Vehicle ambient light control method and device, electronic equipment and storage medium

Abstract

The application provides a vehicle-mounted atmosphere lamp control method and device, electronic equipment and storage medium, and relates to the technical field of automobiles. The method comprises the following steps: collecting and obtaining multi-modal input information, determining the corresponding safety level, scene priority and context correlation according to a preset priority mapping relationship; then determining the dynamic weight of each multi-modal input information, weighting and fusing the multi-modal input information to obtain fusion features, combining a preset atmosphere lamp generation model, outputting lamp effect semantic representation, and mapping the lamp effect semantic representation to structured lamp effect parameters according to a preset atmosphere lamp knowledge base; performing safety constraint adjustment to obtain safety lamp effect parameters to control the vehicle-mounted atmosphere lamp. The method realizes intelligent adaptive dynamic control of the display effect of the vehicle-mounted atmosphere lamp by collecting multi-modal input information and performing dynamic weight fusion and safety constraint adjustment, thereby improving the individualization degree and scene adaptation ability of the vehicle-mounted atmosphere lamp and reducing possible safety hazards.

Description

Technical Field

[0001] This application relates to the field of automotive technology, and in particular to a method, device, electronic device, and storage medium for controlling in-vehicle ambient lighting. Background Technology

[0002] As an important component of the smart cockpit, in-vehicle ambient lighting provides users with an immersive driving experience through changes in color, brightness, and dynamic effects. With the continuous improvement of automotive intelligence, users' demands for personalized, scenario-based, and intelligent ambient lighting are growing daily.

[0003] In existing in-vehicle ambient lighting control solutions, lighting effect switching is typically achieved through a preset library of fixed lighting effects or by triggering preset scenes based on a single sensor signal. For example, the system may preset a "Sports Mode" corresponding to a red breathing light effect, a "Soothing Mode" corresponding to a blue gradient light effect, or perform simple mapping within a preset color range based on vehicle speed. Some solutions allow users to manually adjust the color and brightness or select preset themes via the central control screen, or make limited selections from a preset color name library via voice commands.

[0004] However, the aforementioned in-vehicle ambient lighting control solutions suffer from fixed and highly homogenized lighting effects, failing to intelligently and dynamically adapt to multi-source information (such as vehicle operating status, environmental scenarios, user emotions, and entertainment content). Furthermore, existing solutions lack consideration for driving safety factors; high-brightness or high-frequency flashing lighting effects at night may cause driver glare, affecting driving safety. Therefore, existing in-vehicle ambient lighting control solutions exhibit low personalization, poor scene adaptability, and pose safety hazards. Summary of the Invention

[0005] The main purpose of this application is to propose a method, device, electronic device and storage medium for controlling in-vehicle ambient lighting, which aims to improve the personalization and scene adaptability of in-vehicle ambient lighting and reduce potential safety hazards.

[0006] In a first aspect, the present invention provides a method for controlling in-vehicle ambient lighting, applied to in-vehicle control equipment, the method comprising: Collect and acquire multimodal input information, including: vehicle operation information, environmental perception information, user status information, and entertainment information; Based on the preset priority mapping relationship, the security level, scene priority and context relevance of each of the multimodal input information are determined; Based on the security level, the scene priority, and the context relevance, the dynamic weights of each of the multimodal input information are determined, and the multimodal input information is weighted and fused according to the dynamic weights to obtain fused features; The fused features are input into a preset ambient light generation model to obtain the semantic representation of the lighting effect output by the preset ambient light generation model, and the semantic representation of the lighting effect is mapped into structured lighting effect parameters according to a preset ambient light knowledge base; wherein, the preset ambient light generation model is a machine learning model trained based on the preset ambient light knowledge base and historical lighting effect data. Based on the preset safety constraint model and the multimodal input information, the structured lighting effect parameters are adjusted for safety constraints to obtain safety lighting effect parameters, and the vehicle ambient light outputs the corresponding ambient lighting effect according to the safety lighting effect parameters.

[0007] In an optional implementation, determining the dynamic weights of each of the multimodal input information based on the security level, the scene priority, and the context relevance includes: The weighted calculation is performed on the security level, the scene priority, and the context relevance to obtain the weight score of each of the multimodal input information; The weight scores of each of the multimodal input information are normalized to obtain the dynamic weights of each of the multimodal input information.

[0008] In an optional implementation, before inputting the fused features into a preset ambient light generation model and obtaining the semantic representation of the lighting effect output by the preset ambient light generation model, the method further includes: Based on the multimodal input information and the preset conflict resolution rules, the conflicting multimodal input information is resolved to obtain the resolved multimodal input information. The step of determining the security level, scene priority, and context relevance of each of the multimodal input information according to a preset priority mapping relationship includes: Based on a preset priority mapping relationship, the security level, the scene priority, and the context relevance corresponding to the resolved multimodal input information are determined.

[0009] In an optional implementation, the multimodal input information further includes: emergency event input information; The step of resolving conflicting multimodal input information according to each of the multimodal input information and a preset conflict resolution rule to obtain resolved multimodal input information includes: If both high-security-level and low-security-level multimodal input information exist simultaneously, then the low-security-level multimodal input information is eliminated. If both the vehicle operation information and the entertainment information exist simultaneously, then the entertainment information is deleted; If there is emergency event input information, then the multimodal input information other than the emergency event input information is eliminated to obtain the eliminated multimodal input information.

[0010] In an optional implementation, the structured lighting effect parameters include: a color matrix, spatial mapping, time series, and dynamic change function; mapping the semantic representation of the lighting effect to structured lighting effect parameters according to a preset ambient lighting knowledge base includes: Based on the semantic representation of the lighting effect and the preset ambient lighting knowledge base, the color matrix, the spatial mapping, the time series, and the dynamic change function are generated.

[0011] In an optional implementation, the step of adjusting the structured lighting effect parameters according to the preset safety constraint model and the multimodal input information to obtain safe lighting effect parameters includes: Based on the preset safety constraint model, the vehicle operation information, and the environmental perception information, determine the current upper limit threshold for brightness and the upper limit threshold for frequency. The brightness parameter in the structured lighting effect parameters is adjusted according to the upper limit threshold of brightness, and the dynamic change frequency in the structured lighting effect parameters is adjusted according to the upper limit threshold of frequency, so as to obtain the safety lighting effect parameters.

[0012] In an optional implementation, the method further includes: Collect user feedback on the output ambient lighting effect, including satisfaction rating, usage time, and reuse status. A reward function is constructed based on the feedback information, and the preset ambient light generation model is optimized through reinforcement learning based on the reward function.

