Smart home illumination control method based on scene adaptive adjustment

Abstract

The invention provides a smart home illumination control method based on scene adaptive adjustment, and relates to the technical field of illumination intelligent control, and the method comprises the steps: 1, obtaining indoor data and personnel comfort illumination intensity to form a model data set, 2, carrying out the indoor division, and forming a partition data set, 4, training an illumination intensity prediction model; 5, obtaining balance recommendation illumination intensity; 6, adjusting the balance recommendation illumination intensity; 7, analyzing partitions needing shading; according to the invention, through model construction and partition division, the illumination condition of each scene is accurately adjusted and controlled, and coordinated comfortable illumination and timely adjustment are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent lighting control, and specifically to a smart home lighting control method based on scene adaptive adjustment. Background Art

[0002] Existing indoor lighting systems usually adjust indoor lighting through simple light sensors or preset timing controls. However, this method cannot accurately consider the comprehensive influence of various factors such as indoor and outdoor environmental changes, users' behavior habits, and air quality, resulting in poor lighting effects and difficulty in meeting the personalized needs of different people in different scenarios. At the same time, the existing systems lack the ability to respond to real-time fluctuations in outdoor light intensity and the intelligent analysis of multi-region lighting balance, which easily causes too strong or too weak light in local areas. In addition, the existing methods usually cannot continuously optimize the model and cannot dynamically improve the prediction accuracy according to the actual adjustment behavior of users, restricting the adaptive ability of the system.

[0003] In the prior art, the publication number CN200910108229.6 discloses a subway scene lighting control system and method. The system includes a call control module for triggering the system to run; a scene trigger module for finding the dimming controller address, lighting device group address, and light brightness regulation value corresponding to each scene number from the database according to the preset scene number and operation time period of the system; a data processing module for sending the lighting device group address and light brightness regulation value information to the dimming controller at the corresponding address; and a dimming control module for adjusting the light brightness value of the corresponding lighting device group according to the received lighting device group address and light brightness regulation value.

[0004] Although the above information disclosed in the background art realizes lighting control based on scenes, its control method is not fine and specific enough for the control of corresponding light, and the handling of environmental changes and personnel comfort conditions is not perfect enough. Summary of the Invention

[0005] The purpose of the present invention is to provide a smart home lighting control method based on scene adaptive adjustment to solve the problems raised in the above background art.

[0006] To achieve the above purpose, the present invention provides the following technical solutions:

[0007] A smart home lighting control method based on scene adaptive adjustment, the specific steps include:

[0008] Step 1: Obtain light intensity data, indoor temperature and humidity data, and personnel behavior data, manually adjust the indoor light intensity to the comfortable light intensity for personnel, and synchronously record the comfortable light intensity data of personnel to form a model data set;

[0009] Step 2: Extract features from the model dataset to obtain indoor light intensity features, outdoor light intensity features, indoor temperature feature T, indoor humidity feature H, user behavior features, carbon dioxide concentration features, and PM2.5 concentration features respectively, form a feature set and train it in a linear regression model to construct a light intensity prediction model;

[0010] Step 3: Divide the indoor area, obtain the recommended light intensity for each partition through the data of each partition, identify the partitions that need lighting through the evaluation results of the environment of each partition, and at the same time perform light intensity uniformity analysis on the partitions that need lighting, and balance the partitions with large differences according to the situation of adjacent partitions;

[0011] Step 4: During the continuous operation of the lighting system, analyze the impact of changes in outdoor light intensity on indoor light intensity through the actual difference between indoor and outdoor light. When the outdoor light intensity changes greatly, adjust the startup situation and brightness of the lighting system for each partition in real time;

[0012] Step 5: Compare the light intensity of each partition with the recommended light intensity, and shade the partitions with excessive indoor light intensity;

[0013] Step 6: Adjust the light intensity prediction model according to the actual adjustment of the light intensity by the user, identify the indoor manual light intensity and replace the recommended light intensity.

[0014] Furthermore, obtain light intensity data, which includes outdoor light intensity and indoor light intensity, obtain indoor temperature and humidity data, which includes temperature and humidity, obtain personnel behavior data, obtain indoor air quality data, which includes carbon dioxide concentration and PM2.5 concentration, the behavior data includes four categories: reading, dining, entertainment, and rest. The collection frequencies of the light intensity data, indoor temperature and humidity data, and personnel behavior data are the same, add timestamps and align them in time. Preprocess the data to form a model dataset. Manually adjust the indoor light intensity to the comfortable light intensity for personnel in real time, and synchronously collect the indoor comfortable light intensity data with the model dataset, and add the indoor comfortable light intensity data as labels to the model dataset.

