Building thermal comfort and hvac linkage monitoring method based on behavior recognition

Abstract

The application discloses a building thermal comfort and HVAC linkage monitoring method based on behavior recognition, and relates to the technical field of intelligent buildings; an extreme AI intelligent recognition model is constructed; the extreme AI intelligent recognition model is used for intelligently recognizing an extreme behavior feature set and other behavior data; the other behavior data is subjected to behavior feature splitting to obtain a behavior split feature set; feature level grouping is performed to obtain a behavior feature level group; feature grouping is performed on the feature level group to generate a feature analysis triple; a first layer thermal comfort coefficient is obtained for a first evaluation layer; a second layer thermal comfort coefficient of a target scene area is obtained according to the extreme behavior feature set; and a two-layer coefficient set is formed together with the first layer thermal comfort coefficient; a thermal comfort feedback result is obtained according to a preset thermal comfort investigation two-dimensional code, a third layer coefficient is obtained, and a thermal comfort coefficient generation model is generated in association; and the thermal comfort coefficient of the target scene area is obtained according to the thermal comfort coefficient generation model, and the HVAC equipment is regulated and controlled.

Description

Technical Field

[0001] This invention relates to the field of intelligent building technology, specifically to a method for monitoring building thermal comfort and HVAC linkage based on behavior recognition. Background Technology

[0002] HVAC is an abbreviation for Heating, Ventilation and Air Conditioning. As the core of modern building environmental control, the application of HVAC systems has expanded from traditional temperature regulation to multiple dimensions such as health, energy saving, and intelligence. HVAC systems are widely used in buildings, industry, commerce, and transportation; system design must comprehensively consider energy saving, comfort, safety, and economy, and is deeply integrated with new technologies such as artificial intelligence and the Internet of Things, developing towards intelligent and customized solutions.

[0003] Currently, with the acceleration of urbanization and the popularization of green building concepts, the building sector has put forward higher requirements for the precision and intelligence of thermal comfort control. Traditional HVAC control methods are no longer able to adapt to the dynamic needs of complex scenarios such as large shopping malls and office buildings at different times, in different seasons, with different numbers of people gathered, and different people's temperature perception. They rely solely on a single subjective feedback, which leads to the irrationality of HVAC control.

[0004] While some systems combine behavior recognition for regulation, they typically only use AI to intelligently identify extreme behaviors, ignoring other limb behavior characteristics related to thermal sensation, such as weak limb behavior characteristics and routine limb behavior characteristics. This results in an incomplete and one-sided assessment of thermal comfort, further reducing the rationality of HVAC regulation.

[0005] Therefore, this invention provides a method for monitoring building thermal comfort and HVAC linkage based on behavior recognition. Summary of the Invention

[0006] The purpose of this invention is to provide a method for monitoring building thermal comfort and HVAC linkage based on behavior recognition, so as to solve the problems in the prior art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a method for monitoring building thermal comfort and HVAC linkage based on behavior recognition, the method comprising the following steps:

[0008] Construct an extreme AI intelligent recognition model; use the extreme AI intelligent recognition model to intelligently identify extreme behavioral feature sets and other behavioral data;

[0009] Other behavioral data are segmented into behavioral features to obtain behavioral feature sets; these behavioral feature sets are then grouped by feature level to obtain behavioral feature level groups; and feature level groups are then used to generate feature analysis triples.

[0010] The first-layer thermal comfort coefficient of the target scene area is obtained by using feature analysis triples as the first evaluation layer;

[0011] The second-layer thermal comfort coefficient of the target scene area is obtained based on the extreme behavior feature set; and a second-layer coefficient set is formed with the first-layer thermal comfort coefficient.

[0012] Thermal comfort feedback results are obtained by using a pre-set thermal comfort survey QR code. The feedback results are then filtered to obtain the optimal set of adopted feedback results. The third-level coefficients are then obtained based on the optimal set of adopted feedback results, and the third-level coefficients are associated with the second-level coefficient set to generate a thermal comfort coefficient generation model.

[0013] The thermal comfort coefficient of the target scene area is obtained based on the thermal comfort coefficient generation model, and the HVAC equipment is adjusted according to the thermal comfort coefficient.

[0014] Furthermore, an extreme behavior identification model is constructed; the process of identifying extreme behavior feature sets and other behavioral data through the extreme behavior identification model includes:

[0015] Set up environmental data acquisition nodes for the target scene area to collect regional environmental data for the target scene area in the current season; generate objective thermal sensation for the target scene area in the current season based on the regional environmental data, and set up a set of regular behavioral features corresponding to the objective thermal sensation;

[0016] Set up an extreme recognition behavior feature set corresponding to objective thermal sensation; and set up camera nodes in the target scene area to use AI-based intelligent recognition of the extreme behavior feature set corresponding to the extreme recognition behavior feature set.

[0017] Obtain the complement of the set of normal and extreme behavior features in the target scene region and all behavior feature sets in the target scene region, denoted as other behavior data; then construct an extreme behavior recognition model.

