Adaptive control methods and systems for mood lighting, electronic devices and media

Abstract

This application provides an adaptive control method and system for mood lighting, an electronic device, and a medium. The method includes generating a mood lighting parameter vector based on a mood lighting preference model and controlling the lighting device to output a target light environment; acquiring behavioral feedback events generated by the user in response to the target light environment; performing attribution analysis on the correlation between the behavioral feedback events and the parameter configuration of the target light environment to obtain feedback attribution weights; converting the behavioral feedback events into lighting parameter correction vectors and calculating a fit score; incrementally updating the mood lighting preference model under preset constraints based on the feedback attribution weights, the lighting parameter correction vectors, and the fit score, and using the updated model for subsequent generation of mood lighting parameter vectors. This application solves the technical problem in existing mood lighting systems where the lack of refined analysis of user feedback leads to insufficient adaptive control capabilities and difficulty in dynamically matching users' personalized mood preferences.

Description

Technical Field

[0001] This application belongs to the field of intelligent lighting technology, and relates to an adaptive control method and system for mood lighting, electronic devices and media. Background Technology

[0002] With the development of intelligent lighting technology, lighting systems have gradually evolved from traditional switch control, brightness adjustment, and color temperature adjustment to emotional lighting systems that can generate multi-dimensional light environments by combining user emotional state, scene activities, and environmental conditions.

[0003] However, in real-world usage scenarios, there is no fixed correlation between mood lighting parameters and users' subjective acceptance. On the one hand, different users may have significantly different evaluations of comfort, relaxation, or pleasure in the same mood lighting environment; on the other hand, the same user's preference for the same combination of lighting parameters may change at different times, in different rooms, under different activity states, and at different stages of mood. Therefore, if a mood lighting system relies solely on preset rules, general models, or one-time configured mood-light mapping relationships, it is difficult to continuously adjust the lighting strategy based on real user feedback over long-term use, and it cannot achieve an adaptive control effect that gradually conforms to user preferences as usage progresses.

[0004] Furthermore, while existing smart lighting solutions possess the ability to automatically control lighting based on ambient illuminance, human presence sensing, sleep patterns, or preset scenarios, and some solutions attempt to optimize lighting strategies using user behavior data, they generally suffer from the following limitations: these solutions typically treat user feedback merely as general control logs or macroscopic statistical data, lacking refined semantic expression and correlation analysis specific to the dimensions of mood-related lighting parameters. This coarse-grained data processing approach prevents the system from deeply understanding the intrinsic relationship between user feedback and specific lighting environment parameters, thus limiting the accuracy and intelligence level of mood-related adaptive lighting control. Summary of the Invention

[0005] This application provides an adaptive control method and system for mood lighting, an electronic device, and a medium to solve the technical problem in existing mood lighting systems that lack refined analysis of user feedback, resulting in insufficient adaptive control capabilities and difficulty in dynamically matching users' personalized emotional preferences.

[0006] In a first aspect, this application provides an adaptive control method for mood lighting, comprising: generating a mood lighting parameter vector based on a mood lighting preference model, and controlling a lighting device to output a target light environment based on the mood lighting parameter vector; acquiring behavioral feedback events generated by a user in response to the target light environment; the behavioral feedback events including one or more of manual dimming behavior, scene switching behavior, scene rollback behavior, automatic mode interruption behavior, continuous use behavior, and repeated selection behavior; performing attribution analysis on the correlation between the behavioral feedback events and the parameter configuration of the target light environment to obtain feedback attribution weights; the feedback attribution weights are used to characterize the relationship between the behavioral feedback events and the parameters. The degree of correlation between configurations; converting the behavioral feedback event into a lighting parameter correction vector of the same dimension as the emotional lighting parameter vector, and calculating a suitability score based on the behavioral feedback event; the lighting parameter correction vector is used to characterize the correction direction and magnitude of the behavioral feedback event on each parameter dimension in the emotional lighting parameter vector; the suitability score is used to characterize the user's acceptance of the target lighting environment; based on the feedback attribution weight, the lighting parameter correction vector, and the suitability score, the emotional lighting preference model is incrementally updated under preset constraints, and the updated emotional lighting preference model is used for subsequent emotional lighting parameter vector generation.

[0007] In one implementation of the first aspect, controlling the output of a target light environment by a lighting device based on the emotional lighting parameter vector includes: acquiring the capability information of the lighting device, and constructing a lighting device capability matrix based on the capability information; the lighting device capability matrix is ​​a data structure that parametrically expresses the controllable output capabilities of the lighting device, the controllable output capabilities including luminous flux range, color temperature range, color gamut range, channel structure, dimming depth, spatial coverage area, minimum response time, and maximum rate of change; based on the lighting device capability matrix, under the condition of satisfying preset safe lighting boundaries and device output constraints, constraining and solving the emotional lighting parameter vector to obtain the channel control values ​​of the lighting device; and driving the corresponding lighting device to output the target light environment based on the channel control values.

[0008] In one implementation of the first aspect, calculating the fit score based on the behavioral feedback event includes: acquiring indicators associated with the target light environment within a preset statistical window; the indicators include usage duration, manual intervention frequency, scene rollback probability, and number of automatic mode interruptions; and performing normalized weighted calculations on the usage duration, manual intervention frequency, scene rollback probability, and number of automatic mode interruptions to obtain the fit score, wherein the usage duration corresponds to a positive weight, and the manual intervention frequency, scene rollback probability, and number of automatic mode interruptions correspond to negative weights.

[0009] In one implementation of the first aspect, incrementally updating the emotional lighting preference model under preset constraints based on the feedback attribution weights, the lighting parameter correction vector, and the fit score includes: adding the emotional lighting parameter vector to the lighting parameter correction vector to obtain the user expectation parameter vector; constructing a loss function using the deviation between the emotional lighting parameter vector output by the emotional lighting preference model and the user expectation parameter vector as the loss term; normalizing the fit score to obtain update weights; and updating the parameters of the emotional lighting preference model based on the feedback attribution weights, the loss function, and the update weights.