[0015] Furthermore, perform feature extraction on the model dataset, perform one-hot encoding on the behavior data, convert each behavior category into ternary variables that are different from each other to obtain behavior features, normalize the light intensity data, indoor temperature and humidity data, and air quality data, and obtain indoor light intensity features, outdoor light intensity features, indoor temperature features, indoor humidity features, carbon dioxide concentration features, and PM2.5 concentration features respectively. Extract hour features and month features from the timestamps. The basis formulas are as follows:

[0016]

[0017] Among them, Ho sin is the hour feature value, Ho is the hour value in the timestamp, and Mo sin is the month feature value, and Mo is the month value in the timestamp;

[0018] The extracted features are formed into a model feature set and input into a linear regression analysis model to construct a light intensity prediction model. The formula is as follows:

[0019] y = β0 + β1L in + β2L out + β3T + β4H + β5A user + β6CO + β7PM + β8Ho sin + β9Mo sin

[0020] Among them, y is the recommended light intensity, β0 is the bias term representing the constant value, and β1, β2, β3, β4, β5, β6, β7, β8, β9 are the model coefficients respectively, representing the influence degree of each input feature on the lighting adjustment. L in is the indoor light intensity feature, L out is the outdoor light intensity feature, T is the indoor temperature feature, H is the indoor humidity feature, A user is the user behavior feature, CO is the carbon dioxide concentration feature, PM is the PM2.5 concentration feature, Ho sin is the hour feature, and Mo sin is the month feature.

[0021] Furthermore, before the indoor lighting system needs to be started, the indoor area is divided. The division logic is to divide according to the indoor functional areas. The light intensity data, indoor temperature and humidity data, and behavior data are obtained in real time in each area. Each area forms a partition data set separately. The partition feature sets formed by each partition data set are input into the light intensity prediction model, and the recommended light intensity of each partition is obtained separately. The recommended light intensity of each partition and the partition environment are judged. The formula is as follows:

[0022] ΔL in-i = L in-i - y i

[0023] Among them, ΔL in-i represents the light intensity difference of the i-th partition, L in-i is the indoor light intensity of the i-th partition, and y i$L_i$ is the recommended light intensity for the $i$-th partition, where $i$ is the partition retrieval variable, $i\in N$, and $1\leq i\leq n$;

[0024] When $\Delta L$ in-i $\geq0$, the lighting system for the $i$-th partition is not activated. When $\Delta L$ in-i $<0$, the lighting system for the $i$-th partition is activated and the light intensity of the lighting system is preset to the light intensity difference;

[0025] Count the partitions that need to activate the lighting system, and judge the light intensity difference between the partitions that need to activate the lighting system. The formula is as follows:

[0026]

[0027] where $JFC$ is the mean square deviation of the light intensity between the partitions that need to activate the lighting system, $y$ i is the recommended light intensity for the $i$-th partition that needs to activate the lighting system, is the average value of the recommended light intensities of the partitions that need to activate the lighting system, $k$ is the number of partitions that need to activate the lighting system, and $i$ is the partition retrieval variable of the partitions that need to activate the lighting system, $i\in N$, and $1\leq i\leq k$;

[0028] Judge the change of the light intensity between each partition in the room through the mean square deviation. Because if the light intensity difference between different partitions is too large, it will also cause discomfort to the user. Therefore, it is necessary to judge the light intensity difference between different partitions.

[0029] Conduct an analysis of the light intensity balance in the room, set the threshold of $JFC$. When $JFC$ exceeds the threshold, it means that the light intensity difference between the partitions that need to activate the lighting system in the room is large, and it is necessary to balance the recommended light intensity. When $JFC$ does not exceed the threshold, it means that the difference between the partitions that need to activate the lighting system in the room is small, and there is no need to balance the light intensity. The logic of the light intensity balance is as follows:

[0030] Identify the two partitions with the highest and lowest recommended light intensities, adjust the recommended light intensities of these two partitions to the recommended light intensities of their adjacent partitions respectively, and recalculate the mean square deviation of the light intensity to judge whether the mean square deviation still exceeds the threshold. If it exceeds the threshold, identify the two partitions with the highest and lowest light intensities again, and adjust the recommended light intensities of these two partitions towards the adjacent partitions until the mean square deviation of the light intensity is lower than the threshold, so as to obtain the balanced recommended light intensity;

[0031] Adjust the light intensity difference again according to the balanced recommended light intensity. Each partition that needs to activate the lighting system illuminates the room according to the light intensity difference, while each partition that does not need to activate the lighting system remains in the non-activated state.

