monitoring system are usually independent of each other, forming a technical barrier between perception and execution, resulting in data flow break and decision isolation. Due to the lack of the linkage mechanism, the system is difficult to realize closed-loop optimization from environment information to personalized light environment output, and further improvement of the driving experience intelligence level is severely restricted. Disclosure of Invention In view of the above-mentioned drawbacks of the prior art, the present application provides a vehicle light control method, device, apparatus and medium, so as to solve the above-mentioned technical problems. The application provides a vehicle light control method, which comprises the following steps: Acquiring driving behavior multi-modal data; Based on the driving behavior multi-modal data, carrying out state identification through a pre-trained deep learning model to obtain a state evaluation result of a driver; Generating a lamplight control strategy according to the state evaluation result; and controlling at least one optical parameter of an in-vehicle light system through the light control strategy. In an embodiment of the present application, the state evaluation result includes a driving fatigue index, and the method for generating the driving fatigue index includes: extracting a plurality of first features characterizing a driver fatigue state from the driving behavior multimodal data; combining the plurality of first features to construct a multi-modal feature matrix; Calculating the multi-modal feature matrix based on an attention mechanism to generate dynamic weight coefficients representing each modal data; Weighting a plurality of features in the multi-mode feature matrix according to the dynamic weight coefficient to obtain a weighted fusion feature; and mapping the weighted fusion characteristics through a preset activation function to obtain a driving fatigue index for representing whether the driver is in a fatigue state. In an embodiment of the present application, the calculating the multi-modal feature matrix based on the attention mechanism generates a dynamic weight coefficient representing each modal data, including: according to the multi-mode feature matrix, calculating a query matrix, a key matrix and a value matrix; Calculating an attention score matrix based on the query matrix and the key matrix; Normalizing the attention score matrix to generate an attention weight matrix; multiplying the attention weight matrix with the value matrix to obtain an attention output matrix; The dynamic weighting coefficients are determined based on the attention output matrix. In an embodiment of the present application, the state evaluation result includes an emotion fluctuation index, and the method for generating the emotion fluctuation index includes: extracting a plurality of second features characterizing the emotional state of the driver from the driving behavior multimodal data; fusing the plurality of second features to obtain a comprehensive feature vector; And mapping the comprehensive feature vector through a preset classifier to obtain an emotion fluctuation index for representing whether the driver is in an emotion fluctuation state. In an embodiment of the present application, the at least one optical parameter includes brightness and color, and the generating a light control strategy according to the state evaluation result includes: In response to the driving fatigue index exceeding a preset driving fatigue threshold, executing a first control strategy in a light control strategy, wherein the first control strategy comprises the steps of reducing the brightness of an illumination area in a vehicle based on the driving fatigue index and adjusting the