[0018] Furthermore, the process of obtaining a behavioral feature set by decomposing other behavioral data includes:

[0019] Based on thermal sensation, a split feature classification pool is set, and each split feature classification pool is connected to one or more corresponding feature recognition points; a split feature is set for the split feature classification pool, and the feature significance amplitude of the split feature is set; the split feature and the feature significance amplitude are correlated to form the significant split feature of the feature recognition point; each feature recognition point is connected to a significant split feature.

[0020] AI intelligently identifies other behavioral data based on obvious splitting features, and splits them into corresponding splitting feature classification pools to generate behavioral splitting feature sets; then, other unsplit behavioral data are removed.

[0021] Furthermore, the process of grouping the behavioral feature set into feature levels to obtain behavioral feature level groups, and then generating feature analysis triples from these feature level groups, includes:

[0022] Set corresponding feature visibility levels for each split feature classification pool, and set the feature visibility level corresponding to the feature visibility amplitude threshold range of obvious split features;

[0023] Based on the threshold range of feature significance, the split features corresponding to each feature recognition point in the split feature pool are grouped by feature significance level; thereby obtaining the behavioral feature level group for each feature significance level;

[0024] Set the feature weights for feature level groups; obtain the number of split features for each feature level group, denoted as the number of features for each level group; combine the feature level groups, the number of features for each level, and the feature weights to generate the feature analysis triplet corresponding to the split feature pool.

[0025] Furthermore, the process of obtaining the first-layer thermal comfort coefficient of the target scene region using feature analysis triples as the first evaluation layer includes:

[0026] The product of the number of rank features and the feature weights of the feature analysis triplet is denoted as the feature analysis coefficient of the rank group; then, the feature analysis coefficients of each split data pool are summed to obtain the data pool coefficient of the split data pool;

[0027] Set the characteristic standard coefficients for splitting the data pool; the characteristic standard coefficients include -1, 0 and +1; where "-" and "+" are used to represent the cold tendency value and hot tendency value corresponding to the split data pool, respectively; 0 represents the excessive tendency value; and then obtain the data pool standard coefficients of the data pool coefficients based on the characteristic standard coefficients;

[0028] The sum of the standard coefficients of each data pool that acquires other behavioral data is denoted as the first-level thermal comfort coefficient.

[0029] Furthermore, the process of obtaining the second-layer thermal comfort coefficient of the target scene area based on the extreme behavior feature set, and forming the second-layer coefficient set with the first-layer thermal comfort coefficient, includes:

[0030] The extreme behavior feature set is classified into extreme thermal sensation to obtain extreme heat feature datasets and extreme cold feature datasets; the quantity of corresponding extreme behavior features in the extreme heat feature dataset and extreme cold feature dataset is determined;

[0031] If the number of extreme behavioral features corresponding to the extreme heat feature dataset is greater than the number of extreme behavioral features corresponding to the extreme cold feature dataset, then the extreme heat sensation is marked as the feature standard coefficient + 1, and the difference in quantity is obtained. Then, it is multiplied with the feature standard coefficient to obtain the second layer of thermal comfort coefficient of the target scene area.

[0032] If the number of extreme behavioral features corresponding to the extreme heat feature dataset is less than the number of extreme behavioral features corresponding to the extreme cold feature dataset, then the extreme heat sensation is marked as feature standard coefficient -1, and the difference in quantity is obtained. Then, it is multiplied with the feature standard coefficient to obtain the second layer of thermal comfort coefficient.

[0033] Conversely, the thermal comfort coefficient of the second layer is 0;

[0034] The first-layer thermal comfort coefficient and the second-layer thermal comfort coefficient are integrated to form a two-layer coefficient set.

[0035] Furthermore, the process of obtaining thermal comfort feedback results based on a pre-set thermal comfort survey QR code, and then filtering the feedback results to obtain the optimal set of adopted feedback results includes:

[0036] Deploy thermal comfort survey QR codes at key nodes in the target scene area; set a survey cycle and conduct daily promotion in the target scene area according to the survey cycle; then automatically scan the code to select the corresponding thermal comfort result and generate thermal comfort feedback results.

[0037] The thermal comfort feedback results are classified to obtain the number of feedback results corresponding to each thermal comfort result. The feedback results are then sorted from largest to smallest to obtain a column of feedback numbers, which are then numbered and denoted as the i-th feedback number, where i = 1, 2, ... j, and j is a positive integer.

[0038] Set an excess percentage threshold, and perform the first comparison between the first feedback number in the feedback number column and the second feedback number based on the excess percentage threshold.

[0039] If the percentage of responses exceeding the second set exceeds the threshold, the thermal comfort feedback results corresponding to the first set of responses are retained and marked as the optimal set of adopted feedback results, while other thermal comfort feedback results are discarded; and the loop ends thereafter.

[0040] Conversely, the thermal comfort feedback results corresponding to the first and second feedback quantities are retained together and marked as the first set of adopted feedback results, while other thermal comfort feedback results are discarded; and the second and third feedback quantities are compared.

[0041] If the percentage exceeds the third feedback number, the thermal comfort feedback result corresponding to the third feedback number will be removed, and the first set of adopted feedback results will be retained; then the loop ends.