[0032] Further, in the state where the lighting system in a partition has been turned on, continuously collect the light intensity data, indoor temperature and humidity data, and behavior data of each partition, and make a judgment on the change in outdoor light intensity based on the light intensity data. The formula is as follows:

[0033]

[0034] Among them, ΔLC is the difference between indoor and outdoor light, y i is the recommended light intensity of the i-th partition, L out is the outdoor light intensity, i is the partition number retrieval variable, i ∈ N, 1 ≤ i ≤ n;

[0035] Judge the change in the overall indoor light intensity and outdoor light intensity through the difference. After the indoor lighting system is started, the indoor light intensity data is likely not to change. Simply based on the indoor light intensity as a benchmark, it is difficult to cope with the change in outdoor light intensity. Therefore, calculate the difference between the indoor light intensity and the outdoor light intensity to judge the change in outdoor light intensity.

[0036] Organize the indoor and outdoor light difference of each partition into an indoor and outdoor light difference data set ΔLC. The formula is as follows:

[0037] ΔLC = {Lc1, Lc2, Lc3,..., Lc i ,..., Lc n}

[0038] Among them, Lc i is the indoor and outdoor light difference at the i-th moment, i is the time retrieval variable, i ∈ n, 1 ≤ i ≤ n;

[0039] Identify that k consecutive data in the indoor and outdoor light difference data set ΔLC are greater than 0. Set the continuous change threshold value. When k does not exceed the continuous change threshold, it means that the outdoor light intensity has changed but not much. When k exceeds the continuous change threshold, it means that the outdoor light intensity has changed greatly and the indoor brightness needs to be adjusted. The logic of the brightness adjustment is as follows:

[0040] Compare the recommended light intensity of each partition with the outdoor light intensity. For the partitions where the recommended light intensity is greater than the outdoor light, if the lighting system has not been started, start the lighting system in this area according to the recommended light intensity. If the lighting system has been started, adjust all the partitions where the lighting system has been started according to the balanced light intensity obtained from the indoor light balance analysis;

[0041] Adjust the indoor lighting situation in real time according to the change in outdoor light intensity.

[0042] Further, continuously collect the light intensity data, and adjust the partitions with high indoor light intensity, which is calibrated as the light intensity area adjustment. The logic of the light intensity area adjustment is as follows:

[0043] The recommended light intensity obtained from the light intensity prediction model is compared with the indoor light intensity, and the formula used is:

[0044] ΔLLC = L in -β·y

[0045] Where, ΔLLC is the light intensity comparison difference of this partition, L in is the indoor light intensity of this partition, y is the recommended light intensity of this partition, β is a coefficient. When ΔLLC > 0, it means that the indoor light intensity of this partition is too high, then the light-shielding curtain of this partition is activated to adjust the indoor light intensity so that ΔLLC ≤ 0. When ΔLLC ≤ 0, no adjustment is made.

[0046] Furthermore, during the continuous operation of the lighting system, user adjustment data is collected. The user adjustment data is the light intensity adjustment value, and the light intensity adjustment value is the indoor manual light intensity after the user manually adjusts the indoor light intensity. At the same time, the light intensity data, indoor temperature and humidity data, and behavior data at this time are calibrated, and the data at this time is preprocessed and feature extracted. The expansion ranges of the light intensity data and indoor temperature and humidity data are set. The expansion range is within the upper and lower limits of the calibrated light intensity data and indoor temperature and humidity data floating. When the light intensity prediction model calculates the recommended light intensity again through the calibrated light intensity, indoor temperature and humidity data, and behavior data, the value of the recommended light intensity is replaced with the light intensity adjustment value.

[0047] Compared with the prior art, the beneficial effects of the present invention are:

[0048] The present invention realizes personalized recommended light intensity for different scenarios by constructing a light intensity prediction model. Based on the light adjustment strategy of area division, it compares the indoor and outdoor light changes in real time, realizes coordinated lighting in multiple areas, improves the uniformity of indoor light, and can respond to environmental changes in a timely manner. Brief Description of the Drawings

[0049] Figure 1 It is a schematic diagram of the overall method flow of the present invention. Detailed Embodiments

[0050] In order to make the objectives, technical solutions, and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below in conjunction with specific embodiments.