[0010] In one implementation of the first aspect, the method further includes updating the emotional lighting preference model using a confidence gating mechanism. The confidence gating mechanism includes: obtaining the feedback confidence of the behavioral feedback event, whereby the feedback confidence is used to characterize the reliability of the behavioral feedback event as a model update sample; when at least one of the feedback attribution weight, the fit score, and the feedback confidence is lower than a corresponding preset threshold, only the corresponding behavioral feedback event is recorded as a feedback sample, without immediately updating the emotional lighting preference model; when similar feedback samples repeatedly appear within a preset time range, and the cumulative number reaches a preset cumulative threshold, the emotional lighting preference model is updated based on the similar feedback samples; the similar feedback samples refer to feedback samples with the same feedback type, the same corresponding light parameter dimension, and a scene context that satisfies a preset similarity condition; when the number of feedback samples is lower than a preset sample threshold, within the safe lighting boundary, an exploration perturbation not exceeding a preset perturbation amplitude is applied to at least one of the light parameter dimensions (brightness, color temperature, color, dynamic speed, or transition time), and the user preference gradient in the emotional lighting preference model is estimated based on the change in the fit score before and after the exploration perturbation is applied.

[0011] In one implementation of the first aspect, the emotional lighting preference model includes an individual-level emotional lighting preference model and a family-level emotional lighting collaboration model; it also includes: when multiple family members are in a shared lighting space or the same lighting scene, constructing a family member conflict graph based on the individual-level emotional lighting preference model of each family member; the family member conflict graph uses family members as nodes and uses the preference differences, spatial overlap, activity priority, and safety priority among family members as edge weights; based on the family member conflict graph, generating a shared area lighting parameter vector through the family-level emotional lighting collaboration model; when the solved shared area lighting parameter vector cannot simultaneously satisfy the preference constraints of multiple family members, dividing the shared lighting space into a shared light field and at least one local light field, and generating lighting parameters for the shared light field and the local light field respectively.

[0012] In one implementation of the first aspect, converting the behavioral feedback event into a lighting parameter correction vector of the same dimension as the emotional lighting parameter vector includes: when the behavioral feedback event is the manual dimming behavior, parsing the manual dimming behavior to determine at least one operation content among brightness increase / decrease, color temperature increase / decrease, color shift, saturation change, dynamic speed adjustment, or dynamic amplitude adjustment; mapping the operation content into a lighting parameter correction vector of the same dimension as the emotional lighting parameter vector according to a preset operation and parameter mapping table; when the behavioral feedback event is the scene switching behavior, the scene rewind behavior, or the automatic mode interruption behavior, generating a negative correction amount or suppression weight of the same dimension as the emotional lighting parameter vector; when the behavioral feedback event is the continuous use behavior or the repeated selection behavior, generating a positive correction amount or reinforcement weight of the same dimension as the emotional lighting parameter vector.

[0013] Secondly, this application provides an adaptive control system for mood lighting, comprising: a mood lighting module, used to generate a mood lighting parameter vector based on a mood lighting preference model, and control a lighting device to output a target light environment based on the mood lighting parameter vector; a feedback acquisition module, used to acquire behavioral feedback events generated by a user in response to the target light environment; the behavioral feedback events include one or more of the following: manual dimming behavior, scene switching behavior, scene rewind behavior, automatic mode interruption behavior, continuous use behavior, and repeated selection behavior; and an attribution analysis module, used to perform attribution analysis on the correlation between the behavioral feedback events and the parameter configuration of the target light environment, and obtain feedback attribution weights; the feedback attribution weights are used to characterize the behavioral feedback events. The degree of correlation between the behavioral feedback event and the parameter configuration; the adaptation scoring module, used to convert the behavioral feedback event into a lighting parameter correction vector of the same dimension as the emotional lighting parameter vector, and calculate the adaptation score based on the behavioral feedback event; the lighting parameter correction vector is used to characterize the direction and magnitude of the correction of each parameter dimension in the emotional lighting parameter vector by the behavioral feedback event; the adaptation score is used to characterize the user's acceptance of the target lighting environment; the model update module, used to incrementally update the emotional lighting preference model under preset constraints based on the feedback attribution weight, the lighting parameter correction vector and the adaptation score, and use the updated emotional lighting preference model for subsequent emotional lighting parameter vector generation.

[0014] Thirdly, this application provides an electronic device, comprising: a memory for storing a computer program; and a processor for executing the computer program stored in the memory to cause the electronic device to perform the method described in any of the preceding claims.

[0015] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in any of the preceding claims.

[0016] As described above, the mood lighting adaptive control method and system, electronic device and medium described in this application have the following beneficial effects:

[0017] (1) It can convert the implicit and explicit behavioral feedback of users in real family emotional lighting scenarios into calculable and traceable light parameter learning signals, thereby improving the long-term adaptive capability of emotional lighting parameters and the collaborative control effect of multi-member families.

[0018] (2) By constructing a lighting equipment capability matrix, it is possible to ensure that the target light environment is stably presented while meeting the physical capabilities and safety boundaries of the equipment;

[0019] (3) The confidence gating mechanism improves the stability and reliability of model updates and avoids erroneous shifts;

[0020] (4) It can take into account both the consistency of the basic lighting environment of the shared space and the differences in individual preferences. Attached Figure Description

[0021] Figure 1 The diagram shown is a deployment architecture diagram of an emotion lighting adaptive control system according to an embodiment of this application.

[0022] Figure 2 The flowchart shown is a process for an adaptive control method for mood lighting in one embodiment of this application.

[0023] Figure 3 The diagram shown is an output flowchart of the target light environment in one embodiment of this application.

[0024] Figure 4 The flowchart shown is a process for converting behavioral feedback events into lighting parameter correction vectors in one embodiment of this application.

[0025] Figure 5 The diagram shown is a flowchart of the fit score calculation in one embodiment of this application.

[0026] Figure 6 The flowchart shown is an update flowchart of the mood lighting preference model in one embodiment of this application.

[0027] Figure 7 The diagram shown illustrates the working principle of the confidence gating mechanism in one embodiment of this application.

[0028] Figure 8 The flowchart shown is a process for an adaptive control method for mood lighting according to another embodiment of this application.

[0029] Figure 9 The diagram shown is a structural schematic of an adaptive lighting control system for mood according to an embodiment of this application.

[0030] Figure 10 The diagram shown is a structural schematic of an electronic device according to an embodiment of this application. Detailed Implementation

[0031] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, unless otherwise specified, the following embodiments and features in the embodiments can be combined with each other.

[0032] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. Therefore, the drawings only show the components related to this application and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0033] The following embodiments of this application provide an adaptive control method and system for mood lighting, an electronic device, and a medium. By establishing a refined correlation mechanism between user behavior feedback and mood lighting parameters, continuous learning and adaptive optimization of the light environment strategy are achieved. This method is widely applicable to mood lighting application scenarios such as homes, hotels, elderly care spaces, commercial spaces, and other scenarios that require long-term personalized light environment adjustment.

[0034] Please see Figure 1 The diagram shows the deployment architecture of an adaptive lighting control system for mood lighting in one embodiment of this application.