[0013] Secondly, the present invention provides a vehicle ambient lighting control device, comprising: The acquisition module is used to acquire multimodal input information, which includes: vehicle operation information, environmental perception information, user status information, and entertainment information. The determination module is used to determine the security level, scene priority, and context relevance corresponding to each of the multimodal input information according to a preset priority mapping relationship; determine the dynamic weight of each of the multimodal input information according to the security level, scene priority, and context relevance; and perform weighted fusion on each of the multimodal input information according to the dynamic weight to obtain fused features. The generation module is used to input the fused features into a preset ambient light generation model, obtain the lighting effect semantic representation output by the preset ambient light generation model, and map the lighting effect semantic representation into structured lighting effect parameters according to a preset ambient light knowledge base; wherein, the preset ambient light generation model is a machine learning model trained based on the preset ambient light knowledge base and historical lighting effect data. The control module is used to adjust the structured lighting effect parameters according to the preset safety constraint model and the multimodal input information to obtain safety lighting effect parameters, and control the vehicle ambient light to output the corresponding ambient light effect according to the safety lighting effect parameters.

[0014] Thirdly, the present invention provides an electronic device, comprising: a processor, a storage medium, and a bus, wherein the storage medium stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the storage medium via the bus, and the processor executes the machine-readable instructions to perform the steps of any of the methods described in the foregoing embodiments.

[0015] Fourthly, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the method as described in any of the foregoing embodiments.

[0016] The beneficial effects of this application are: The vehicle ambient lighting control method provided in this application includes: acquiring multimodal input information; determining the security level, scene priority, and context relevance of each multimodal input information according to a preset priority mapping relationship; determining the dynamic weight of each multimodal input information; weighting and fusing the multimodal input information accordingly to obtain fusion features; inputting the fusion features into a preset ambient lighting generation model to obtain the lighting effect semantic representation output by the preset ambient lighting generation model; mapping the lighting effect semantic representation into structured lighting effect parameters according to a preset ambient lighting knowledge base; adjusting the structured lighting effect parameters according to a preset safety constraint model and multimodal input information to obtain safety lighting effect parameters; and controlling the vehicle ambient lighting to output the corresponding ambient lighting effect according to the safety lighting effect parameters. In this embodiment, by collecting multimodal input information and performing dynamic weight fusion based on a preset priority mapping relationship, the fused feature input is used to generate a preset ambient light generation model trained based on a preset ambient light knowledge base and historical lighting effect data to generate a semantic representation of the lighting effect and map it into structured lighting effect parameters. Furthermore, the structured lighting effect parameters are adjusted for safety constraints according to a preset safety constraint model and multimodal input information. This achieves intelligent adaptive dynamic control of the display effect of the vehicle ambient light, thereby improving the personalization and scene adaptability of the vehicle ambient light and reducing potential safety hazards. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart illustrating a method for controlling in-vehicle ambient lighting according to an embodiment of this application; Figure 2 A flowchart illustrating a vehicle ambient lighting control method according to another embodiment of this application; Figure 3 This is a schematic diagram of the structure of a vehicle ambient lighting control device provided in an embodiment of this application; Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0020] Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0021] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0022] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0023] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0024] In existing in-vehicle ambient lighting control solutions, lighting effect switching is typically achieved through a preset library of fixed lighting effects or by triggering preset scenes based on a single sensor signal. For example, the system may preset a "Sports Mode" corresponding to a red breathing light effect, a "Soothing Mode" corresponding to a blue gradient light effect, or perform simple mapping within a preset color range based on vehicle speed. Some solutions allow users to manually adjust the color and brightness or select preset themes via the central control screen, or make limited selections from a preset color name library via voice commands.

[0025] However, the aforementioned in-vehicle ambient lighting control solutions suffer from fixed and highly homogenized lighting effects, failing to intelligently and dynamically adapt to multi-source information (such as vehicle operating status, environmental scenarios, user emotions, and entertainment content). Furthermore, existing solutions lack consideration for driving safety factors; high-brightness or high-frequency flashing lighting effects at night may cause driver glare, affecting driving safety. Therefore, existing in-vehicle ambient lighting control solutions exhibit low personalization, poor scene adaptability, and pose safety hazards.

[0026] To address the aforementioned issues, the main objective of this application is to propose a method for controlling in-vehicle ambient lighting, aiming to improve the personalization and scene adaptability of in-vehicle ambient lighting, and reduce potential safety hazards.

[0027] Figure 1 This is a flowchart illustrating a method for controlling in-vehicle ambient lighting according to an embodiment of this application. The executing entity of this method can be, for example, a vehicle infotainment system or other device with computing power, but is not limited thereto. Please refer to... Figure 1 The method includes: S101. Collect and acquire multimodal input information, including vehicle operation information, environmental perception information, user status information, and entertainment information.

[0028] For example, the aforementioned vehicle operation information may include data reflecting the current driving status of the vehicle, such as vehicle speed information, emergency braking information, steering information, and gear information. This vehicle operation information can be obtained from various sensors and electronic control units through the vehicle's internal bus.

[0029] The aforementioned environmental perception information may include, for example, the type of external environment (such as tunnels, urban roads, highways, seaside, etc.), lighting conditions, weather conditions, etc., and can be obtained through external cameras, photosensitive sensors, navigation maps, etc.

[0030] The aforementioned user status information may include, for example, user emotion information, fatigue status information, attention status information, etc., and can be obtained through driver monitoring systems, passenger monitoring systems, voice recognition systems, etc.

[0031] The aforementioned entertainment information may include, for example, the rhythm, style, and volume information of the currently playing music, and this entertainment information can be obtained through the in-vehicle entertainment system.

[0032] It is understandable that the collection frequency of the above multimodal input information can be set according to actual needs. For example, it can be collected once per second, or it can be triggered by specific events (such as emergency braking events, music switching events, etc.). No specific restrictions are imposed here.

[0033] S102. Based on the preset priority mapping relationship, determine the security level, scene priority and context relevance corresponding to each of the above multimodal input information.

[0034] For example, the aforementioned preset priority mapping relationship may refer to a correspondence table or mapping function between various types of input information and security level, scenario priority and context relevance, which is established in advance based on experimental measurement data, evaluation data, etc. However, the specific method and form of establishing the preset priority mapping relationship are not specifically limited here.

[0035] The aforementioned safety level can be used, for example, to characterize the degree of impact of the input information on driving safety. The higher the degree of impact, the higher the safety level can be. For example, the safety level of emergency braking information is higher than that of music style information.

[0036] The aforementioned scenario priority can be used to characterize the importance of the input information in the current driving scenario. For example, the scenario priority of driving information (such as vehicle operation information) is higher than that of entertainment information.

[0037] The aforementioned contextual relevance can be used to characterize the degree of correlation between the input information and the current driving scenario. For example, in a tunnel scenario, the contextual relevance of information about the tunnel in the environmental perception information can be high, while the contextual relevance of information in a seaside scenario can be low.

[0038] Based on the aforementioned preset priority mapping relationship, the aforementioned in-vehicle systems and other devices with computing capabilities can determine the security level, scene priority, and context relevance corresponding to each of the currently collected multimodal input information. To facilitate subsequent participation in related calculations, the security level, scene priority, and context relevance can be represented, for example, in the form of a score value. For instance, the security level can be represented in the form of a security level score, the scene priority can be represented in the form of a scene priority score, and the context relevance can be represented in the form of a relevance score, but this is not a limitation.