[0042] Conversely, the thermal comfort feedback results corresponding to the third feedback quantity are retained and added to the first adoption feedback result set to generate the second adoption feedback result set, while other thermal comfort feedback results are removed; the third feedback quantity is compared with the fourth feedback quantity; and so on, until the loop ends to obtain the i-th adoption feedback result set, which is denoted as the optimal adoption feedback result set.

[0043] Furthermore, the process of obtaining the third-level coefficients based on the optimal adoption feedback result set and associating the third-level coefficients with the second-level coefficient set to generate the thermal comfort coefficient generation model includes:

[0044] Set the comfort weights for the thermal comfort results, obtain the third-layer coefficients for the target scene region based on the comfort weights, and then associate the third-layer coefficients as judgment coefficients with the second-layer coefficient set to generate a thermal comfort coefficient generation model.

[0045] Furthermore, the process of obtaining the thermal comfort coefficient of the target scene area based on the thermal comfort coefficient generation model, and then adjusting the HVAC equipment based on the thermal comfort coefficient includes:

[0046] Set the threshold range for the third-layer coefficients, and the coefficient enhancement value within the threshold range; obtain the coefficient enhancement value for the third-layer coefficients based on the threshold range.

[0047] The sum of the first-level coefficient and the second-level coefficient is recorded as the first thermal comfort coefficient; the sum of the coefficient enhancement value and the first thermal comfort coefficient is recorded as the thermal comfort coefficient.

[0048] Set the basic control coefficient, then obtain the control parameter set of thermal comfort coefficient based on the basic control coefficient, and control the HVAC equipment according to the control parameters.

[0049] Compared with the prior art, the beneficial effects of the present invention are:

[0050] 1. Construct an extreme AI intelligent recognition model; use the extreme AI intelligent recognition model to intelligently identify extreme behavioral feature sets and other behavioral data; perform behavioral feature decomposition on other behavioral data to obtain behavioral feature sets; group the behavioral feature sets by feature level to obtain behavioral feature level groups; perform feature composition on feature level groups to generate feature analysis triplet groups; effectively improve the comprehensiveness of body behavior feature collection in complex scenarios and avoid single body behavior features.

[0051] 2. The first-layer thermal comfort coefficient of the target scene area is obtained using feature analysis triples as the first evaluation layer; the second-layer thermal comfort coefficient of the target scene area is obtained based on the extreme behavior feature set; and a second-layer coefficient set is formed with the first-layer thermal comfort coefficient; thermal comfort feedback results are obtained based on a preset thermal comfort survey QR code, and the feedback results are filtered to obtain the optimal adoption feedback result set; then, a third-layer coefficient is obtained based on the optimal adoption feedback result set, and the third-layer coefficient is associated with the second-layer coefficient set to generate a thermal comfort coefficient generation model; the thermal comfort coefficient of the target scene area is obtained based on the thermal comfort coefficient generation model, and the HVAC equipment is adjusted through the thermal comfort coefficient; effectively improving the irrationality of HVAC adjustment. Attached Figure Description

[0052] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0053] Figure 1 This is a flowchart of the building thermal comfort and HVAC linkage monitoring method based on behavior recognition according to the present invention.

[0054] Figure 2 The flowchart illustrates the construction of the extreme identification model for this invention.

[0055] Figure 3 This is a flowchart for generating feature analysis triples for this invention.

[0056] Figure 4 The flowchart illustrates the process of constructing a thermal comfort coefficient generation model for this invention. Detailed Implementation

[0057] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0058] Please see Figure 1 As shown, a method for monitoring building thermal comfort and HVAC linkage based on behavior recognition includes the following steps:

[0059] Step S1: Construct an extreme AI intelligent recognition model; use the extreme AI intelligent recognition model to intelligently identify extreme behavioral feature sets and other behavioral data;

[0060] Step S2: Perform behavioral feature segmentation on other behavioral data to obtain a behavioral segmentation feature set; group the behavioral segmentation feature set by feature level to obtain behavioral feature level groups; and generate feature analysis triplet groups from the feature level groups.

[0061] Step S3: Obtain the first-layer thermal comfort coefficient of the target scene area using the feature analysis triplet as the first evaluation layer;

[0062] Step S4: Obtain the second-layer thermal comfort coefficient of the target scene area based on the extreme behavior feature set; and form a second-layer coefficient set with the first-layer thermal comfort coefficient;

[0063] Step S5: Obtain thermal comfort feedback results based on the preset thermal comfort survey QR code, filter the feedback results to obtain the optimal set of adopted feedback results; then obtain the third-level coefficient based on the optimal set of adopted feedback results, and associate the third-level coefficient with the second-level coefficient set to generate a thermal comfort coefficient generation model;

[0064] Step S6: Obtain the thermal comfort coefficient of the target scene area based on the thermal comfort coefficient generation model, and adjust the HVAC equipment according to the thermal comfort coefficient.