[0035] like Figure 1 As shown, this application can be flexibly deployed on one or more devices, including local devices, home gateways, edge computing nodes, or cloud servers, or it can adopt a distributed deployment architecture where the above devices work together. In specific implementation, a cloud-edge-device collaborative strategy can be adopted: the local device or edge computing node performs real-time lighting control to ensure the immediate responsiveness of the user experience; the home gateway or cloud server undertakes computationally intensive tasks such as model management, user preference updates, and long-term behavioral data analysis. Through this collaborative deployment method, this application can not only ensure the real-time performance and stability of lighting control, but also continuously mine the value of user data during long-term use, and continuously improve the personalized adaptive capabilities and intelligence level of the mood lighting system.

[0036] The following will describe in detail the principle and implementation of an adaptive control method and system for mood lighting, electronic equipment and medium of this embodiment, so that those skilled in the art can understand the adaptive control method and system for mood lighting, electronic equipment and medium of this embodiment without creative effort.

[0037] Please see Figure 2 The above is a flowchart of an adaptive control method for mood lighting in one embodiment of this application.

[0038] like Figure 2 As shown, this embodiment provides an adaptive control method for mood lighting, including the following steps S100 to S500.

[0039] In step S100, an emotional lighting parameter vector is generated based on the emotional lighting preference model, and the lighting device is controlled to output the target light environment based on the emotional lighting parameter vector.

[0040] In this embodiment, the emotional lighting preference model may include an individual-level emotional lighting preference model and a family-level emotional lighting synergy model.

[0041] The individual-level emotion lighting preference model is used to characterize the personalized mapping relationship between an individual user and emotion lighting parameters under different emotional states, different scene contexts, and different time periods. This model enables the formation of long-term stable emotion lighting preferences for specific users and generates emotion lighting parameter vectors that approximate these preferences.

[0042] The family-level mood lighting coordination model is used to coordinate the individual lighting preferences of different family members when multiple family members are in a shared lighting space or the same lighting scene, and to generate shared area lighting parameters, or generate combined parameters of shared light field and local light field.

[0043] In one embodiment of this application, generating an emotional lighting parameter vector based on an emotional lighting preference model includes generating an emotional lighting parameter vector that matches the user's current emotional state, usage scenario, and long-term preferences based on the user's emotional state parameters, scene context parameters, and historical behavioral feedback.

[0044] The emotional state parameters can be obtained from mobile terminals, cameras, voice interaction, wearable devices, wall panels, historical behavior logs, or the fusion of the above-mentioned multi-source data, and at least include emotion type, emotion intensity, emotion duration, and emotion change trend.

[0045] The scene context parameters include at least one or more of the following: room type, time period, activity type, ambient illuminance, user location, current lighting scene identifier, and current lighting parameter version number.

[0046] The mood lighting parameter vector is used to characterize the target configuration of the target light environment in multiple lighting parameter dimensions, including at least one or more of the following: illuminance or brightness, correlated color temperature, chromaticity coordinates, saturation, shortwave or target spectral proportion, dynamic amplitude, dynamic change rate, transition time, and spatial output weight.

[0047] Specifically, the emotion lighting parameter vector can be represented as:

[0048] P_t=[E_lux,CCT,x,y,S,ρ_s,A_d,V_d,T_g,W_s],

[0049] Where P_t represents the mood lighting parameter vector, E_lux represents illuminance or brightness, CCT represents correlated color temperature, x and y represent chromaticity coordinates, S represents saturation, ρ_s represents the proportion of shortwave or target spectrum, A_d represents the dynamic change amplitude, V_d represents the dynamic change rate, T_g represents the transition time, and W_s represents the spatial output weight.

[0050] It should be noted that the above parameters are only examples and can be expanded, tailored or replaced according to the type of lighting equipment, control precision, application scenario and safety lighting requirements in actual applications.

[0051] Please see Figure 3 The output flowchart of the target light environment is shown in one embodiment of this application.

[0052] like Figure 3 As shown, controlling the output target light environment of the lighting device based on the emotion lighting parameter vector includes the following steps S101 to S103.

[0053] In step S101, the capability information of the lighting device is obtained, and a lighting device capability matrix is ​​constructed based on the capability information of the lighting device.

[0054] The lighting equipment capability matrix is ​​a data structure that parametrically expresses the controllable output capabilities of lighting equipment. These controllable output capabilities include luminous flux range, color temperature range, color gamut range, channel structure, dimming depth, spatial coverage area, minimum response time, and maximum rate of change. When controlling the output of the target lighting environment, the system can allocate lighting parameters according to the actual output capabilities of each lighting device, avoiding the generation of control commands that exceed the device's capability range or do not conform to safe lighting boundaries.

[0055] In step S102, based on the lighting equipment capability matrix, under the condition of satisfying the preset safety lighting boundary and equipment output constraints, the emotional lighting parameter vector is constrained and solved to obtain the channel control value of the lighting equipment.

[0056] In one embodiment of this application, the following objective function can be constructed and solved:

[0057] J(U)=||G_dU-P_t||^2+λ||U-U_{t-1}||^2+μR_safe(U);

[0058] Where U represents the channel control value at time t, i.e. the channel control value to be solved; G_d represents the device capability matrix; P_t represents the emotional lighting parameter vector; U_{t-1} represents the channel control value at time t-1 (i.e. the previous time); R_safe represents the safe lighting boundary penalty term; λ and μ represent weighting coefficients; the safe lighting boundary includes one or more of the following: upper limit of illuminance, glare constraint, shortwave ratio constraint, color temperature range constraint, dynamic flicker frequency constraint, dynamic change rate constraint, and stimulation threshold constraint in child or elderly mode.

[0059] By solving the above objective function, we can approximate the emotional lighting parameter vector as closely as possible while ensuring that the obtained channel control values ​​meet the requirements of the lighting equipment's physical output capability, safe lighting boundary, and continuous output smoothness.

[0060] In step S103, based on the channel control value, the corresponding lighting device is driven to output the target light environment.

[0061] In this embodiment, the preference learning results corresponding to the user's emotional state, scene context, and historical behavior feedback can be actually implemented as the controllable output of the lighting device, so that the target light environment can be stably presented while meeting the device's output capabilities and safe lighting boundaries.

[0062] For example, when a user is stressed or fatigued and is in a bedroom, the mood lighting preference model can generate a combination of parameters with a low dynamic change rate, a long transition time, low saturation, and moderate low illuminance. After constraint solving, the system can control the main bedroom light, bedside lamp, or light strip to output a soft and stable target light environment.