[0039] S103. Based on the above security level, the above scenario priority and the above context relevance, determine the dynamic weight of each of the above multimodal input information, and perform weighted fusion on each of the above multimodal input information according to the above dynamic weight to obtain fusion features.

[0040] For example, after determining the security level, scene priority, and context relevance corresponding to each of the aforementioned multimodal input information, the aforementioned in-vehicle infotainment system or other device with computing capabilities can calculate the dynamic weight of each of the aforementioned multimodal input information based on the values ​​of the above three dimensions, that is, based on the security level, scene priority, and context relevance expressed in the form of score values. The aforementioned dynamic weight can, for example, be used to describe the degree of influence of each of the aforementioned multimodal input information on the final lighting effect output in the current scene. The higher the security level, the higher the scene priority, or the higher the context relevance of the input information, the higher its corresponding dynamic weight can also be. However, the specific method of determining the dynamic weight is not limited here.

[0041] After calculating and obtaining the dynamic weights of each of the aforementioned multimodal input information, the aforementioned vehicle-mounted system and other devices with computing capabilities can perform weighted fusion of the feature vectors corresponding to each of the aforementioned multimodal input information according to the aforementioned dynamic weights. That is, the feature vectors of each input information are multiplied by their corresponding dynamic weights and then comprehensively processed to obtain the aforementioned fused features. The aforementioned fused features integrate multi-source input information and adaptively adjust the contribution ratio of each input information according to the current scene, providing a more accurate feature foundation for subsequent lighting effect generation.

[0042] S104. Input the aforementioned fused features into the preset ambient light generation model, obtain the semantic representation of the lighting effect output by the preset ambient light generation model, and map the semantic representation of the lighting effect into structured lighting effect parameters according to the preset ambient light knowledge base. The preset ambient light generation model is a machine learning model trained based on the preset ambient light knowledge base and historical lighting effect data.

[0043] For example, the aforementioned preset ambient lighting knowledge base may include knowledge of color matching principles, the impact of ambient lighting color and brightness on the eyes in dark environments, and examples of artificially matched ambient lighting effects. The aforementioned historical lighting effect data may be, for example, historically collected data on ambient lighting effects that satisfied users. After training with the aforementioned preset ambient lighting knowledge base and the aforementioned historical lighting effect data, the aforementioned preset ambient lighting generation model possesses the algorithmic logic to intelligently generate lighting effect semantic representations that conform to human aesthetics based on the input fusion features. The aforementioned lighting effect semantic representation may, for example, be a natural language description or semantic encoding of attributes such as lighting effect style, color tendency, and rhythm of change.

[0044] After obtaining the aforementioned semantic representation of lighting effects, devices with computing capabilities, such as vehicle-mounted systems, can map this semantic representation into structured lighting effect parameters based on the preset ambient lighting knowledge base. These structured lighting effect parameters can be, for example, a standardized set of parameters that quantifies various attributes of the lighting effect. This set is used to convert the abstract semantic representation of lighting effects into specific control parameters that the ambient lighting controller can recognize and execute, enabling the corresponding lighting effect to be precisely executed by the ambient lighting controller.

[0045] S105. Based on the preset safety constraint model and the above multimodal input information, the structured lighting effect parameters are adjusted for safety constraints to obtain safety lighting effect parameters, and the vehicle ambient light output is controlled to produce the corresponding ambient light effect based on the above safety lighting effect parameters.

[0046] For example, after obtaining the above-mentioned structured lighting effect parameters, the above-mentioned vehicle-mounted system and other devices with computing capabilities may not directly use the above-mentioned structured lighting effect parameters to control the vehicle ambient lighting. Instead, they may adjust the above-mentioned structured lighting effect parameters according to the preset safety constraint model and the above-mentioned multimodal input information to obtain the above-mentioned safety lighting effect parameters that meet the safety requirements.

[0047] The aforementioned preset safety constraint model may include, for example, safety constraints that ambient lighting effects should meet under different driving scenarios. These safety constraints may involve aspects such as the brightness range, rate of change, and flashing frequency of the ambient lighting. The aforementioned in-vehicle infotainment system or other devices with computing capabilities determine the current driving scenario based on the vehicle operation information and environmental perception information from the aforementioned multimodal input information. They then invoke the corresponding safety constraints from the aforementioned preset safety constraint model to adjust or restrict the relevant parameters in the aforementioned structured lighting effect parameters. This ensures that the adjusted safety lighting effect parameters retain the original lighting effect style and semantic expression while meeting the driving safety requirements of the current driving scenario. For example, in scenarios requiring high driver attention, such as at night or at high speeds, the aforementioned safety constraint adjustments may appropriately reduce the overall brightness of the lighting effect or slow down the rate of change of the lighting effect to avoid glare or distraction for the driver, but this is not the only possibility.

[0048] After obtaining the aforementioned safety lighting effect parameters, the aforementioned vehicle-mounted system and other devices with computing capabilities can generate corresponding control commands based on the aforementioned safety lighting effect parameters, and send the aforementioned control commands to the vehicle ambient lighting controller to control each ambient lighting unit to output the corresponding ambient lighting effect according to the aforementioned safety lighting effect parameters.

[0049] It should be noted that the specific methods of adjusting the above-mentioned safety constraints, the specific content of the above-mentioned safety constraints, and the specific form of the above-mentioned control commands can all be adjusted and determined according to the actual situation, and no specific restrictions are imposed here.

[0050] The vehicle ambient lighting control method provided in this application includes: acquiring multimodal input information; determining the safety level, scene priority, and context relevance of each multimodal input information according to a preset priority mapping relationship; determining the dynamic weight of each multimodal input information; weighting and fusing the multimodal input information to obtain a fusion feature; inputting the fusion feature into a preset ambient lighting generation model to obtain the lighting effect semantic representation output by the preset ambient lighting generation model; mapping the lighting effect semantic representation into structured lighting effect parameters according to a preset ambient lighting knowledge base; adjusting the structured lighting effect parameters according to a preset safety constraint model and the multimodal input information to obtain safety lighting effect parameters; and controlling the vehicle ambient lighting to output the corresponding ambient lighting effect according to the safety lighting effect parameters. In this embodiment, by collecting multimodal input information and performing dynamic weight fusion based on a preset priority mapping relationship, the fused feature input is used to generate a preset ambient light generation model trained based on a preset ambient light knowledge base and historical lighting effect data to generate a semantic representation of the lighting effect and map it into structured lighting effect parameters. Furthermore, the structured lighting effect parameters are adjusted for safety constraints according to a preset safety constraint model and multimodal input information. This achieves intelligent adaptive dynamic control of the display effect of the vehicle ambient light, thereby improving the personalization and scene adaptability of the vehicle ambient light and reducing potential safety hazards.