[0051] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should have the ordinary meanings understood by those with ordinary skills in the field to which the present invention pertains. The "first", "second" and similar terms used in the present invention do not denote any order, quantity or importance, but are only used to distinguish different components. The terms such as "comprising" or "including" mean that the elements or objects appearing before this word cover the elements or objects listed after this word and their equivalents, without excluding other elements or objects. The terms such as "connected" or "linked" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The terms such as "upper", "lower", "left" and "right" are only used to represent relative position relationships. When the absolute position of the object being described changes, the relative position relationship may also change accordingly.

[0052] Embodiment:

[0053] Please refer to Figure 1 , the present invention provides a technical solution:

[0054] A smart home lighting control method based on scene adaptive adjustment, the specific steps include:

[0055] Step 1: Obtain light intensity data, indoor temperature and humidity data, and personnel behavior data. Manually adjust the indoor light intensity to the comfortable light intensity for personnel in actuality, and synchronously record the comfortable light intensity data for personnel to form a model data set.

[0056] The said Step 1 includes:

[0057] Obtain light intensity data, the light intensity data includes outdoor light intensity and indoor light intensity, obtain indoor temperature and humidity data, the indoor temperature and humidity data includes temperature and humidity, obtain personnel behavior data, obtain indoor air quality data, the indoor air quality data includes carbon dioxide concentration and PM2.5 concentration, the behavior data includes four categories of reading, dining, entertainment and rest, the personnel behavior data is recognized by a wearable device, the collection frequencies of the light intensity data, indoor temperature and humidity data and personnel behavior data are the same, add time stamps and align them in time, preprocess the data and form a model data set, manually adjust the indoor light intensity to the comfortable light intensity for personnel in real time, and synchronously collect the comfortable light intensity data for the indoor with the model data set, and add the indoor comfortable light intensity data as a label into the model data set.

[0058] Obtain the indoor light intensity data. According to the actual light intensity analysis, it can be determined how much additional light intensity is needed to reach the light intensity comfortable for the human body, achieving an energy-saving effect. Moreover, in different indoor environments, such as the levels of temperature and humidity and the condition of indoor air quality, the light intensity comfortable for the human body will also vary. Therefore, when calculating the light intensity, not only the scenarios of the personnel but also the realistic factors need to be considered, taking into account all environmental factors that may affect human comfort, so as to obtain the recommended light intensity closest to the real scenario when training the model according to various specific and real situations.

[0059] Step 2: Extract features from the model dataset to obtain indoor light intensity features, outdoor light intensity features, T as indoor temperature features, H as indoor humidity features, user behavior features, carbon dioxide concentration features, and PM2.5 concentration features respectively, form a feature set and train it in a linear regression model to construct a light intensity prediction model;

[0060] The said Step 2 includes:

[0061] Step 201: Extract features from the model dataset, perform one-hot encoding on the behavior data, convert each behavior category into ternary variables that are mutually different to obtain behavior features. Reading is encoded as [0, 0, 1], dining is encoded as [0, 1, 0], entertainment is encoded as [1, 0, 0], rest is encoded as [0, 1, 1]. Normalize the light intensity data, indoor temperature and humidity data, and air quality data to obtain indoor light intensity features, outdoor light intensity features, indoor temperature features, indoor humidity features, carbon dioxide concentration features, and PM2.5 concentration features respectively. Extract hour features and month features from the timestamp, and the formulas are as follows:

[0062]

[0063] Among them, Ho sin is the hour feature value, Ho is the hour value in the timestamp, Mo sin is the month feature value, Mo is the month value in the timestamp;

[0064] The periodic changes in hours and months reflect the periodic characteristics of outdoor light intensity. As a reference for fully considering outdoor light intensity at different time points during model training, one-hot encoding is performed on the behavior categories to convert the text into specific data for use in model training. The light intensity data, indoor temperature and humidity data, and air quality data are normalized to obtain indoor light intensity features, outdoor light intensity features, indoor temperature features, indoor humidity features, carbon dioxide concentration features, and PM2.5 concentration features, avoiding affecting the results of model training due to the numerical differences of the data itself, enabling the model to obtain more accurate data while reducing the computational amount.