[0065] Please see Figure 2 As shown, step S1 requires further refinement, including the construction of an extreme behavior identification model. The process of identifying extreme behavior feature sets and other behavioral data through the extreme behavior identification model includes:

[0066] Step S101: Set up environmental data acquisition nodes for the target scene area to collect regional environmental data of the target scene area corresponding to the current season; generate objective thermal sensation of the target scene area corresponding to the current season based on the regional environmental data, and set the routine behavioral feature set corresponding to the objective thermal sensation;

[0067] Step S102: Set the extreme recognition behavior feature set corresponding to the objective thermal sensation; and set the camera nodes in the target scene area for AI-based intelligent recognition of the extreme behavior feature set corresponding to the extreme recognition behavior feature set.

[0068] Step S103: Obtain the complement of the set of normal behavior features and extreme behavior features in the target scene area, which is denoted as other behavior data; and then construct an extreme behavior recognition model.

[0069] In the above embodiments, steps S101-S103 need further clarification. The target scene area includes, but is not limited to, a target scene area targeting a key area in a large scene (large shopping mall, office building, factory production space), a target scene area targeting a floor, a target scene area targeting a room, etc.; corresponding sensors are set at the environmental data acquisition nodes, and corresponding camera devices are set at the camera nodes; thermal sensation refers to the subjective perception of "cold" or "hot" in the surrounding environment, mainly driven by changes in skin temperature, reflecting the instantaneous intensity of cold and hot stimulation; including cold, cool, slightly cool, neutral, slightly warm, warm, and hot; objective thermal sensation is used to represent the objective thermal sensation of the current environmental temperature, which has universal applicability to the objective cold and hot state of the environment and does not depend on individual subjective feelings; it is calculated by combining data such as temperature, humidity, wind speed, and average radiant temperature collected by the environmental data acquisition nodes with the PMV thermal comfort model; the conventional behavioral feature set is used for This refers to routine behavioral characteristics under objective thermal sensation, where it's impossible to discern whether these characteristics indicate a clear intention to regulate body temperature. It's important to further clarify that even with identical objective thermal sensations in the same region across different seasons, the corresponding sets of routine behavioral characteristics can differ significantly. For example, when the objective thermal sensation is consistently "neutral and comfortable" (corresponding to environmental data: temperature 24℃, relative humidity 50%, wind speed 0.2m / s), in summer, the routine behavioral characteristics include wearing short-sleeved clothing, naturally relaxed limbs, occasionally raising a hand to adjust hair, and a relaxed sitting posture with arms naturally placed on a table. In winter, the routine behavioral characteristics include wearing a thick coat or sweater, hands naturally placed in pockets or on a table, slightly tightened limbs (without a clear intention to keep warm), and an upright sitting posture with arms close to the torso. These differences primarily stem from variations in clothing worn in different seasons and the varying degrees of adaptation of the human body to the same objective thermal environment, leading to significant differences in routine behavioral manifestations. Furthermore, the extreme behavioral feature set is used to represent extreme behavioral characteristics exhibited by the human body that deviate from or exceed the normal response level of objective thermal sensation corresponding to regional environmental data. For example, when the objective thermal sensation is "neutral and comfortable," extreme behavioral data of overheating or overcooling may occur; such as frequent and continuous wiping of sweat, repeatedly pulling clothing to dissipate heat, frequently fanning oneself with objects, or exhibiting extreme heat dissipation behaviors; or exhibiting violent curling of the body, continuous rubbing of hands and arms, frequent stomping of feet to warm up, tightly hugging the body with arms, obvious trembling, and actively and repeatedly approaching heat sources. The entire behavioral feature set is used to represent all thermal sensation-related behavioral characteristics of the target scene area collected by the camera node.

[0070] Please see Figure 3 As shown, step S2 needs further clarification. The process of obtaining a behavior segmentation feature set by splitting other behavior data includes:

[0071] Step S201: Set up a split feature classification pool based on thermal sensation, wherein each split feature classification pool is connected to one or more corresponding feature recognition points;

[0072] Step S202: The feature recognition point corresponds to a clearly split feature;

[0073] Step S203: AI intelligently identify other behavioral data based on the obvious splitting features, and split them into the corresponding splitting feature classification pool to generate a behavioral splitting feature set; then remove other unsplit behavioral data.

[0074] The process for obtaining the obvious splitting features in step S202 includes:

[0075] Step S2021: Set the splitting features of the splitting feature classification pool, and set the feature significance of the splitting features;

[0076] Step S2022: Associate the splitting features with the obvious amplitude of the features to form obvious splitting features of the feature recognition points;

[0077] In the above embodiments, steps S201-S203 need to be further defined as follows: each of the splitting feature classification pools corresponds to a type of splitting feature with the same core limb behavior attributes and related to thermal sensation recognition, used to achieve accurate splitting and classification of other behavioral data with weak cold-heat correlation and only involving limb movements; furthermore, the splitting feature classification pool set according to thermal sensation is a feature lower than the behavioral feature level group; it can be divided into limb heat dissipation tendency type splitting feature classification pool, limb slight heat dissipation tendency type splitting feature classification pool, limb warmth retention tendency type splitting feature classification pool, limb slight warmth retention tendency type splitting feature classification pool, and limb thermal adaptation related type splitting feature classification pool;

[0078] For example, the classification pool of minor heat dissipation tendency of limbs is used to include the classification features related to minor heat dissipation of the human body that do not reach the level of extreme heat dissipation behavior, which indirectly reflects the minor heat dissipation tendency of the human body. The corresponding classification features are set as slightly rolled up sleeves (not fully rolled up), slightly open collar (not fully open), and slightly extended arms (the range is less than that of extreme heat dissipation actions), etc. The obvious range of the feature can be set as slight movement range, no continuous repetition, and not reaching the extreme reaction standard. The corresponding obvious classification features are "3.5cm≤sleeve roll-up range≤6cm, 1.5s≤s collar adjustment single duration≤2s, 6cm<arm extension range<10cm".