[0063] In step S200, the user's behavioral feedback events in response to the target light environment are acquired.

[0064] Specifically, during or after the output of the target lighting environment, the system collects behavioral feedback events generated by the user in response to the current target lighting environment. These behavioral feedback events include one or more of the following: manual dimming behavior, scene switching behavior, scene rollback behavior, automatic mode interruption behavior, continuous use behavior, and repeated selection behavior.

[0065] The manual dimming behavior can include the user manually adjusting one or more of the following parameters: brightness, color temperature, color, saturation, dynamic change speed, dynamic change amplitude, or transition time, via a mobile terminal application, wall panel, voice interaction, remote control, or other interactive devices. The scene switching behavior can include the user switching from the currently automatically generated mood lighting scene to other preset scenes, custom scenes, or historical scenes. The scene rollback behavior can include the user reverting the current target lighting environment to the default scene, the previous lighting scene, or a historically saved scene. The automatic mode interruption behavior can include the user turning off the automatic mood lighting mode, pausing adaptive control, or switching to manual control mode. The continuous use behavior can include the user maintaining the current target lighting environment continuously within a preset time range without manual intervention, scene switching, or automatic mode interruption. The repeated selection behavior can include the user repeatedly calling, confirming, saving, or reselecting the same lighting environment parameter configuration under the same or similar emotional state and scene context.

[0066] In one embodiment of this application, to facilitate subsequent attribution analysis and parameterization of the behavioral feedback event, the behavioral feedback event can be structurally represented as follows:

[0067] e_k={type_k,t_k,space_k,obj_k,op_k,uid_k,conf_k};

[0068] Where e_k represents the behavior feedback event, type_k represents the feedback type, t_k represents the feedback occurrence time, space_k represents the feedback occurrence space, obj_k represents the operation object, op_k represents the operation content, uid_k represents the user identifier, and conf_k represents the user identification confidence level.

[0069] In one embodiment of this application, to avoid misinterpreting changes in user behavior caused by non-lighting factors as feedback on the current target lighting environment, the system can establish a corresponding lighting parameter version number and a valid attribution window for each target lighting environment output. Subsequently, the system filters user behavior feedback events within the preset valid attribution window.

[0070] Specifically, if the target lighting environment is output at time t_0, the effective attribution window can be represented as: [t_0+τ_min, t_0+τ_max], where τ_min represents the minimum attribution delay time, used to exclude immediate accidental touches or irrational operations that occur before the user has fully perceived the target lighting environment; τ_max represents the maximum attribution delay time, used to exclude delayed behaviors with a weak correlation to the current target lighting environment or behavioral changes caused by other scene factors. τ_min and τ_max can be set or dynamically adjusted according to room type, activity type, mood lighting scene, and user's historical behavioral characteristics.

[0071] In step S300, attribution analysis is performed on the correlation between the behavioral feedback event and the parameter configuration of the target light environment to obtain the feedback attribution weight.

[0072] In real-world applications, user-generated behavioral feedback events in response to a target lighting environment typically exhibit strong temporal, spatial, object-related, and user-identity-related characteristics. Whether a user action should be considered valid feedback for the current target lighting environment generally depends on the target lighting environment's output time, output space, lighting parameter version, object of operation, and the confidence level of user identification. Therefore, this embodiment does not directly use all collected behavioral feedback events as the basis for model updates. Instead, it first performs attribution analysis on the correlation between behavioral feedback events and the parameter configuration of the current target lighting environment. The feedback attribution weights obtained through attribution analysis are used to characterize the degree of correlation between the behavioral feedback events and the parameter configuration.

[0073] In one embodiment of this application, the feedback attribution weight corresponding to the user behavior feedback event can be calculated based on the output time, output space, lighting parameter version, operation object, and user identification confidence of the target light environment.

[0074] Specifically, the feedback attribution weights can be calculated as follows:

[0075] q_k=f_t(t_k-t_0)·f_s(space_k,space_t)·f_o(obj_k,P_t)·conf_k;

[0076] Where q_k represents the feedback attribution weight corresponding to the k-th behavioral feedback event; f_t represents the time decay function, used to characterize the temporal correlation between the occurrence time of the behavioral feedback event and the output time of the target light environment; t_k represents the occurrence time of the k-th behavioral feedback event; t_0 represents the output time of the target light environment; f_s is the spatial consistency function, used to characterize the consistency between the feedback occurrence space and the target light environment output space; space_k represents the occurrence space of the k-th behavioral feedback event; space_t represents the output space of the target light environment; f_o is the operation object consistency function, used to characterize the correspondence between the user operation object and the current target light environment parameter configuration; obj_k represents the operation object corresponding to the k-th behavioral feedback event; P_t represents the emotion lighting parameter vector; and conf_k represents the user identity recognition confidence corresponding to the k-th behavioral feedback event.

[0077] In this implementation, feedback attribution weights are calculated based on the effective attribution window, spatial consistency, operation object consistency, and user identity recognition confidence. This allows for the filtering and weighting of behavioral feedback events before model updates, preventing user behaviors caused by non-current target lighting environments from being mistakenly identified as valid feedback for the current lighting parameter configuration. This improves the reliability of subsequent lighting parameter correction vector generation and incremental updates of the emotional lighting preference model.

[0078] In step S400, the behavioral feedback event is converted into a lighting parameter correction vector with the same dimension as the emotion lighting parameter vector, and an adaptation score is calculated based on the behavioral feedback event.

[0079] The lighting parameter correction vector is used to characterize the direction and magnitude of the correction of each parameter dimension in the emotional lighting parameter vector by the behavioral feedback event. The suitability score is used to characterize the user's acceptance of the target lighting environment.

[0080] In this implementation, by converting behavioral feedback events into lighting parameter correction vectors and fitness scores, the system can transform users' explicit operational behaviors and implicit usage behaviors into computationally calculable, accumulative learning signals that can be used for model updates.

[0081] Please see Figure 4 The diagram shows a flowchart of converting behavioral feedback events into lighting parameter correction vectors in one embodiment of this application.

[0082] like Figure 4 As shown, converting the behavioral feedback event into a lighting parameter correction vector with the same dimension as the emotion lighting parameter vector includes the following steps S401 to S404.

[0083] In step S401, when the behavior feedback event is the manual dimming behavior, the manual dimming behavior is analyzed to determine at least one of the following operations: brightness increase / decrease, color temperature increase / decrease, color shift, saturation change, dynamic speed adjustment, or dynamic amplitude adjustment.