[0051] Furthermore, in the above Figure 1 Based on the embodiments, the dynamic weights of each of the above-mentioned multimodal input information are determined according to the security level, the scenario priority, and the contextual relevance, including: The weighted scores of the above-mentioned security levels, scenario priorities, and contextual relevance are calculated to obtain the weighted scores of each of the above-mentioned multimodal input information.

[0052] The weight scores of the aforementioned multimodal input information are normalized to obtain the dynamic weights of the aforementioned multimodal input information.

[0053] For example, after obtaining the security level, scene priority, and context relevance corresponding to each of the above-mentioned multimodal input information, the aforementioned in-vehicle system and other devices with computing capabilities may, for example, first perform weighted calculations on the values ​​of the above three dimensions to obtain the weight scores of each of the above-mentioned multimodal input information.

[0054] In one possible implementation, corresponding weighting coefficients (e.g., denoted as α, β, and γ) can be pre-set for the aforementioned security level, scenario priority, and context relevance, respectively. Then, the security level value, scenario priority value, and context relevance value corresponding to each of the aforementioned multimodal input information are multiplied by their respective weighting coefficients and summed to obtain the weight score for each of the aforementioned multimodal input information. The specific values ​​of the aforementioned weighting coefficients can be determined based on the actual application scenario and experimental calibration results. For example, when driving safety is the primary consideration, the weighting coefficient α corresponding to the security level can be set relatively high, the weighting coefficient β corresponding to the scenario priority can be second, and the weighting coefficient γ corresponding to the context relevance can be relatively low, but this is not a limitation.

[0055] After calculating the weight scores of each of the aforementioned multimodal input information, the aforementioned vehicle-mounted system or other device with computing capabilities normalizes the weight scores of each of the aforementioned multimodal input information to obtain the dynamic weights of each of the aforementioned multimodal input information. The purpose of the above normalization process is to make the sum of the dynamic weights of each of the aforementioned multimodal input information equal to 1, thereby facilitating subsequent weighted fusion calculations. In one possible implementation, the above normalization process can be implemented using the Softmax function, that is, using each of the aforementioned weight scores as input to the Softmax function and outputting the normalized dynamic weights. In another possible implementation, the above normalization process can also be implemented by dividing each weight score by the sum of all weight scores. Of course, the specific method of the above normalization process can be selected and determined according to actual needs, and no specific restrictions are imposed here.

[0056] Based on the examples above, assume that the multimodal input information is represented as the following set: X = {x1, x2, ..., x n}, Here, X can represent, for example, a set of multimodal input information, x1, x2, ..., x... n For example, it can represent various multimodal input information from different sources and with different semantics, such as vehicle operation information, environmental perception information, user status information, and entertainment information.

[0057] The aforementioned dynamic weights can be expressed, for example, as the following formula: W i = Softmax(α·S i + β·P i + γ·C i ), Among them, the aforementioned W i For the i-th multimodal input information x i The corresponding dynamic weights, the S mentioned above iP i C i Let x be the i-th multimodal input information. i The corresponding security level, scenario priority, and context relevance can be expressed in the form of a score, but are not limited to this.

[0058] Through the aforementioned weighted calculation and normalization process, the aforementioned in-vehicle systems and other devices with computing capabilities can transform abstract safety levels, scene priorities, and contextual relevance into quantifiable and comparable dynamic weights, providing an accurate weight basis for the subsequent weighted fusion of multimodal input information.

[0059] Furthermore, continuing with the examples above, the process of determining the dynamic weights of each of the multimodal input information based on the security level, scenario priority, and contextual relevance, and then weighting and fusing the multimodal input information according to these dynamic weights to obtain the fused features, can be represented by the following formula: V = Σ (W i · v i ), Wherein, V refers to the aforementioned fusion feature, and v refers to the aforementioned fusion feature. i Represents the i-th multimodal input information x i The corresponding feature vector is a vectorized representation obtained after feature extraction of the input information, used to characterize the key features of the input information. The above Σ (W i · v i ) represents the feature vector v corresponding to each multimodal input information. i Multiplied by its dynamic weight W i Then, all the weighted feature vectors are summed to obtain the final fused feature V.

[0060] Optionally, in the foregoing Figure 1 Based on the embodiments, before inputting the aforementioned fused features into the preset ambient light generation model and obtaining the semantic representation of the lighting effect output by the preset ambient light generation model, the method further includes: Based on the aforementioned multimodal input information and the preset conflict resolution rules, the conflicting multimodal input information is resolved to obtain the resolved multimodal input information.

[0061] The above-mentioned determination of the security level, scene priority, and context relevance corresponding to each of the aforementioned multimodal input information based on the preset priority mapping relationship includes: Based on the preset priority mapping relationship, the security level, scenario priority, and context relevance corresponding to the above-mentioned resolved multimodal input information are determined.

[0062] For example, in real-world applications, the lighting effect targets favored by the various multimodal input information collected simultaneously by devices with computing capabilities, such as in-vehicle systems, may be inconsistent or conflicting. For instance, emergency braking information in vehicle operation information tends to generate warning lighting effects, while soothing music information in entertainment information tends to generate calming lighting effects; the two have conflicting demands for lighting effects. If all conflicting multimodal input information is directly used for subsequent dynamic weight calculation and weighted fusion, it may lead to problems such as chaotic semantic representations or inconsistent styles in the generated lighting effects.

[0063] To address the aforementioned issues, the aforementioned in-vehicle infotainment system and other devices with computing capabilities can, before inputting the fused features into the preset ambient lighting generation model, first resolve conflicting multimodal input information based on each of the aforementioned multimodal input information and preset conflict resolution rules, thus obtaining resolved multimodal input information. The preset conflict resolution rules can, for example, be a pre-defined set of priority rules for resolving multi-input conflicts. The core idea could be, for example, to prioritize the retention of input information more important to driving safety or the current driving scenario, while resolving or suppressing relatively less important input information or input information that contradicts the current primary needs; however, the specific content and form are not limited here.

[0064] After completing the above-mentioned resolution process, the aforementioned in-vehicle infotainment system and other devices with computing capabilities no longer perform subsequent processing based on the original collected multimodal input information, but instead perform subsequent processing based on the resolved multimodal input information. Correspondingly, the determination of the security level, scene priority, and context relevance corresponding to each of the aforementioned multimodal input information according to a preset priority mapping relationship can specifically include: determining the security level, scene priority, and context relevance corresponding to the resolved multimodal input information according to the preset priority mapping relationship. Then, the aforementioned in-vehicle infotainment system and other devices with computing capabilities calculate dynamic weights and perform weighted fusion based on the security level, scene priority, and context relevance corresponding to the resolved multimodal input information to obtain the fused features, and input the fused features into the preset ambient light generation model to generate a semantic representation of the lighting effect.