[0079] It should be further explained that the process of obtaining behavioral feature level groups by grouping the behavioral feature set into feature levels, and generating feature analysis triples from the feature level groups includes:

[0080] Step S204: Set the corresponding feature visibility level for each split feature classification pool, and set the feature visibility level corresponding to the feature visibility amplitude threshold range of obvious split features;

[0081] Step S205: Based on the threshold range of feature significance, group the split features of each feature recognition point in the split feature pool according to the feature significance level; and then obtain the behavioral feature level group for each feature significance level.

[0082] Step S206: Set the feature weights of the feature level groups; obtain the number of split features for each feature level group, and record it as the number of features in the level group;

[0083] Step S207: Construct the feature level group, the number of level features, and the feature weights to generate the feature analysis triplet corresponding to the split feature pool, denoted as (feature level group, number of level features, feature weights).

[0084] In the above embodiments, steps S204-S207 need to be further defined so that each splitting feature in the feature level group of the feature analysis triplet is connected one-to-one with the corresponding splitting feature classification pool; in preparation for subsequent analysis.

[0085] Step S3 needs further clarification. The process of obtaining the first-layer thermal comfort coefficient of the target scene region using the feature analysis triplet as the first evaluation layer includes:

[0086] Step S301: Obtain the product of the number of rank features and the feature weights of the feature analysis triplet, and record it as the feature analysis coefficient of the feature rank group; then obtain the feature analysis coefficients of each split data pool and sum them to obtain the data pool coefficients of the split data pool;

[0087] Step S203: Set the characteristic standard coefficients for splitting the data pool; the characteristic standard coefficients include -1, 0 and +1; where "-" and "+" are used to represent the cold tendency value and hot tendency value corresponding to the split data pool, respectively; 0 represents the excessive tendency value; and then obtain the data pool standard coefficients of the data pool coefficients based on the characteristic standard coefficients.

[0088] Step S204: Obtain the sum of the standard coefficients of each data pool of other behavioral data and record it as the first-level thermal comfort coefficient;

[0089] It should be further explained that steps S301-S304 require further clarification. The feature standard coefficient is used to represent the thermal sensation tendency corresponding to the split feature classification pool, and is a benchmark coefficient for quantifying the direction of cold and heat correlation. "+1" corresponds to thermal tendency, mainly related to the data pool related to heat dissipation, such as the limb heat dissipation tendency split feature classification pool and the limb slight heat dissipation tendency split feature classification pool; "-1" corresponds to cold tendency, mainly related to the data pool related to warmth, such as the limb warmth retention tendency split feature classification pool and the limb slight warmth retention tendency split feature classification pool; and 0 represents the limb thermal adaptation related split feature classification pool. Since there is no clear cold and heat tendency, its feature standard coefficient is set to 0, which is used to distinguish the thermal sensation correlation direction of different split pools and provide a benchmark for the calculation of the data pool standard coefficient.

[0090] Step S4 requires further refinement. The process of obtaining the second-layer thermal comfort coefficient of the target scene area based on the extreme behavior feature set and forming the second-layer coefficient set with the first-layer thermal comfort coefficient includes:

[0091] Step S401: Classify the extreme behavior feature set according to extreme heat sensation to obtain the extreme heat feature dataset and the extreme cold feature dataset; determine the quantity of corresponding extreme behavior features in the extreme heat feature dataset and the extreme cold feature dataset;

[0092] Step S402: If the number of extreme behavioral features corresponding to the extreme heat feature dataset is greater than the number of extreme behavioral features corresponding to the extreme cold feature dataset, then mark the extreme heat sensation as feature standard coefficient + 1, obtain the difference in quantity, and then multiply it with the feature standard coefficient to obtain the second layer thermal comfort coefficient of the target scene area.

[0093] Step S403: If the number of extreme behavioral features corresponding to the extreme heat feature dataset is less than the number of extreme behavioral features corresponding to the extreme cold feature dataset, then mark the extreme heat sensation as feature standard coefficient -1, obtain the difference in quantity, and then multiply it with the feature standard coefficient to obtain the second layer of thermal comfort coefficient.

[0094] Conversely, the thermal comfort coefficient of the second layer is 0;

[0095] The first-layer thermal comfort coefficient and the second-layer thermal comfort coefficient are integrated to form a two-layer coefficient set.