[0084] In step S402, according to the preset operation and parameter mapping table, the operation content is mapped to a lighting parameter correction vector with the same dimension as the emotion lighting parameter vector.

[0085] The operation and parameter mapping table is used to represent the correspondence between different user operation content and the parameter dimensions in the emotion lighting parameter vector.

[0086] In step S403, when the behavioral feedback event is the scene switching behavior, the scene rollback behavior, or the automatic mode interruption behavior, a negative correction amount or suppression weight with the same dimension as the emotion lighting parameter vector is generated.

[0087] In step S404, when the behavioral feedback event is the continuous use behavior or the repeated selection behavior, a positive correction amount or reinforcement weight with the same dimension as the emotion lighting parameter vector is generated.

[0088] For example, if a user increases the brightness from 40% to 55% after the target lighting environment is output, the system can generate a positive correction in the brightness dimension of the lighting parameter correction vector; if a user adjusts the color temperature from 6500K to 3500K, the system can generate a negative correction in the relevant color temperature dimension; if a user turns off dynamic effects, the system can generate negative corrections in both the dynamic change amplitude and dynamic change speed dimensions.

[0089] Please see Figure 5 The diagram shows a flowchart of the fit score calculation in one embodiment of this application.

[0090] like Figure 5 As shown, calculating the fit score based on the behavioral feedback event includes the following steps S405 and S406.

[0091] In step S405, within a preset statistics window, indicators associated with the target light environment are obtained; the indicators include usage duration, frequency of manual intervention, probability of scene rollback, and number of times automatic mode is interrupted.

[0092] Specifically, the usage duration characterizes the length of time a user continuously uses the current target lighting environment without switching scenes, reverting to previous scenes, or interrupting the automatic mode. The manual intervention frequency characterizes the frequency with which the user manually adjusts lighting parameters such as brightness, color temperature, color, saturation, dynamic change speed, or dynamic change amplitude within the preset statistical window. The scene revert probability characterizes the likelihood of the user reverting from the current target lighting environment to the default scene, the previous scene, or a historically saved scene. The number of automatic mode interruptions characterizes the number of times the user turns off, pauses, or exits the automatic mood lighting mode.

[0093] In step S406, the usage duration, the frequency of manual intervention, the probability of scene rollback, and the number of automatic mode interruptions are normalized and weighted to obtain the fit score.

[0094] Specifically, the usage duration corresponds to a positive weight, while the manual intervention frequency, the scene rollback probability, and the number of automatic mode interruptions correspond to negative weights.

[0095] In step S500, based on the feedback attribution weight, the lighting parameter correction vector, and the fitness score, the emotional lighting preference model is incrementally updated under preset constraints, and the updated emotional lighting preference model is used for subsequent emotional lighting parameter vector generation.

[0096] Please see Figure 6 The above is a flowchart showing the update process of the mood lighting preference model in one embodiment of this application.

[0097] like Figure 6 As shown, the incremental update of the emotional lighting preference model based on the feedback attribution weight, the lighting parameter correction vector, and the fitness score under preset constraints includes the following steps S501 to S504.

[0098] In step S501, the emotion lighting parameter vector is added to the lighting parameter correction vector to obtain the user expectation parameter vector.

[0099] In step S502, a loss function is constructed using the deviation between the emotion lighting parameter vector output by the emotion lighting preference model and the user expectation parameter vector as the loss term.

[0100] In step S503, the fitness score is normalized to obtain the updated weight.

[0101] In step S504, the parameters of the emotion lighting preference model are updated based on the feedback attribution weights, the loss function, and the update weights.

[0102] Specifically, the parameters of the mood lighting preference model are updated according to the following formula:

[0103] ;

[0104] Where θ_i^{n} represents the parameters of the mood lighting preference model before the nth update; θ_i^{n+1} represents the parameters of the mood lighting preference model after the (n+1)th update; Proj_Ω represents projecting the update result onto the safe and feasible region Ω; the safe and feasible region Ω is determined by the parameter change smoothness constraint and the safe lighting boundary constraint; η represents the learning rate; q_k represents the feedback attribution weight; r_k represents the update weight; and L represents the loss function. This represents the gradient of user preferences.

[0105] In this implementation, by performing constrained incremental updates on the emotion lighting preference model, the system can gradually transform user behavioral feedback during real-world use into changes in model parameters. This makes the subsequently generated emotion lighting parameter vectors more consistent with the user's personalized preferences under specific emotional states, scene contexts, and time conditions. Simultaneously, because the model update process is jointly constrained by feedback attribution weights, update weights, and the safe and feasible region, it avoids learning low-association feedback, occasional operations, or unsafe parameters as stable preferences, thereby improving the stability, reliability, and long-term personalization capabilities of the emotion lighting adaptive control.

[0106] In one embodiment of this application, the emotional lighting preference model is updated using a confidence gating mechanism.

[0107] Please see Figure 7 The diagram shows the working principle of the confidence gating mechanism in one embodiment of this application.

[0108] like Figure 7 As shown, the confidence gating mechanism includes the following steps S601 to S604.

[0109] In step S601, the feedback confidence level of the behavioral feedback event is obtained.

[0110] The feedback confidence score is used to characterize the reliability of the behavioral feedback event as a model update sample. The feedback confidence score can be determined based on one or more of the following: user identity recognition confidence score, feedback collection quality, feedback event completeness, feedback type reliability, consistency with historical similar feedback, and scene context matching degree.

[0111] In step S602, when at least one of the feedback attribution weight, the fit score, and the feedback confidence is lower than the corresponding preset threshold, only the corresponding behavioral feedback event is recorded as a feedback sample, and the emotion lighting preference model is not updated immediately.

[0112] In this way, the system can avoid unstable model updates caused by single low-confidence feedback, accidental touches, temporary operations, or operations with unclear identities.

[0113] In step S603, when similar feedback samples repeatedly appear within a preset time range and the cumulative number reaches a preset cumulative threshold, the emotion lighting preference model is updated based on the similar feedback samples.

[0114] The similar feedback samples refer to feedback samples with the same feedback type, the same corresponding light parameter dimension, and a scene context that meets preset similarity conditions. The scene context meeting the preset similarity conditions may include one or more of the following: the same or similar room type, the same or similar time period, the same or similar activity type, the same or similar emotional state type, and a corresponding relationship between the current lighting parameter version or lighting scene type.

[0115] For example, if a user manually reduces the dynamic rate of change in a relaxing bedroom scenario over multiple nights, and the cumulative number of such feedback samples reaches a preset cumulative threshold, the system can consider this feedback to reflect the user's stable preference in the corresponding scenario, and update the emotional lighting preference model based on these feedback samples. Thus, the system can gradually accumulate recurring low-intensity feedback into stable learning signals, while reducing the impact of single, occasional feedback on model updates.