[0065] Step 202:

[0066] The extracted features are formed into a model feature set and input into a linear regression analysis model to construct a light intensity prediction model. The formula is as follows:

[0067] y = β0 + β1L in + β2L out + β3T + β4H + β5A user + β6CO + β7PM + β8Ho sin + β9Mo sin

[0068] Among them, y is the recommended light intensity, β0 is the bias term representing the constant value, and β1, β2, β3, β4, β5, β6, β7, β8, β9 are the model coefficients respectively, representing the influence degree of each input feature on the lighting adjustment. L in is the indoor light intensity feature, L out is the outdoor light intensity feature, T is the indoor temperature feature, H is the indoor humidity feature, A user is the user behavior feature, CO is the carbon dioxide concentration feature, PM is the PM2.5 concentration feature, Ho sin is the hour feature, Mo sin is the month feature.

[0069] Step 3: Divide the indoor area, obtain the recommended light intensity for each partition through the data of each partition, identify the partitions that need lighting through the evaluation results of the environments of each partition, and at the same time perform light intensity uniformity analysis on the partitions that need lighting, and balance the partitions with large differences according to the situation of adjacent partitions;

[0070] The said Step 3 includes:

[0071] Step 301: Before starting the lighting system indoors, divide the indoor area. The division logic is to divide according to the indoor functional areas. Obtain the light intensity data, indoor temperature and humidity data, and behavior data in real time for each area. Each area forms a separate partition data set. Input the partition feature set formed by each partition data set into the light intensity prediction model to obtain the recommended light intensity for each partition respectively. Judge the recommended light intensity and the partition environment for each partition. The basis formula is as follows:

[0072] ΔL in-i =L in-i -y i

[0073] Among them, ΔL in-i represents the light intensity difference of the i-th partition, L in-i is the indoor light intensity of the i-th partition, y i is the recommended light intensity of the i-th partition, i is the partition retrieval variable, i ∈ N, 1 ≤ i ≤ n;

[0074] When ΔL in-i ≥ 0, the lighting system of the i-th partition is not started. When ΔL in-i < 0, the lighting system of the i-th partition is started and the light intensity of the lighting system is preset as the light intensity difference;

[0075] Step 302: Count the partitions that need to start the lighting system, and judge the light intensity difference between the partitions that need to start the lighting system. The basis formula is as follows:

[0076]

[0077] Among them, JFC is the mean square deviation of the light between the partitions that need to start the lighting system, y i is the recommended light intensity of the i-th partition that needs to start the lighting system, is the average value of the recommended light intensities of the partitions that need to start the lighting system, k is the number of partitions that need to start the lighting system, i is the partition retrieval variable of the partitions that need to start the lighting system, i ∈ N, 1 ≤ i ≤ k;

[0078] Step 303: Conduct an analysis of the indoor lighting uniformity. Set the threshold of JFC. When JFC exceeds the threshold, it means that the light difference between the partitions that need to start the lighting system indoors is large, and the recommended light intensity needs to be balanced. When JFC does not exceed the threshold, it means that the difference between the partitions that need to start the lighting system indoors is small, and there is no need to balance the light intensity. The logic of the light intensity balance is as follows:

[0079] Identify the two zones with the highest and lowest recommended light intensities, adjust the recommended light intensities of these two zones to the recommended light intensities of their adjacent zones respectively, and recalculate the light intensity mean square deviation. Determine whether the mean square deviation still exceeds the threshold. If it exceeds the threshold, identify the two zones with the highest and lowest light intensities again, and adjust the recommended light intensities of these two zones towards the adjacent zones until the light intensity mean square deviation is lower than the threshold, thereby obtaining the balanced recommended light intensity;

[0080] Adjust the light intensity difference again according to the balanced recommended light intensity. Each zone that needs to start the lighting system illuminates the interior according to the light intensity difference, while each zone that does not need to start the lighting system remains in the non-start state.

[0081] Step 4: During the continuous operation of the lighting system, analyze the influence of outdoor light intensity changes on indoor light intensity through the actual internal and external light difference. When the outdoor light intensity changes greatly, adjust the start-up situation and brightness of the lighting system in each zone in real time;

[0082] The said Step 4 includes:

[0083] Step 401: Under the state of the zones where the lighting system has been turned on, continuously collect the light intensity data, indoor temperature and humidity data, and behavior data of each zone, and make a judgment on the change of outdoor light intensity based on the light intensity data. The formula is as follows:

[0084]

[0085] Among them, ΔLC is the internal and external light difference, y i is the recommended light intensity of the i-th zone, L out is the outdoor light intensity, i is the zone number retrieval variable, i ∈ N, 1 ≤ i ≤ n;

[0086] Organize the internal and external light differences of each zone into an internal and external light difference data set ΔLC. The formula is as follows:

[0087] ΔLC = {Lc1, Lc2, Lc3,..., Lc i ,..., Lc n}

[0088] Among them, Lc i is the internal and external light difference at the i-th moment, i is the time retrieval variable, i ∈ n, 1 ≤ i ≤ n;

[0089] Step 402: Identify that k consecutive data within the indoor-outdoor light difference dataset ΔLC are greater than 0, and set the value of the consecutive change threshold. When k does not exceed the consecutive change threshold, it indicates that the outdoor light intensity has changed but not significantly. When k exceeds the consecutive change threshold, it indicates that the outdoor light intensity has changed significantly, and indoor brightness adjustment is required. The logic of the brightness adjustment is as follows:

[0090] Compare the recommended light intensity of each partition with the outdoor light intensity. For the partitions where the recommended light intensity is greater than the outdoor light, if the lighting system has not been started, start the lighting system of this area according to the recommended light intensity. If the lighting system has already been started, adjust all the partitions where the lighting system has been started according to the balanced light intensity obtained from the indoor light balance analysis;

[0091] Adjust the indoor lighting situation in real time according to the change of the outdoor light intensity.

[0092] By judging the change of the outdoor light intensity through the comparison of the indoor and outdoor light intensities, the recommended indoor light intensity can be adjusted in time to keep the indoor light intensity always near the comfortable environment for the human body.

[0093] Step 5: Compare the light intensities of each partition through the recommended light intensity, and shade the partitions with excessive indoor light intensity;

[0094] The said Step 5 includes: continuously collecting light intensity data, adjusting the partitions with high indoor light intensity, and calibrating it as the light intensity area adjustment. The logic of the light intensity area adjustment is as follows:

[0095] Compare the recommended light intensity obtained from the light intensity prediction model with the indoor light intensity. The formula is as follows:

[0096] ΔLLC = L in -β·y

[0097] Among them, ΔLLC is the light intensity comparison difference of this partition, L in is the indoor light intensity of this partition, y is the recommended light intensity of this partition, β is a coefficient set by the personnel. When ΔLLC>0, it means that the indoor light intensity of this partition is too high, then start the light-shielding curtain of this partition to adjust the indoor light intensity so that ΔLLC≤0. When ΔLLC≤0, no adjustment is made.

[0098] Step 6: Adjust the light intensity prediction model through the user's actual adjustment of the light intensity, identify the indoor manual light intensity and replace the recommended light intensity.

[0099] Step 6 includes: during the continuous operation of the lighting system, collecting user adjustment data, where the user adjustment data is the light intensity adjustment value, and the light intensity adjustment value is the indoor manual light intensity after the user manually adjusts the indoor light intensity. At the same time, calibrating the light intensity data, indoor temperature and humidity data, and behavior data at this time, preprocessing and feature extraction of the data at this time, setting the expansion range of the light intensity data and indoor temperature and humidity data, where the expansion range is within the upper and lower limits of the calibrated light intensity data and indoor temperature and humidity data floating. When the light intensity prediction model calculates the recommended light intensity again through the calibrated light intensity, indoor temperature and humidity data, and behavior data, replace the value of the recommended light intensity with the light intensity adjustment value.

[0100] The above formulas are all dimensionless and take their numerical calculations. The formulas are obtained by collecting a large amount of data for software simulation to obtain a formula that is closest to the actual situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0101] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. Those skilled in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in this article can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0102] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units. They can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0103] As described above, the above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in this application, and all of them should be covered by the protection scope of this application.

Claims

1. A smart home lighting control method based on scene adaptive adjustment, characterized in that, The specific steps include: Step 1: Obtain multiple sets of light intensity data, indoor temperature and humidity data, and personnel behavior data. Manually adjust the indoor light intensity to the comfortable light intensity for personnel in actuality, and synchronously record the comfortable light intensity data for personnel to form a model data set. Step 2: Divide the indoor area into functional areas to generate each sub-area, obtain the light intensity data, indoor temperature and humidity data, and personnel behavior data of each sub-area to form a partition data set. Step 3: Extract features from the model data set and the partition data set respectively to obtain indoor light intensity features, outdoor light intensity features, indoor temperature features, indoor humidity features, carbon dioxide concentration features, PM2.5 concentration features, hour features, and month features. Form a model feature set from the model data set and a partition feature set from the partition data set. Step 4: Input the obtained model feature set into a linear regression model, use the recommended light intensity as the dependent variable, and perform training to obtain a light intensity prediction model. Step 5: Input the partition feature set into the light intensity prediction model to obtain the recommended light intensity for each partition. By comparing the recommended light with the actual light intensity of the partition, count the partitions that need to start the lighting system, and balance the light intensity between partitions by analyzing the light intensity differences between partitions to obtain a balanced recommended light intensity. Step 6: According to the change of outdoor light intensity, compare the change of indoor recommended light intensity, and adjust the balanced recommended light intensity according to the comparison result. Step 7: Each partition continuously collects light intensity data, and shade the partitions where the outdoor light intensity is greater than the indoor recommended light intensity. Step 8: Obtain the data manually adjusted by the user during use and mark the light intensity data, indoor temperature and humidity data, and behavior data at the time of adjustment. In the light intensity prediction model, when the input features are the marked light intensity data, indoor temperature and humidity data, and behavior data, replace the recommended light intensity of the intensity prediction model with the value manually adjusted by the user.