[0096] In the above embodiments, steps S401-S403 need further clarification. The extreme thermal sensation includes extreme heat and extreme cold. Extreme behavioral features, whether extreme heat or extreme cold, correspond to extreme reactions in the target scene area, and can therefore be considered to have the same weight. Thus, the feature standard coefficient is determined by the number of extreme heat or extreme cold events. For example, after classifying the extreme behavioral feature set, if the number of extreme behavioral features in the extreme heat feature dataset is 6 and the number in the extreme cold feature dataset is 3, then the corresponding feature standard coefficient is +1; the difference in number is 6 − 3 = 3, so the second-layer thermal comfort coefficient is 3 × (+1) = +3. If the number of extreme heat behaviors is 2 and the number of extreme cold behaviors is 9, then the corresponding feature standard coefficient is −1, the difference in number is 7, and the second-layer thermal comfort coefficient is 7 × (−1) = −7.

[0097] Please see Figure 4 As shown, step S5 requires further refinement. The process of obtaining thermal comfort feedback results based on a pre-defined thermal comfort survey QR code, filtering the feedback results to obtain the optimal set of adopted feedback results, and then obtaining the third-level coefficients based on the optimal set of adopted feedback results, and finally associating the third-level coefficients with the second-level coefficient set to generate the thermal comfort coefficient generation model, includes:

[0098] Step S501: The thermal comfort results set by the thermal comfort survey QR code include very comfortable, comfortable, somewhat comfortable, no feeling, uncomfortable, and very uncomfortable;

[0099] Step S502: Deploy the thermal comfort survey QR code at key nodes in the target scene area; set the survey period and conduct daily promotion in the target scene area according to the survey period; then the QR code can be automatically scanned to select the corresponding thermal comfort result and generate thermal comfort feedback results.

[0100] Step S503: Classify the thermal comfort feedback results to obtain the feedback quantity corresponding to each thermal comfort result, and sort them from largest to smallest according to the feedback quantity to obtain the feedback quantity column, and number them as the i-th feedback quantity, where i=1,2,...j, and j takes the value of a positive integer;

[0101] Step S504: Set the percentage threshold, and make the first comparison between the first feedback number in the feedback number column and the second feedback number based on the percentage threshold.

[0102] Step S505: If the percentage of the number of feedbacks exceeding the second feedback quantity exceeds the percentage threshold, then the thermal comfort feedback results corresponding to the first feedback quantity are retained and marked as the optimal adoption feedback result set, and other thermal comfort feedback results are discarded; then the loop ends.

[0103] Step S506: Conversely, retain the thermal comfort feedback results corresponding to the first and second feedback quantities respectively, and mark them as the first set of adopted feedback results, and discard other thermal comfort feedback results; and compare the second and third feedback quantities.

[0104] Step S507: If the percentage threshold for the number of third feedbacks is exceeded, the thermal comfort feedback results corresponding to the number of third feedbacks will be removed, and the set of first-time adopted feedback results will be retained; then the loop ends.

[0105] Step S508: Conversely, retain the thermal comfort feedback results corresponding to the third feedback quantity and add them to the first adoption feedback result set to generate the second adoption feedback result set, and remove other thermal comfort feedback results; compare the third feedback quantity with the fourth feedback quantity; and so on, until the loop ends to obtain the i-th adoption feedback result set, which is denoted as the optimal adoption feedback result set;

[0106] Step S509: Set the comfort weight of the thermal comfort result, and obtain the third-level coefficient of the target scene area of ​​the optimal adoption feedback result set according to the comfort weight; then associate the third-level coefficient as the judgment coefficient with the second-level coefficient set to generate a thermal comfort coefficient generation model;

[0107] It needs further clarification that steps S501-S509 require further clarification. The percentage threshold is a pre-set proportional threshold used to determine whether the difference between adjacent feedback quantities is sufficiently significant to decide whether to continue including them in the subsequent feedback result set. For example, it can be set to 60%, meaning the current feedback quantity must exceed 60% of the next feedback quantity to be considered as having a significant advantage. The process of setting the comfort weight includes: setting the comfort weight to decrease progressively with the subjective comfort level; the higher the comfort level, the larger the comfort weight value, and the lower the comfort level, the smaller the comfort weight value. The comfort weight sign is consistent with the hot / cold tendency of the two thermal comfort coefficients: assigning positive values ​​to the more comfortable range and negative values ​​to the more uncomfortable range; no sensation corresponds to a weight of 0, ensuring that the subjective feedback result and the behavior recognition result are consistent in numerical meaning. For example: very comfortable corresponds to 2.0, comfortable corresponds to 1.8, fairly comfortable corresponds to 0.6, no sensation corresponds to 0, uncomfortable corresponds to -0.9, and very uncomfortable corresponds to -1.5.

[0108] Step S6 requires further clarification. The process of obtaining the thermal comfort coefficient of the target scene area based on the thermal comfort coefficient generation model, and then adjusting the HVAC equipment using the thermal comfort coefficient, includes:

[0109] Step S601: Set the threshold range of the third-layer coefficients and the coefficient enhancement value of the threshold range of the third-layer coefficients; obtain the coefficient enhancement value of the third-layer coefficients based on the threshold range of the third-layer coefficients;

[0110] Step S602: Obtain the sum of the first-level coefficient and the second-level coefficient, and record it as the first thermal comfort coefficient; obtain the sum of the coefficient enhancement value and the first thermal comfort coefficient, and record it as the thermal comfort coefficient;

[0111] Step S603: Set the basic control coefficient, then obtain the control parameter set of thermal comfort coefficient based on the basic control coefficient, and control the HVAC equipment according to the control parameters.