[0116] In step S604, when the number of feedback samples is lower than a preset sample threshold, within the safe lighting boundary, an exploration perturbation of no more than a preset perturbation amplitude is applied to at least one of the light parameter dimensions of brightness, color temperature, color, dynamic speed, or transition time, and the user preference gradient in the emotional lighting preference model is estimated based on the change in fit score before and after the exploration perturbation is applied.

[0117] For example, the system can make minor perturbations to brightness or color temperature, provided that it does not cause glare, does not exceed the color temperature change rate threshold, does not exceed the dynamic change rate threshold, and does not exceed the stimulation threshold in children's or elderly modes. If the fit score increases after the perturbation, it can be considered that the perturbation direction is closer to the user's preference; if the fit score decreases after the perturbation, or if negative feedback such as manual intervention, scene rollback, or automatic mode interruption occurs, the system can immediately stop the perturbation and restore the lighting parameter configuration before the perturbation.

[0118] In this implementation, through the aforementioned confidence gating mechanism, this application can perform model updates when the feedback samples are reliable, delay updates or only record samples when the feedback samples are unreliable, and estimate user preference gradients through small-scale exploration within a safe range when the feedback samples are insufficient. This improves the stability and reliability of incremental updates to the emotion lighting preference model, avoiding erroneous shifts caused by low-confidence feedback, sporadic behavior, or insufficient samples during the cold start phase.

[0119] Please see Figure 8 The above is a flowchart of an adaptive control method for mood lighting in another embodiment of this application.

[0120] like Figure 8 As shown, the mood lighting adaptive control method described in this application further includes the following steps S701 to S703.

[0121] In step S701, when multiple family members are in a shared lighting space or the same lighting scene, a family member conflict map is constructed based on the individual-level emotional lighting preference model of each family member.

[0122] In one embodiment of this application, the individual-level emotion lighting preference model includes an emotion-light mapping matrix M_i, a parameter sensitivity matrix S_i, and a feedback confidence matrix C_i. The emotion-light mapping matrix M_i can be indexed according to emotion type, emotion intensity, scene type, and time period, and is used to output candidate light parameters for user i under specific conditions. The parameter sensitivity matrix S_i records the user's sensitivity to each light parameter dimension. For example, if a user is highly sensitive to dynamic speed, the system will limit the exploration range of this dimension in subsequent updates. The feedback confidence matrix C_i records the credibility of feedback samples under different conditions, and is used to prevent low-confidence feedback from leading to over-updates.

[0123] The family member conflict diagram uses family members as nodes and the preference differences, spatial overlap, activity priority, and safety priority among family members as edge weights. The preference differences are used to characterize the degree of difference between different family members in lighting parameters such as brightness, color temperature, color, saturation, dynamic change rate, dynamic change amplitude, transition time, or spatial weight. The spatial overlap is used to characterize the degree of overlap between the locations, activity areas, or lighting coverage areas of different family members. The activity priority is used to characterize the priority of different activity types in terms of lighting environment requirements. The safety priority is used to characterize the safety lighting requirements for children, the elderly, sleep scenarios, eye-use scenarios, or other special scenarios.

[0124] In one implementation, when the preference differences between two family members are significant and the shared space overlaps considerably, the edge weight between them can be increased to indicate that the two family members have a high degree of lighting preference conflict in the current shared space or the same lighting scene, requiring conflict resolution in subsequent steps. Conversely, when the preference differences between two family members are small, or when they are in different local lighting areas, the edge weight between them can be decreased to indicate that the lighting preference conflict between the two family members is weak.

[0125] For example, in a living room movie-watching scenario, if user A prefers higher color saturation and stronger dynamic ambient lighting, while user B prefers lower brightness and slower dynamic changes, the system can assign a higher edge weight to the edge between user A and user B in the family member conflict graph G_h based on the difference between their candidate mood lighting parameter vectors. The system can then use this edge weight to determine whether a compromise shared light field needs to be generated, or further generate a combination of parameters for the shared light field and local light fields, to accommodate the lighting needs of multiple family members.

[0126] In step S702, based on the family member conflict diagram, a shared area lighting parameter vector is generated through the family-level emotional lighting collaboration model.

[0127] In one embodiment of this application, based on the family member conflict diagram, a family-level emotional lighting collaboration model is used to generate shared area lighting parameters within a safe and feasible region. The shared area lighting parameters are then solved to minimize the following group optimization objective:

[0128] ;

[0129] Where P represents the shared area lighting parameters to be solved; P_i represents the candidate lighting parameters of the i-th family member; P_j represents the candidate lighting parameters of the j-th family member; w_i represents the member weight of the i-th family member; a_{ij} represents the edge weight between the i-th and j-th family members in the family member conflict graph G_h; R_safe represents the safe lighting boundary penalty term; β and γ represent the weight coefficients.

[0130] Among the aforementioned group optimization objectives, The shared area lighting parameters P are used to constrain the lighting parameters P of each family member to be as close as possible to the candidate mood lighting parameter vectors P_i of each family member, so as to improve the overall adaptability of the shared lighting environment to the preferences of multiple members. It is used to characterize the intensity of preference conflicts among family members and to enable the system to perform more careful collaborative solutions to lighting parameters in shared areas when the conflict is strong.

[0131] If the obtained shared area lighting parameters P can simultaneously satisfy the basic preferences, safety lighting boundaries, and device output constraints of multiple family members, the system can generate a shared light field based on the shared area lighting parameters and output the shared light field through main lights, downlights, surface light sources, or other lighting devices covering the shared area.

[0132] In step S703, when the solved shared area lighting parameter vector cannot simultaneously satisfy the preference constraints of multiple family members, the shared lighting space is divided into a shared light field and at least one local light field, and lighting parameters for the shared light field and the local light field are generated respectively.

[0133] The shared light field is used to cover a basic area where multiple family members engage in common activities, and outputs a low-conflict, stable basic light environment that meets safety lighting boundaries. The local light field is used to cover the location of a specific family member or a specific activity area, and outputs lighting parameters that are more tailored to the individual preferences of the corresponding family member. The local light field can be output by local luminaires, light strips, reading lights, desktop lights, background lights, or other lighting devices with local lighting capabilities.

[0134] Through the above methods, this application can achieve consistency, safety and stability of the basic lighting environment of a shared space by using a combination of shared light field and local light field output when the lighting parameters of a single shared area are difficult to take into account the preferences of multiple family members. At the same time, it can take into account the differences in the individual emotional lighting preferences of different family members, thereby improving the practical usability of adaptive control of emotional lighting in multi-member family scenarios.