[0065] By resolving the conflicts mentioned above, we can effectively avoid the problem of chaotic lighting effects caused by inconsistent goals of multiple input information sources, making the final ambient lighting effect more coordinated and unified, while also ensuring that driving safety-related input information plays a dominant role in lighting effect decision-making.

[0066] Furthermore, based on the above embodiments, the multimodal input information also includes: emergency event input information.

[0067] Based on the aforementioned multimodal input information and preset conflict resolution rules, the conflicting multimodal input information is resolved to obtain the resolved multimodal input information, including: If both high-security-level and low-security-level multimodal input information exist simultaneously, then the low-security-level multimodal input information is eliminated.

[0068] If both the aforementioned vehicle operation information and the aforementioned entertainment information exist simultaneously, then the aforementioned entertainment information will be deleted.

[0069] If there is emergency event input information, then the above multimodal input information other than the emergency event input information is cancelled to obtain the cancelled multimodal input information.

[0070] For example, the above resolution rules can be encapsulated as a resolution function. Then, the resolution process for the conflicting multimodal input information to obtain the resolved multimodal input information can be expressed as, for example, by the following formula: L = F_conflict(X'), Wherein, L is the multimodal input information after resolution, F_conflict() is the resolution function, and X' is a set containing conflicting multimodal input information.

[0071] In addition, Figure 1 Based on the embodiments, the above-mentioned structured lighting effect parameters include: color matrix, spatial mapping, time series, and dynamic change function. The mapping of the above-mentioned lighting effect semantic representation into structured lighting effect parameters according to a preset ambient lighting knowledge base includes: Based on the above semantic representation of lighting effects and the above preset ambient lighting knowledge base, the above color matrix, the above spatial mapping, the above time series, and the above dynamic change function are generated.

[0072] For example, after obtaining the lighting effect semantic representation output by the preset ambient lighting generation model, the aforementioned in-vehicle infotainment system or other device with computing power can convert the abstract semantic description into structured lighting effect parameters that the ambient lighting controller can directly recognize and execute, based on the aforementioned lighting effect semantic representation and the aforementioned preset ambient lighting knowledge base. The aforementioned structured lighting effect parameters may include, for example, color matrices, spatial mappings, time series, and dynamic change functions, but are not limited thereto. Specifically, the aforementioned color matrix can be used to define the color values ​​that each ambient light unit should display in the current lighting effect. In one possible implementation, the color matrix can be a multi-dimensional array, where each element corresponds to the color value of an ambient light unit or a group of ambient light units. These color values ​​can be represented in formats such as RGB (Red-Green-Blue) color space or HSV (Hue-Saturation-Value) color space. The aforementioned in-vehicle infotainment system or other device with computing power, based on the description of color tendencies in the aforementioned lighting effect semantic representation (e.g., "warm orange," "cool blue gradient," etc.), and combined with the color matching principles in the aforementioned preset ambient light knowledge base, assigns appropriate color values ​​to each ambient light unit, generating the aforementioned color matrix.

[0073] The aforementioned spatial mapping can, for example, be used to define the distribution of colors or lighting effects within the vehicle's interior space. In one possible implementation, the spatial mapping can describe the color transition relationships between ambient lighting units, the direction of lighting effect flow (e.g., from front to back, from left to right, etc.), or the emphasis rendering method for specific areas. The aforementioned in-vehicle infotainment system or other device with computing power generates the aforementioned spatial mapping based on the descriptions of spatial distribution or dynamic direction (e.g., "flow," "surround effect," etc.) in the aforementioned lighting effect semantic representation, combined with spatial layout knowledge and lighting effect instances from the aforementioned preset ambient lighting knowledge base.

[0074] The aforementioned time series can be used, for example, to define the time nodes of lighting effect changes, the duration of each stage, and the switching timing between different lighting effect stages. In one possible implementation, the aforementioned time series may include a series of timestamps and corresponding lighting effect status identifiers, used to control the phased changes of the ambient lighting effect over time. The aforementioned in-vehicle infotainment system or other devices with computing capabilities generate corresponding time series parameters based on the descriptions of change rhythm, cycle, or sequence (e.g., "slow fading," "rapid flashing," "periodic breathing," etc.) in the aforementioned lighting effect semantic representation.

[0075] The aforementioned dynamic change function can, for example, be used to define the gradation pattern of the lighting effect within a single change cycle. In one possible implementation, the dynamic change function can describe the mathematical curves of how parameters such as color and brightness change over time, such as linear gradation functions, sinusoidal gradation functions, exponential gradation functions, etc. The aforementioned in-vehicle infotainment system or other device with computational processing capabilities selects or generates corresponding dynamic change functions for each stage based on the description of the change pattern in the aforementioned lighting effect semantic representation (e.g., "gentle change," "rapid change," etc.) and in conjunction with the dynamic effect knowledge in the aforementioned preset ambient lighting knowledge base.

[0076] Based on this, the above-mentioned structured lighting effect parameters can be expressed as, for example, the following formula: E = {M_color, M_space, T_seq, F_dyn(t)}, Wherein, E is the structured lighting effect parameter, M_color is the color matrix, M_space is the spatial mapping, T_seq is the time series, and F_dyn(t) is the dynamic change function at time t.

[0077] It is understandable that the specific representations, generation algorithms, and coordination methods among the aforementioned color matrices, spatial mappings, time series, and dynamic change functions can all be adjusted and determined based on actual application requirements, ambient lighting hardware configurations, and computing resources, and no specific restrictions are imposed here. Through the generation of these structured lighting effect parameters, devices with computing capabilities, such as vehicle-mounted systems, can transform the semantic-level lighting effect descriptions output by large models into precise and controllable low-level control parameters, providing a standardized and quantifiable data foundation for subsequent safety constraint adjustments and lighting effect execution.

[0078] Figure 2 This application provides a flowchart illustrating the control process for in-vehicle ambient lighting, as shown in another embodiment. Figure 2 As shown, optionally, in the foregoing Figure 1 Based on the embodiments, the structured lighting effect parameters are adjusted according to the preset safety constraint model and the multimodal input information to obtain safe lighting effect parameters, including: S201. Based on the aforementioned preset safety constraint model, the aforementioned vehicle operation information, and the aforementioned environmental perception information, determine the current upper limit threshold for brightness and the upper limit threshold for frequency.

[0079] For example, after acquiring the structured lighting effect parameters, the aforementioned in-vehicle infotainment system and other devices with computing capabilities do not directly use the structured lighting effect parameters to control the in-vehicle ambient lighting. Instead, they determine the upper limit threshold of brightness and the upper limit threshold of frequency that should be applied in the current driving scenario based on the aforementioned preset safety constraint model, the aforementioned vehicle operation information in the aforementioned multimodal input information, and the aforementioned environmental perception information.