2. The smart home lighting control method based on scene adaptive adjustment according to claim 1, wherein: Obtain light intensity data, where the light intensity data includes outdoor light intensity and indoor light intensity, obtain indoor temperature and humidity data, where the indoor temperature and humidity data includes temperature and humidity, obtain personnel behavior data, obtain indoor air quality data, where the indoor air quality data includes carbon dioxide concentration and PM2.5 concentration, the behavior data includes four categories of reading, dining, entertainment, and rest, the collection frequencies of the light intensity data, indoor temperature and humidity data, and personnel behavior data are the same, add time stamps and align them in time, preprocess the data and form a model data set, manually adjust the indoor light intensity to the comfortable light intensity for personnel in real time, and synchronously collect the indoor comfortable light intensity data with the model data set, and add the indoor comfortable light intensity data as a label to the model data set.

3. The smart home lighting control method based on scene adaptive adjustment according to claim 2, wherein: Extract features from the model dataset, perform one-hot encoding on the behavior data, convert each behavior category into ternary variables that are mutually distinct to obtain behavior features, normalize the light intensity data, indoor temperature and humidity data, and air quality data to obtain indoor light intensity features, outdoor light intensity features, indoor temperature features, indoor humidity features, carbon dioxide concentration features, and PM2.5 concentration features respectively, and extract hour features and month features from the timestamp. The formulas are as follows: Among them, Ho sin is the hour feature value, and Ho is the hour value in the timestamp. Mo sin is the month feature value, and Mo is the month value in the timestamp; Form a model feature set from the extracted features and input it into a linear regression analysis model to construct a light intensity prediction model. The formula is as follows: y = β0 + β1L in + β2L out + β3T + β4H + β5A user + β6CO + β7PM + β8Ho sin + β9Mo sin Among them, y is the recommended light intensity, β0 is the bias term representing a constant value, and β1, β2, β3, β4, β5, β6, β7, β8, β9 are model coefficients representing the influence degree of each input feature on the light adjustment, L in is the indoor light intensity feature, L out is the outdoor light intensity feature, T is the indoor temperature feature, H is the indoor humidity feature, A user is the user behavior feature, CO is the carbon dioxide concentration feature, PM is the PM2.5 concentration feature, Ho sin is the hour feature, Mo sin is the month feature.

4. The smart home lighting control method based on scene adaptive adjustment according to claim 3, characterized in that: Before the lighting system needs to be started indoors, divide the indoor area. The division logic is to divide it according to the indoor functional areas. Real-time obtain the light intensity data, indoor temperature and humidity data, and behavior data in each area. Each area forms a partition dataset separately. Input the partition feature set formed by each partition dataset into the light intensity prediction model to obtain the recommended light intensity for each partition respectively, and judge the recommended light intensity for each partition and the partition environment. The formula is as follows: ΔL in-i = L in-i - y i Among them, ΔL in-i represents the light intensity difference of the i-th partition, and L in-i is the indoor light intensity of the i-th partition, y i is the recommended light intensity of the i-th partition, and i is the partition retrieval variable, i ∈ N, 1 ≤ i ≤ n; When ΔL in-i ≥ 0, the lighting system of the i-th partition is not activated. When ΔL in-i < 0, the lighting system of the i-th partition is activated and the light intensity of the lighting system is preset to the light intensity difference; Count the partitions that need to start the lighting system and judge the light intensity difference between the partitions that need to start the lighting system. The formula is as follows: Among them, JFC is the mean square error of illumination between partitions where the lighting system needs to be started, and y i is the recommended illumination intensity of the i-th partition where the lighting system needs to be started, is the average value of the recommended illumination intensities of the partitions where the lighting system needs to be started, k is the number of partitions where the lighting system needs to be started, i is the partition retrieval variable of the partitions where the lighting system needs to be started, i ∈ N, 1 ≤ i ≤ k; Conduct an analysis of the indoor lighting balance, set the threshold of JFC. When JFC exceeds the threshold, it indicates that the light difference between the partitions that need to start the lighting system indoors is large and the recommended light intensity needs to be balanced. When JFC does not exceed the threshold, it indicates that the difference between the partitions that need to start the lighting system indoors is small and there is no need to balance the light intensity. The logic of the light intensity balance is as follows: Identify the two partitions with the highest and lowest recommended light intensities, adjust the recommended light intensities of these two partitions to the recommended light intensities of their adjacent partitions respectively, and calculate the light mean square error again. Judge whether the mean square error still exceeds the threshold. If it exceeds the threshold, identify the two partitions with the highest and lowest light intensities again and adjust the recommended light intensities of these two partitions towards the adjacent partitions until the light mean square error is lower than the threshold, thereby obtaining the balanced recommended light intensity; Adjust the light intensity difference again according to the balanced recommended light intensity. Each partition that needs to start the lighting system illuminates the indoor area according to the light intensity difference, while each partition that does not need to start the lighting system remains in the non-start state.