[0135] In one embodiment of this application, the individual-level emotion lighting preference model and the family-level emotion lighting collaborative model adopt a multi-timescale update mechanism, wherein the multi-timescale update mechanism includes: recording behavioral feedback events and fit scores at the minute-level timescale; updating stable preferences based on feedback samples that meet confidence requirements at the daily-level timescale; detecting user preference drift at the weekly or monthly-level timescale; and performing version evaluation on the model update results, and performing model version rollback when preset rollback conditions are met.

[0136] Through the aforementioned multi-timescale update mechanism, this application can record real-time feedback at the minute level, form stable preferences at the daily level, identify long-term preference drift at the weekly or monthly level, and avoid mislearning occasional behaviors, short-term emotional fluctuations, or low-confidence feedback as long-term preferences through model version evaluation and rollback mechanisms, thereby improving the stability, reliability, and long-term personalized effect of the emotion lighting adaptive control process.

[0137] In one embodiment of this application, the mood lighting adaptive control method further includes step S800.

[0138] In step S800, the updated individual-level emotion lighting preference model, family-level emotion lighting collaboration model, or a summary of their model versions are stored for subsequent emotion lighting parameter generation, self-learning optimization, and model version rollback.

[0139] In this implementation, through the aforementioned model storage, version summary management, verification evaluation, and rollback mechanisms, this application can maintain the traceability of the model update process while continuously learning emotion lighting preferences, and promptly restore to a more stable historical version when the model update effect declines, thereby improving the long-term reliability, stability, and practical usability of the emotion lighting adaptive control method.

[0140] It should be noted that the scope of protection of the emotion lighting adaptive control method described in the embodiments of this application is not limited to the execution order of the steps listed in this embodiment. Any solution implemented by adding, subtracting, or replacing steps in the prior art based on the principles of this application is included within the scope of protection of this application.

[0141] Please see Figure 9 The diagram shown is a structural schematic of an adaptive lighting control system for mood according to an embodiment of this application.

[0142] like Figure 9 As shown, this application provides an adaptive lighting control system for mood lighting, including:

[0143] The mood lighting module is used to generate a mood lighting parameter vector based on a mood lighting preference model, and to control the output of the target light environment of the lighting device based on the mood lighting parameter vector.

[0144] The feedback acquisition module is used to acquire behavioral feedback events generated by the user in response to the target lighting environment; the behavioral feedback events include one or more of the following: manual dimming behavior, scene switching behavior, scene rollback behavior, automatic mode interruption behavior, continuous use behavior, and repeated selection behavior.

[0145] The attribution analysis module is used to perform attribution analysis on the correlation between the behavioral feedback event and the parameter configuration of the target light environment to obtain the feedback attribution weight; the feedback attribution weight is used to characterize the degree of correlation between the behavioral feedback event and the parameter configuration.

[0146] The adaptation scoring module is used to convert the behavioral feedback event into a lighting parameter correction vector with the same dimension as the emotional lighting parameter vector, and to calculate the adaptation score based on the behavioral feedback event; the lighting parameter correction vector is used to characterize the direction and magnitude of the correction of each parameter dimension in the emotional lighting parameter vector by the behavioral feedback event; the adaptation score is used to characterize the user's acceptance of the target lighting environment.

[0147] The model update module is used to incrementally update the emotional lighting preference model based on the feedback attribution weight, the lighting parameter correction vector, and the fitness score under preset constraints, and to use the updated emotional lighting preference model for subsequent emotional lighting parameter vector generation.

[0148] It should be noted that the structure and principle of the emotion lighting module, the feedback acquisition module, the attribution analysis module, the adaptation scoring module and the model update module described in this embodiment correspond one-to-one with the steps in the above-mentioned emotion lighting adaptive control method, so they will not be repeated here.

[0149] Please see Figure 10 The image shown is a schematic diagram of the structure of an electronic device according to an embodiment of this application.

[0150] like Figure 10 As shown, this application provides an electronic device, including:

[0151] The memory is used to store computer programs;

[0152] A processor, the processor being configured to execute a computer program stored in the memory, so as to cause the electronic device to perform the method described in any of the preceding descriptions.

[0153] In some embodiments, the processor may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The memory may be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0154] In the embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, or methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of modules / units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or units may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of apparatuses or modules or units may be electrical, mechanical, or other forms.

[0155] The modules / units described as separate components may or may not be physically separate. The components shown as modules / units may or may not be physical modules; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules / units can be selected to achieve the objectives of the embodiments of this application, depending on actual needs. For example, the functional modules / units in the various embodiments of this application may be integrated into one processing module, or each module / unit may exist physically separately, or two or more modules / units may be integrated into one module / unit.

[0156] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0157] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the methods described in any of the above embodiments. Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing a processor. The program can be stored in a computer-readable storage medium, which is a non-transitory medium, such as random access memory, read-only memory, flash memory, hard disk, solid-state drive, magnetic tape, floppy disk, optical disk, and any combination thereof. The storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., digital video disc (DVD)), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0158] The descriptions of the processes or structures corresponding to the above figures each have their own emphasis. For parts of a process or structure that are not described in detail, please refer to the relevant descriptions of other processes or structures.

[0159] The above embodiments are merely illustrative of the principles and effects of this application and are not intended to limit this application. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of this application. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this application should still be covered by the claims of this application.

Claims

1. An adaptive control method for mood lighting, characterized in that, include: An emotional lighting parameter vector is generated based on an emotional lighting preference model, and the lighting equipment is controlled to output the target light environment based on the emotional lighting parameter vector. Acquire user behavior feedback events in response to the target lighting environment; the behavior feedback events include one or more of the following: manual dimming behavior, scene switching behavior, scene rollback behavior, automatic mode interruption behavior, continuous use behavior, and repeated selection behavior; Attribution analysis is performed on the correlation between the behavioral feedback event and the parameter configuration of the target light environment to obtain the feedback attribution weight; The feedback attribution weight is used to characterize the degree of correlation between the behavioral feedback event and the parameter configuration; The behavioral feedback event is converted into a lighting parameter correction vector with the same dimension as the emotion lighting parameter vector, and an fit score is calculated based on the behavioral feedback event. The lighting parameter correction vector is used to characterize the direction and magnitude of the correction of each parameter dimension in the emotional lighting parameter vector by the behavioral feedback event; The compatibility score is used to characterize the user's acceptance of the target lighting environment; Based on the feedback attribution weight, the lighting parameter correction vector, and the fitness score, the emotional lighting preference model is incrementally updated under preset constraints, and the updated emotional lighting preference model is used for subsequent emotional lighting parameter vector generation.