[0080] The aforementioned preset safety constraint model may, for example, pre-store the correspondence between different driving scenarios and safety constraint thresholds. The aforementioned driving scenarios may be determined by comprehensively judging, for example, based on vehicle speed information, emergency braking information, etc. in the aforementioned vehicle operation information, and lighting conditions, weather conditions, and external environment type, etc. in the aforementioned environmental perception information.

[0081] For example, when the vehicle operation information indicates a high current speed (e.g., exceeding 80 km / h) and the environmental perception information indicates a low-light environment such as nighttime or a tunnel, the current driving scenario can be determined to be one that demands high driver attention. In this case, a relatively low upper limit threshold for brightness and frequency should be selected from the preset safety constraint model to prevent high-brightness or high-frequency flashing lights from dazzling or distracting the driver. Conversely, when the vehicle is stationary or in a well-lit daytime scenario, the upper limit thresholds for brightness and frequency can be relaxed accordingly.

[0082] In one possible implementation, the aforementioned upper limit threshold for brightness and upper limit threshold for frequency can be expressed, for example, in the following form: B_max = f_B(V_env, D_veh), F_max = f_F(V_env, D_veh), Wherein, B_max is the brightness upper limit threshold, F_max is the frequency upper limit threshold, V_env is the feature parameter corresponding to the environmental perception information, D_veh is the feature parameter corresponding to the vehicle operation information, and f_B() and f_F() are the mapping functions used to calculate the brightness upper limit threshold and the frequency upper limit threshold in the preset safety constraint model, respectively. Of course, the above formula is only one possible representation; in actual implementation, other methods such as table lookup and rule matching can be used to determine the brightness upper limit threshold and the frequency upper limit threshold according to specific needs, and no specific restrictions are imposed here.

[0083] S202. Adjust the brightness parameter in the structured lighting effect parameters according to the brightness upper limit threshold, and adjust the dynamic change frequency in the structured lighting effect parameters according to the frequency upper limit threshold to obtain the safety lighting effect parameters.

[0084] For example, after determining the above-mentioned upper limit threshold for brightness and the above-mentioned upper limit threshold for frequency, the above-mentioned vehicle-mounted system or other device with computing power adjusts the parameters related to brightness and the parameters related to dynamic frequency in the above-mentioned structured lighting effect parameters to obtain the above-mentioned safety lighting effect parameters that meet safety requirements.

[0085] Specifically, for the brightness parameters in the aforementioned structured lighting effect parameters (e.g., the brightness components of each color value in the aforementioned color matrix, the overall brightness coefficient, etc.), the aforementioned in-vehicle infotainment system or other devices with computing capabilities compare the brightness parameters with the aforementioned upper limit threshold: if the brightness parameter is less than or equal to the upper limit threshold, the original brightness parameter can be retained unchanged. If the brightness parameter is greater than the upper limit threshold, the brightness parameter can be adjusted to the upper limit threshold or below. The adjustment can be done by scaling proportionally, that is, reducing each brightness parameter proportionally to meet the threshold requirement. Alternatively, it can be done by truncation, that is, directly limiting the portion exceeding the threshold to the threshold value. The specific adjustment method can be selected and determined according to actual needs.

[0086] For the dynamic change frequency in the aforementioned structured lighting effect parameters (e.g., the stage switching frequency in the aforementioned time series, the oscillation frequency in the aforementioned dynamic change function, etc.), the aforementioned in-vehicle infotainment system or other device with computing capabilities compares the aforementioned dynamic change frequency with the aforementioned upper frequency threshold: if the aforementioned dynamic change frequency is less than or equal to the aforementioned upper frequency threshold, the original dynamic change frequency can be retained unchanged. If the aforementioned dynamic change frequency is greater than the aforementioned upper frequency threshold, the aforementioned dynamic change frequency can be adjusted to the aforementioned upper frequency threshold or below.

[0087] In one possible implementation, the above-mentioned security constraint adjustment process can be expressed as, for example, the following optimization problem: E_safe = argmin ‖E' - E‖, The constraints are: Brightness(E') ≤ B_max, Frequency(E') ≤ F_max, Wherein, E represents the structured lighting effect parameter before adjustment, E' represents the candidate safety lighting effect parameter during the adjustment process, E_safe represents the final safety lighting effect parameter, Brightness(E') represents the brightness parameter of the candidate safety lighting effect parameter E', Frequency(E') represents the dynamic change frequency of the candidate safety lighting effect parameter E', and ‖E' - E‖ represents the degree of difference between the candidate safety lighting effect parameter E' and the original structured lighting effect parameter E. The meaning of the above optimization problem is: under the constraints of brightness not exceeding the above brightness upper limit threshold and dynamic change frequency not exceeding the above frequency upper limit threshold, find the safety lighting effect parameter with the smallest difference from the original structured lighting effect parameter, so as to preserve the style and semantic expression of the original lighting effect as much as possible while ensuring driving safety.

[0088] It is understandable that the specific methods for adjusting the aforementioned security constraints, the specific extraction and adjustment methods for the aforementioned brightness parameters and dynamic change frequencies, and the algorithm for solving the aforementioned optimization problem can all be determined based on actual application needs and computing resource allocation, and no specific restrictions are imposed here.

[0089] Through the design of this embodiment, the aforementioned in-vehicle systems and other devices with computing capabilities can, after generating structured lighting effect parameters, reasonably constrain the brightness and dynamic change frequency of the lighting effect according to the safety requirements of the current driving scenario, ensuring that the final output ambient lighting effect will not have a negative impact on driving safety, thus achieving a balance and unity between personalized lighting effect experience and driving safety assurance.

[0090] Optionally, based on any of the foregoing embodiments, the method may further include: Collect user feedback on the output ambient lighting effects, including satisfaction ratings, usage duration, and reuse rate.

[0091] A reward function is constructed based on the feedback information, and the preset ambient light generation model is optimized using reinforcement learning based on the reward function.

[0092] For example, after the vehicle's infotainment system or other devices with computing capabilities control the output of the ambient lighting effect, the system can further collect user feedback on the output ambient lighting effect to continuously optimize the preset ambient lighting generation model.

[0093] In one possible implementation, the feedback information may include, for example, the user's satisfaction rating for the ambient lighting effect, the actual usage duration of the ambient lighting effect, and whether the user actively reused or saved the ambient lighting effect. The satisfaction rating can be actively input by the user through voice interaction, touchscreen operation, or automatically assessed using emotion recognition technology based on the user's facial expressions and tone of voice. The usage duration may refer to the duration of the ambient lighting effect from its initial output until it is manually switched off or automatically terminated by the user. Reuse may refer to whether the user saves the ambient lighting effect as a frequently used theme or selects the ambient lighting effect again during subsequent driving, but is not limited to these limitations.