5. The smart home lighting control method based on scenario adaptive adjustment according to claim 4, wherein: Under the state of the partitions where the lighting system has been turned on, continuously collect the light intensity data, indoor temperature and humidity data, and behavior data of each partition, and make a judgment on the change of the outdoor light intensity according to the light intensity data. The formula is as follows: where ΔLC is the internal and external light difference, and y i is the recommended light intensity of the i-th zone, L out is the outdoor light intensity, i is the zoning number retrieval variable, i ∈ N, 1 ≤ i ≤ n; Organize the internal and external light differences of each partition into an internal and external light difference dataset ΔLC. The formula is as follows: ΔLC = {Lc1, Lc2, Lc3, …, Lc i , …, Lc n} Among them, Lc i is the internal and external light difference at the i-th moment, where i is the time retrieval variable, i ∈ N, and 1 ≤ i ≤ n; Identify that k consecutive data in the indoor and outdoor light difference dataset ΔLC are greater than 0, and set the continuous change threshold value. When k does not exceed the continuous change threshold, it indicates that the outdoor light intensity has changed but not much. When k exceeds the continuous change threshold, it indicates that the outdoor light intensity has changed greatly and the indoor brightness needs to be adjusted. The logic of the brightness adjustment is as follows: Compare the recommended light intensity for each zone with the outdoor light intensity. For zones where the recommended light intensity is greater than the outdoor light intensity, if the lighting system has not been started, start the lighting system for this area according to the recommended light intensity. If the lighting system has already been started, adjust the lighting intensity of all zones where the lighting system has been started according to the balanced light intensity obtained from the indoor light balance analysis. Adjust the indoor lighting situation in real time according to the change of outdoor light intensity.

6. The smart home lighting control method based on scene adaptive adjustment according to claim 5, wherein: Continuously collect light intensity data and adjust the zones with high indoor light intensity, which is designated as light intensity zone adjustment. The logic of the light intensity zone adjustment is as follows: Compare the recommended light intensity obtained from the light intensity prediction model with the indoor light intensity. The formula used is: ΔLLC = L in -β·y Among them, ΔLLC is the light intensity contrast difference of this partition, L in is the indoor light intensity of this partition, y is the recommended light intensity of this partition, β is a coefficient. When ΔLLC > 0, it indicates that the indoor light intensity of this partition is too high, then the light-shading curtain of this partition is activated to adjust the indoor light intensity so that ΔLLC ≤ 0. When ΔLLC ≤ 0, no adjustment is made.

7. The smart home lighting control method based on scenario adaptive adjustment according to claim 6, characterized in that: During the continuous operation of the lighting system, collect user adjustment data, which is the light intensity adjustment value. The light intensity adjustment value is the indoor manual light intensity after the user manually adjusts the indoor light intensity. At the same time, calibrate the light intensity data, indoor temperature and humidity data, and behavior data at this time, preprocess and extract features from the data at this time, and set the expansion range of the light intensity data and indoor temperature and humidity data. The expansion range is within the upper and lower limits of the floating of the calibrated light intensity data and indoor temperature and humidity data. When the light intensity prediction model calculates the recommended light intensity again through the calibrated light intensity, indoor temperature and humidity data, and behavior data, replace the value of the recommended light intensity with the light intensity adjustment value.

Images