2. The method according to claim 1, characterized in that, Controlling the output target light environment of the lighting device based on the aforementioned mood lighting parameter vector includes: The capability information of the lighting device is obtained, and a lighting device capability matrix is ​​constructed based on the capability information of the lighting device. The lighting device capability matrix is ​​a data structure that parametrically expresses the controllable output capability of the lighting device. The controllable output capability includes luminous flux range, color temperature range, color gamut range, channel structure, dimming depth, spatial coverage area, minimum response time, and maximum rate of change. Based on the lighting equipment capability matrix, under the conditions of satisfying the preset safety lighting boundary and equipment output constraints, the constraint solution of the mood lighting parameter vector is performed to obtain the channel control value of the lighting equipment; Based on the channel control value, the corresponding lighting device is driven to output the target light environment.

3. The method according to claim 1, characterized in that, The fitness score is calculated based on the aforementioned behavioral feedback events, including: Within a preset statistical window, acquire indicators associated with the target light environment; these indicators include usage duration, frequency of manual intervention, probability of scene rollback, and number of times automatic mode is interrupted. The adaptation score is obtained by normalizing and weighting the usage duration, the manual intervention frequency, the scene rollback probability, and the number of automatic mode interruptions, wherein the usage duration corresponds to a positive weight, and the manual intervention frequency, the scene rollback probability, and the number of automatic mode interruptions correspond to negative weights.

4. The method according to claim 1, characterized in that, Based on the feedback attribution weights, the lighting parameter correction vectors, and the fitness score, incremental updates to the emotion lighting preference model under preset constraints include: The user expectation parameter vector is obtained by adding the emotion lighting parameter vector to the lighting parameter correction vector. A loss function is constructed using the deviation between the emotional lighting parameter vector output by the emotional lighting preference model and the user's expected parameter vector as the loss term; The fitness score is normalized to obtain the updated weights; The parameters of the emotion lighting preference model are updated based on the feedback attribution weights, the loss function, and the update weights.

5. The method according to claim 1, characterized in that, It also includes updating the emotion lighting preference model using a confidence gating mechanism; the confidence gating mechanism includes: Obtain the feedback confidence score of the behavioral feedback event, which is used to characterize the reliability of the behavioral feedback event as a model update sample; When at least one of the feedback attribution weight, the fit score and the feedback confidence is lower than the corresponding preset threshold, only the corresponding behavioral feedback event is recorded as a feedback sample, and the emotion lighting preference model is not updated immediately. When similar feedback samples appear repeatedly within a preset time range and the cumulative number reaches a preset cumulative threshold, the emotion lighting preference model is updated based on the similar feedback samples; the similar feedback samples refer to feedback samples with the same feedback type, the same corresponding light parameter dimension, and the scene context satisfying preset similarity conditions. When the number of feedback samples is lower than a preset sample threshold, within the safe lighting boundary, an exploration perturbation of no more than a preset perturbation amplitude is applied to at least one of the light parameter dimensions of brightness, color temperature, color, dynamic speed, or transition time, and the user preference gradient in the emotional lighting preference model is estimated based on the change in fit score before and after the application of the exploration perturbation.

6. The method according to claim 1, characterized in that, The emotional lighting preference model includes an individual-level emotional lighting preference model and a family-level emotional lighting synergy model; it also includes: When multiple family members are in a shared lighting space or the same lighting scene, a family member conflict graph is constructed based on the individual-level emotional lighting preference model of each family member. The family member conflict graph uses family members as nodes and the preference differences, spatial overlap, activity priority, and safety priority among family members as edge weights. Based on the family member conflict diagram, a shared area lighting parameter vector is generated through the family-level emotional lighting collaboration model. When the obtained shared area lighting parameter vector cannot simultaneously satisfy the preference constraints of multiple family members, the shared lighting space is divided into a shared light field and at least one local light field, and lighting parameters for the shared light field and the local light field are generated respectively.

7. The method according to claim 1, characterized in that, Converting the behavioral feedback event into a lighting parameter correction vector of the same dimension as the emotion lighting parameter vector includes: When the behavior feedback event is the manual dimming behavior, the manual dimming behavior is analyzed to determine at least one of the following operations: brightness increase / decrease, color temperature increase / decrease, color shift, saturation change, dynamic speed adjustment, or dynamic amplitude adjustment. According to the preset operation and parameter mapping table, the operation content is mapped to a lighting parameter correction vector with the same dimension as the emotion lighting parameter vector; When the behavioral feedback event is the scene switching behavior, the scene rollback behavior, or the automatic mode interruption behavior, a negative correction amount or suppression weight with the same dimension as the emotion lighting parameter vector is generated. When the behavioral feedback event is the continuous use behavior or the repeated selection behavior, a positive correction amount or reinforcement weight with the same dimension as the emotion lighting parameter vector is generated.

8. An adaptive control system for mood lighting, characterized in that, include: The mood lighting module is used to generate a mood lighting parameter vector based on a mood lighting preference model, and to control the lighting device to output a target light environment based on the mood lighting parameter vector; The feedback acquisition module is used to acquire behavioral feedback events generated by the user in response to the target lighting environment; the behavioral feedback events include one or more of the following: manual dimming behavior, scene switching behavior, scene rollback behavior, automatic mode interruption behavior, continuous use behavior, and repeated selection behavior. The attribution analysis module is used to perform attribution analysis on the correlation between the behavioral feedback event and the parameter configuration of the target light environment, and obtain the feedback attribution weight. The feedback attribution weight is used to characterize the degree of correlation between the behavioral feedback event and the parameter configuration; The adaptation scoring module is used to convert the behavioral feedback event into a lighting parameter correction vector with the same dimension as the emotional lighting parameter vector, and to calculate the adaptation score based on the behavioral feedback event. The lighting parameter correction vector is used to characterize the direction and magnitude of the correction of each parameter dimension in the emotional lighting parameter vector by the behavioral feedback event; The compatibility score is used to characterize the user's acceptance of the target lighting environment; The model update module is used to incrementally update the emotional lighting preference model based on the feedback attribution weight, the lighting parameter correction vector, and the fitness score under preset constraints, and to use the updated emotional lighting preference model for subsequent emotional lighting parameter vector generation.

9. An electronic device, characterized in that, include: The memory is used to store computer programs; A processor for executing a computer program stored in the memory to cause the electronic device to perform the method of any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 7.

Images