[0094] After collecting the aforementioned feedback information, the in-vehicle system and other devices with computing capabilities can construct a reward function based on this feedback. This reward function quantifies the overall user satisfaction with the ambient lighting effect output; a higher reward value indicates greater user satisfaction with the ambient lighting effect. In one possible implementation, the reward function can be expressed as, for example, the following formula: R = w1·S_sat + w2·T_use + w3·N_reuse, Wherein, R is the reward value corresponding to this ambient lighting effect, S_sat is the satisfaction score, T_use is the normalized usage time, N_reuse is the quantified value of the reuse status (e.g., 1 for saving or reusing, 0 for not saving or not reusing), and w1, w2, and w3 are the weight coefficients of the corresponding dimensions, which are preset values ​​used to balance the contribution of each feedback dimension to the final reward value.

[0095] After constructing the aforementioned reward function and calculating the reward value corresponding to the current ambient lighting effect, the aforementioned in-vehicle system and other devices with computing power can perform reinforcement learning optimization on the aforementioned preset ambient lighting generation model based on the reward function. For example, the aforementioned fused features can be used as states, the lighting effect semantic representation or structured lighting effect parameters generated by the aforementioned preset ambient lighting generation model can be used as actions, and the aforementioned reward value can be used as a feedback signal. The model parameters of the aforementioned preset ambient lighting generation model can be updated using a reinforcement learning algorithm, so that the aforementioned preset ambient lighting generation model tends to generate lighting effect semantic representations that can obtain higher reward values ​​(i.e., higher user satisfaction) in subsequent generation processes. The aforementioned reinforcement learning algorithm can be, for example, a policy gradient algorithm or a deep Q-network algorithm, but the specific choice can still be determined based on the actual model structure and computing resources.

[0096] Through the aforementioned reinforcement learning optimization mechanism based on user feedback, the preset ambient lighting generation model can continuously learn the user's personalized preferences during use, making the generated ambient lighting effects more in line with the user's actual aesthetic needs, and further improving the personalization of the in-vehicle ambient lighting and user satisfaction.

[0097] Figure 3 This is a schematic diagram of a vehicle ambient lighting control device according to an embodiment of this application. This device can execute the aforementioned vehicle ambient lighting control method and can be integrated into devices with computing capabilities, such as vehicle infotainment systems. Figure 3 As shown, the device may include: The acquisition module 310 is used to acquire multimodal input information, including vehicle operation information, environmental perception information, user status information, and entertainment information.

[0098] The determination module 320 is used to determine the security level, scene priority, and context relevance corresponding to each of the aforementioned multimodal input information according to a preset priority mapping relationship. Based on the aforementioned security level, scene priority, and context relevance, the dynamic weights of each of the aforementioned multimodal input information are determined, and the aforementioned multimodal input information is weighted and fused according to the aforementioned dynamic weights to obtain fused features.

[0099] The generation module 330 is used to input the aforementioned fused features into a preset ambient light generation model, obtain the semantic representation of the lighting effect output by the preset ambient light generation model, and map the semantic representation of the lighting effect into structured lighting effect parameters according to a preset ambient light knowledge base. The preset ambient light generation model is a machine learning model trained based on the preset ambient light knowledge base and historical lighting effect data.

[0100] The control module 340 is used to adjust the structured lighting effect parameters according to the preset safety constraint model and the multimodal input information to obtain the safety lighting effect parameters, and control the vehicle ambient light to output the corresponding ambient light effect according to the safety lighting effect parameters.

[0101] The vehicle ambient lighting control method provided in this application includes: acquiring multimodal input information; determining the safety level, scene priority, and context relevance of each multimodal input information according to a preset priority mapping relationship; determining the dynamic weight of each multimodal input information; weighting and fusing the multimodal input information to obtain a fusion feature; inputting the fusion feature into a preset ambient lighting generation model to obtain the lighting effect semantic representation output by the preset ambient lighting generation model; mapping the lighting effect semantic representation into structured lighting effect parameters according to a preset ambient lighting knowledge base; adjusting the structured lighting effect parameters according to a preset safety constraint model and the multimodal input information to obtain safety lighting effect parameters; and controlling the vehicle ambient lighting to output the corresponding ambient lighting effect according to the safety lighting effect parameters. In this embodiment, by collecting multimodal input information and performing dynamic weight fusion based on a preset priority mapping relationship, the fused feature input is used to generate a preset ambient light generation model trained based on a preset ambient light knowledge base and historical lighting effect data to generate a semantic representation of the lighting effect and map it into structured lighting effect parameters. Furthermore, the structured lighting effect parameters are adjusted for safety constraints according to a preset safety constraint model and multimodal input information. This achieves intelligent adaptive dynamic control of the display effect of the vehicle ambient light, thereby improving the personalization and scene adaptability of the vehicle ambient light and reducing potential safety hazards.

[0102] Optionally, the determining module 320 is specifically used to perform weighted calculations on the security level, the scenario priority, and the context relevance to obtain weight scores for each of the multimodal input information. The weight scores of each of the multimodal input information are then normalized to obtain the dynamic weights of each of the multimodal input information.

[0103] Optionally, the above-mentioned vehicle ambient lighting control device may further include: a resolution module, used to resolve conflicting multimodal input information according to each of the above-mentioned multimodal input information and a preset conflict resolution rule, to obtain resolved multimodal input information.

[0104] The aforementioned determining module 320 is specifically used to determine the aforementioned security level, the aforementioned scenario priority, and the aforementioned contextual relevance corresponding to the aforementioned resolved multimodal input information based on a preset priority mapping relationship.

[0105] Optionally, the aforementioned multimodal input information may also include: emergency event input information.

[0106] The aforementioned resolution module is specifically used to resolve the multimodal input information with lower safety levels if both high-level and low-level multimodal input information exist simultaneously. If both vehicle operation information and entertainment information exist simultaneously, the entertainment information is resolved. If emergency event input information exists, all multimodal input information except for the emergency event input information is resolved, resulting in the resolved multimodal input information.

[0107] Optionally, the above-mentioned structured lighting effect parameters include: color matrix, spatial mapping, time series, and dynamic change function.

[0108] The aforementioned generation module 330 is specifically used to generate the aforementioned color matrix, spatial mapping, time series, and dynamic change function based on the aforementioned lighting effect semantic representation and the aforementioned preset ambient light knowledge base.

[0109] Optionally, the control module 340 is specifically used to determine the current upper limit threshold for brightness and the upper limit threshold for frequency based on the preset safety constraint model, the vehicle operation information, and the environmental perception information. The brightness parameter in the structured lighting effect parameters is adjusted according to the upper limit threshold for brightness, and the dynamic change frequency in the structured lighting effect parameters is adjusted according to the upper limit threshold for frequency, to obtain the aforementioned safety lighting effect parameters.

[0110] Optionally, the aforementioned in-vehicle ambient lighting control device may further include: an optimization module, used to collect user feedback information on the output ambient lighting effect, including satisfaction rating, usage duration, and reuse status. A reward function is constructed based on the feedback information, and the preset ambient lighting generation model is optimized using reinforcement learning based on the reward function.

[0111] The above-described device is used to execute the method provided in the foregoing embodiments, and its implementation principle and technical effect are similar, so they will not be described again here.