[0112] Further explanation is needed regarding steps S601-S603; the aforementioned basic control coefficient represents the unit control coefficient of each control parameter of the HVAC equipment based on the thermal comfort coefficient. Specifically, it can be set according to the type and area of ​​the target scene area, the current season, and the rated power of the HVAC equipment. The ratio of the thermal comfort coefficient to the basic control coefficient is then recorded as the number of adjustments to the basic control coefficient. The number of adjustments is rounded up, with any fraction less than one adjustment counted as one. The HVAC control parameters, basic control coefficient, and number of adjustments are then considered. Control includes upward and downward adjustments, specifically determined based on the type, area, and current season of the target scene area. If the thermal comfort coefficient is positive, it indicates that heat dissipation is needed, and the control parameters should be adjusted to heat dissipation mode. If the thermal comfort coefficient is negative, it indicates that warmth is needed, and the control parameters should be adjusted to heating mode.

[0113] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for monitoring building thermal comfort and HVAC linkage based on behavior recognition, characterized in that, The method includes the following steps: Construct an extreme AI intelligent recognition model; use the extreme AI intelligent recognition model to intelligently identify extreme behavioral feature sets and other behavioral data; Other behavioral data are segmented into behavioral features to obtain behavioral feature sets; these behavioral feature sets are then grouped by feature level to obtain behavioral feature level groups; and feature level groups are then used to generate feature analysis triples. The first-layer thermal comfort coefficient of the target scene area is obtained by using feature analysis triples as the first evaluation layer; The second-layer thermal comfort coefficient of the target scene area is obtained based on the extreme behavior feature set; and a second-layer coefficient set is formed with the first-layer thermal comfort coefficient. Thermal comfort feedback results are obtained by using a pre-set thermal comfort survey QR code. The feedback results are then filtered to obtain the optimal set of adopted feedback results. The third-level coefficients are then obtained based on the optimal set of adopted feedback results, and the third-level coefficients are associated with the second-level coefficient set to generate a thermal comfort coefficient generation model. The thermal comfort coefficient of the target scene area is obtained based on the thermal comfort coefficient generation model, and the HVAC equipment is adjusted according to the thermal comfort coefficient.

2. The building thermal comfort and HVAC linkage monitoring method based on behavior recognition according to claim 1, characterized in that, Construct an extreme identification model; The process of identifying extreme behavioral feature sets and other behavioral data through extreme identification models includes: Set up environmental data acquisition nodes for the target scene area to collect regional environmental data for the target scene area in the current season; generate objective thermal sensation for the target scene area in the current season based on the regional environmental data, and set up a set of regular behavioral features corresponding to the objective thermal sensation; Set up an extreme recognition behavior feature set corresponding to objective thermal sensation; and set up camera nodes in the target scene area to use AI-based intelligent recognition of the extreme behavior feature set corresponding to the extreme recognition behavior feature set. Obtain the complement of the set of normal and extreme behavior features in the target scene region and all behavior feature sets in the target scene region, denoted as other behavior data; then construct an extreme behavior recognition model.

3. The building thermal comfort and HVAC linkage monitoring method based on behavior recognition according to claim 2, characterized in that, The process of obtaining a behavioral feature set by splitting other behavioral data includes: Based on thermal sensation, a split feature classification pool is set, and each split feature classification pool is connected to one or more corresponding feature recognition points; a split feature is set for the split feature classification pool, and the feature significance amplitude of the split feature is set; the split feature and the feature significance amplitude are correlated to form the significant split feature of the feature recognition point; each feature recognition point is connected to a significant split feature. AI intelligently identifies other behavioral data based on obvious splitting features, and splits them into corresponding splitting feature classification pools to generate behavioral splitting feature sets; then, other unsplit behavioral data are removed.

4. The building thermal comfort and HVAC linkage monitoring method based on behavior recognition according to claim 3, characterized in that, The behavioral feature set is grouped by feature level to obtain behavioral feature level groups; The process of generating feature analysis triples from feature level groups includes: Set corresponding feature visibility levels for each split feature classification pool, and set the feature visibility level corresponding to the feature visibility amplitude threshold range of obvious split features; Based on the threshold range of feature significance, the split features corresponding to each feature recognition point in the split feature pool are grouped by feature significance level; thereby obtaining the behavioral feature level group for each feature significance level; Set the feature weights for feature level groups; obtain the number of split features for each feature level group, denoted as the number of features for each level group; combine the feature level groups, the number of features for each level, and the feature weights to generate the feature analysis triplet corresponding to the split feature pool.