[0112] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device can be a device with computing power, such as the aforementioned vehicle infotainment system. Figure 4 As shown, the device 400 includes: The processor 410, storage medium 420, and bus 430 are connected and communicate with each other via bus 430.

[0113] The storage medium 420 stores machine-readable instructions that can be executed by the processor 410. When the electronic device is running, the processor 410 executes the machine-readable instructions to perform the above-mentioned vehicle ambient lighting control method.

[0114] It should be understood that, Figure 4 The structure shown is only a schematic diagram of an electronic device; the electronic device may also include components that are larger than those shown. Figure 4 The more or fewer components shown, or having the same Figure 4 The different configurations shown. Figure 4 The components shown can be implemented using hardware, software, or a combination thereof.

[0115] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the vehicle ambient lighting control method described in the above method embodiments.

[0116] Computer-readable storage media can be electronic storage devices such as flash memory, EEPROM (Electrically Erasable Programmable Read-Only Memory), EPROM, hard disk, or ROM. Optionally, computer-readable storage media includes non-transitory computer-readable storage medium. The computer-readable storage medium has storage space for program code that performs any of the method steps described above. This program code can be read from or written to one or more computer program exhibits. The program code can be compressed, for example, in a suitable form.

[0117] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can also be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program exhibits according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0118] In addition, the functional modules in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0119] If the functionality is implemented as a software module and sold or used as an independent exhibit, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software exhibit. This computer software exhibit is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0120] The above description is merely a preferred embodiment of this application and does not limit the patent scope of this application. Any equivalent structural transformations made based on the inventive concept of this application and the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included within the patent protection scope of this application.

Claims

1. A method for controlling in-vehicle ambient lighting, characterized in that, Applied to in-vehicle control equipment, the method includes: Collect and acquire multimodal input information, including: vehicle operation information, environmental perception information, user status information, and entertainment information; Based on the preset priority mapping relationship, the security level, scene priority and context relevance of each of the multimodal input information are determined; Based on the security level, the scene priority, and the context relevance, the dynamic weights of each of the multimodal input information are determined, and the multimodal input information is weighted and fused according to the dynamic weights to obtain fused features; The fused features are input into a preset ambient light generation model to obtain the semantic representation of the lighting effect output by the preset ambient light generation model, and the semantic representation of the lighting effect is mapped into structured lighting effect parameters according to a preset ambient light knowledge base; wherein, the preset ambient light generation model is a machine learning model trained based on the preset ambient light knowledge base and historical lighting effect data. Based on the preset safety constraint model and the multimodal input information, the structured lighting effect parameters are adjusted for safety constraints to obtain safety lighting effect parameters, and the vehicle ambient light outputs the corresponding ambient lighting effect according to the safety lighting effect parameters.

2. The method according to claim 1, characterized in that, The step of determining the dynamic weights of each of the multimodal input information based on the security level, the scene priority, and the context relevance includes: The weighted calculation is performed on the security level, the scene priority, and the context relevance to obtain the weight score of each of the multimodal input information; The weight scores of each of the multimodal input information are normalized to obtain the dynamic weights of each of the multimodal input information.

3. The method according to claim 1, characterized in that, Before inputting the fused features into a preset ambient light generation model and obtaining the semantic representation of the lighting effect output by the preset ambient light generation model, the method further includes: Based on the multimodal input information and the preset conflict resolution rules, the conflicting multimodal input information is resolved to obtain the resolved multimodal input information. The step of determining the security level, scene priority, and context relevance of each of the multimodal input information according to a preset priority mapping relationship includes: Based on a preset priority mapping relationship, the security level, the scene priority, and the context relevance corresponding to the resolved multimodal input information are determined.

4. The method according to claim 3, characterized in that, The multimodal input information also includes: emergency event input information; The step of resolving conflicting multimodal input information according to each of the multimodal input information and a preset conflict resolution rule to obtain resolved multimodal input information includes: If both high-security-level and low-security-level multimodal input information exist simultaneously, then the low-security-level multimodal input information is eliminated. If both the vehicle operation information and the entertainment information exist simultaneously, then the entertainment information is deleted; If there is emergency event input information, then the multimodal input information other than the emergency event input information is eliminated to obtain the eliminated multimodal input information.

5. The method according to claim 1, characterized in that, The structured lighting effect parameters include: color matrix, spatial mapping, time series, and dynamic change function; mapping the semantic representation of the lighting effect into structured lighting effect parameters according to a preset ambient lighting knowledge base includes: Based on the semantic representation of the lighting effect and the preset ambient lighting knowledge base, the color matrix, the spatial mapping, the time series, and the dynamic change function are generated.

6. The method according to claim 1, characterized in that, The step of adjusting the structured lighting effect parameters according to the preset safety constraint model and the multimodal input information to obtain safe lighting effect parameters includes: Based on the preset safety constraint model, the vehicle operation information, and the environmental perception information, determine the current upper limit threshold for brightness and the upper limit threshold for frequency. The brightness parameter in the structured lighting effect parameters is adjusted according to the upper limit threshold of brightness, and the dynamic change frequency in the structured lighting effect parameters is adjusted according to the upper limit threshold of frequency, so as to obtain the safety lighting effect parameters.

7. The method according to any one of claims 1-6, characterized in that, The method further includes: Collect user feedback on the output ambient lighting effect, including satisfaction rating, usage time, and reuse status. A reward function is constructed based on the feedback information, and the preset ambient light generation model is optimized through reinforcement learning based on the reward function.

8. A vehicle ambient lighting control device, characterized in that, include: The acquisition module is used to acquire multimodal input information, which includes: vehicle operation information, environmental perception information, user status information, and entertainment information. The determination module is used to determine the security level, scene priority, and context relevance corresponding to each of the multimodal input information according to a preset priority mapping relationship; determine the dynamic weight of each of the multimodal input information according to the security level, scene priority, and context relevance; and perform weighted fusion on each of the multimodal input information according to the dynamic weight to obtain fused features. The generation module is used to input the fused features into a preset ambient light generation model, obtain the lighting effect semantic representation output by the preset ambient light generation model, and map the lighting effect semantic representation into structured lighting effect parameters according to a preset ambient light knowledge base; wherein, the preset ambient light generation model is a machine learning model trained based on the preset ambient light knowledge base and historical lighting effect data. The control module is used to adjust the structured lighting effect parameters according to the preset safety constraint model and the multimodal input information to obtain safety lighting effect parameters, and control the vehicle ambient light to output the corresponding ambient light effect according to the safety lighting effect parameters.

9. An electronic device, characterized in that, include: The device includes a processor, a storage medium, and a bus, wherein the storage medium stores machine-readable instructions executable by the processor, and when the electronic device is in operation, the processor communicates with the storage medium via the bus, and the processor executes the machine-readable instructions to perform the steps of the method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the steps of the method as described in any one of claims 1-7.

Images