5. The building thermal comfort and HVAC linkage monitoring method based on behavior recognition according to claim 4, characterized in that, The process of obtaining the first-layer thermal comfort coefficient of the target scene region using feature analysis triples as the first evaluation layer includes: The product of the number of rank features and the feature weights of the feature analysis triplet is denoted as the feature analysis coefficient of the rank group; then, the feature analysis coefficients of each split data pool are summed to obtain the data pool coefficient of the split data pool; Set the characteristic standard coefficients for splitting the data pool; the characteristic standard coefficients include -1, 0 and +1; where "-" and "+" are used to represent the cold tendency value and hot tendency value corresponding to the split data pool, respectively; 0 represents the excessive tendency value; and then obtain the data pool standard coefficients of the data pool coefficients based on the characteristic standard coefficients. The sum of the standard coefficients of each data pool that acquires other behavioral data is denoted as the first-level thermal comfort coefficient.

6. The building thermal comfort and HVAC linkage monitoring method based on behavior recognition according to claim 5, characterized in that, The second-layer thermal comfort coefficient of the target scene area is obtained based on the extreme behavior feature set; The process of forming a second-layer coefficient set with the first-layer thermal comfort coefficient includes: The extreme behavior feature set is classified into extreme thermal sensation to obtain extreme heat feature datasets and extreme cold feature datasets; the quantity of corresponding extreme behavior features in the extreme heat feature dataset and extreme cold feature dataset is determined; If the number of extreme behavioral features corresponding to the extreme heat feature dataset is greater than the number of extreme behavioral features corresponding to the extreme cold feature dataset, then the extreme heat sensation is marked as the feature standard coefficient + 1, and the difference in quantity is obtained. Then, it is multiplied with the feature standard coefficient to obtain the second layer of thermal comfort coefficient of the target scene area. If the number of extreme behavioral features corresponding to the extreme heat feature dataset is less than the number of extreme behavioral features corresponding to the extreme cold feature dataset, then the extreme heat sensation is marked as feature standard coefficient -1, and the difference in quantity is obtained. Then, it is multiplied with the feature standard coefficient to obtain the second layer of thermal comfort coefficient. Conversely, the thermal comfort coefficient of the second layer is 0; The first-layer thermal comfort coefficient and the second-layer thermal comfort coefficient are integrated to form a two-layer coefficient set.

7. The building thermal comfort and HVAC linkage monitoring method based on behavior recognition according to claim 6, characterized in that, The process of obtaining thermal comfort feedback results based on a pre-set thermal comfort survey QR code, and then filtering the feedback results to obtain the optimal set of adopted feedback results includes: Deploy thermal comfort survey QR codes at key nodes in the target scene area; set a survey cycle and conduct daily promotion in the target scene area according to the survey cycle; then automatically scan the code to select the corresponding thermal comfort result and generate thermal comfort feedback results. The thermal comfort feedback results are classified to obtain the number of feedbacks corresponding to each thermal comfort result. The feedbacks are then sorted from largest to smallest to obtain a column of feedbacks, which is then numbered and denoted as the i-th feedback number, where i = 1, 2, ... j, and j is a positive integer. Set an excess percentage threshold, and perform the first comparison between the first feedback number in the feedback number column and the second feedback number based on the excess percentage threshold. If the percentage of responses exceeding the second set exceeds the threshold, the thermal comfort feedback results corresponding to the first set of responses are retained and marked as the optimal set of adopted feedback results, while other thermal comfort feedback results are discarded; and the loop ends thereafter. Conversely, the thermal comfort feedback results corresponding to the first and second feedback quantities are retained and marked as the first set of adopted feedback results, while other thermal comfort feedback results are discarded; and the second and third feedback quantities are compared. If the percentage exceeds the third feedback number, the thermal comfort feedback result corresponding to the third feedback number will be removed, and the first set of adopted feedback results will be retained; then the loop ends. Conversely, the thermal comfort feedback results corresponding to the third feedback quantity are retained and added to the first adoption feedback result set to generate the second adoption feedback result set, while other thermal comfort feedback results are removed; the third feedback quantity is compared with the fourth feedback quantity; and so on, until the loop ends to obtain the i-th adoption feedback result set, which is denoted as the optimal adoption feedback result set.

8. The building thermal comfort and HVAC linkage monitoring method based on behavior recognition according to claim 7, characterized in that, The process of obtaining the third-level coefficients based on the optimal adoption feedback result set and associating the third-level coefficients with the second-level coefficient set to generate the thermal comfort coefficient generation model includes: Set the comfort weights for the thermal comfort results, obtain the third-layer coefficients for the target scene region based on the comfort weights, and then associate the third-layer coefficients as judgment coefficients with the second-layer coefficient set to generate a thermal comfort coefficient generation model.

9. The building thermal comfort and HVAC linkage monitoring method based on behavior recognition according to claim 8, characterized in that, The process of obtaining the thermal comfort coefficient of the target scene area based on the thermal comfort coefficient generation model, and then adjusting the HVAC equipment based on the thermal comfort coefficient includes: Set the threshold range for the third-layer coefficients, and the coefficient enhancement value within the threshold range; obtain the coefficient enhancement value for the third-layer coefficients based on the threshold range. The sum of the first-level coefficient and the second-level coefficient is recorded as the first thermal comfort coefficient; the sum of the coefficient enhancement value and the first thermal comfort coefficient is recorded as the thermal comfort coefficient. Set the basic control coefficient, then obtain the control parameter set of thermal comfort coefficient based on the basic control coefficient, and control the HVAC equipment according to the control parameters.

Images