A smart home environment management system based on data analysis

Abstract

The present invention discloses a smart home environment management system based on data analysis, which relates to the field of smart homes and includes: a data acquisition module, a data analysis module, and a control module; the data acquisition module acquires travel behavior data of each family member; acquires home environment data of each family member at each time when the family member is at home; the data analysis module predicts the predicted return time of each family member based on the travel behavior data; analyzes the suitable environmental parameter range for each family member based on the home environment data; and the control module determines the target home environment parameter range for each time period based on the predicted return time of each family member and the suitable environmental parameter range, controls the operation of each household appliance, and adjusts the indoor environment. The present invention can more intelligently manage the home environment and comprehensively improve the experience of each family member.

Description

Technical Field

[0001] The present invention relates to the field of smart home technology, and more particularly to a smart home environment management system based on data analysis. Background Art

[0002] A smart home is a system that uses the home as a platform and utilizes integrated wiring, network communications, and automated control technologies to integrate various devices and systems within the home, thereby enhancing the intelligence, comfort, and safety of home life. The environmental management system is a core component of a smart home, automatically monitoring and controlling various environmental indicators in the home, such as temperature, humidity, and air quality, to ensure an optimal home environment.

[0003] However, existing smart home environment management systems often only control individual appliances according to different preset modes based on sensor-monitored parameters, user-set parameters, and time. However, this control method is relatively fixed for the home environment and ignores the different comfort levels experienced by different family members in response to different environmental parameters, resulting in some family members being in an uncomfortable home environment. Furthermore, existing smart home environment management systems activate control only after detecting a person, which prevents people from immediately creating a comfortable home environment after returning home. Constantly maintaining the indoor environment increases unnecessary energy consumption.

[0004] Therefore, how to manage the home environment more intelligently is an issue that those skilled in the art urgently need to solve. Summary of the Invention

[0005] In view of this, the present invention provides an intelligent home environment management system based on data analysis. By analyzing the travel data and home environment data of each family member, it predicts the time for each member to return home and the appropriate environmental parameter range, and intelligently controls the home environment.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] The present invention discloses a smart home environment management system based on data analysis, comprising: a data acquisition module, a data analysis module and a control module;

[0008] The data acquisition module acquires the travel behavior data of each family member; acquires the home environment data of each family member at each time when they are at home;

[0009] The data analysis module predicts the predicted return home time of each family member based on the travel behavior data; and analyzes the suitable environmental parameter range of each family member based on the home environment data;

[0010] The control module determines the target home environment parameter range for each time period based on the predicted return time of each family member and the suitable environmental parameter range, controls the operation of each household appliance, and adjusts the indoor environment.

[0011] Furthermore, the data acquisition module obtains the travel displacement data of each family member through the mobile terminal worn by each family member; determines the time of departure and return of each family member through the face recognition device installed at the door and the door opening and closing sensor on the door, and compiles the departure time series and return time series of each family member;

[0012] The travel behavior data includes: the travel displacement data, the outgoing time series and the returning home time series;

[0013] The data acquisition module detects the temperature, humidity, carbon dioxide concentration and particulate matter concentration of the indoor air at each moment through a temperature sensor, a humidity sensor, a carbon dioxide concentration sensor and a laser particle counter to obtain the home environment data.

[0014] Furthermore, the data analysis module includes a home time prediction unit, which trains a corresponding home time prediction model based on each member's historical travel displacement data and each member's historical departure time series and return time series;

[0015] The home time prediction unit predicts the home time of each family member based on the travel displacement data after the current family member's trip, as well as the departure time sequence and home time sequence of the set time length before the current moment, using the trained home time prediction model of each family member.

[0016] Furthermore, the homecoming time prediction model is a TreeCNN-LSTM network model, including a TreeCNN network, a first LSTM network, a second LSTM network, and a fully connected network;

[0017] The TreeCNN network performs convolution and feature extraction on the trajectory tree in the input sample, and outputs a feature sequence to the first LSTM network in the time order of the features;

[0018] The first LSTM network outputs a first predicted time according to the input feature sequence;

[0019] The second LSTM network outputs a second predicted time based on the input outgoing time sequence and returning home time sequence;

[0020] The fully connected network performs weighted calculation on the first prediction time and the second prediction time, and outputs a final prediction time.

[0021] Furthermore, the TreeCNN network includes an input layer, a tree convolution layer, and a fully connected layer connected in sequence; the input layer is used to input the trajectory tree of a family member; the tree convolution layer convolves each data node in the trajectory tree to obtain a feature representation of each data node in the trajectory tree; the fully connected layer generates the feature sequence based on the time information of the feature representation;

[0022] The first LSTM network and the second LSTM network include an input layer, an LSTM layer, a fully connected layer, and an output layer connected in sequence; the input layer is used to input the feature sequence, or the departure time sequence and the return time sequence; the LSTM layer processes the input sequence and extracts features; the fully connected layer maps the extracted features to the final predicted return time, and the output layer outputs the results to the fully connected network;

[0023] The fully connected network maps the predicted home time of the first LSTM network and the second LSTM network to obtain a final predicted home time.

[0024] Furthermore, the trajectory tree is obtained by:

[0025] The travel displacement data of a family member's trip is sampled to obtain three-dimensional tensor data, where the first dimension represents the sampling time, the second dimension represents the two-dimensional plane coordinates of the family member at the sampling time, and the third dimension is the speed of the family member at the sampling time;

[0026] A gridded plane coordinate system is used to divide each slice into grids according to the second dimension of each slice in the three-dimensional tensor data, and the slice density of each grid is calculated;

[0027] According to the set density threshold, key grids are selected from all grids as child nodes of the trajectory tree;

[0028] Taking home as the root node of the trajectory tree, the connection relationship of each node is determined based on whether there is a slice with a sampling time difference less than a set threshold between two nodes, and the nodes are connected in time sequence to obtain the trajectory tree.

[0029] Furthermore, the data analysis module further includes a suitable environment parameter analysis unit, which performs sampling based on the home environment data and the time period of each family member at home according to a set step length to obtain a plurality of environmental parameter point data of each family member, and then obtain an environmental parameter point set of each member;

[0030] Eliminate points in the set of environmental parameter points that have been manually controlled within a set time period before the sampling moment;

[0031] Perform Kmeans clustering on the set of environmental parameter points after the elimination operation, use the k cluster centers as the key points of the environmental parameter point set, and use the maximum and minimum values ​​of air temperature, humidity, carbon dioxide concentration, and particulate matter concentration represented by all key points to determine the appropriate environmental parameter range for the family members corresponding to the environmental parameter point set;

[0032] Eliminate and cluster each set of environmental parameter points one by one to determine the appropriate environmental parameter range for each family member.

[0033] Furthermore, determining the target home environment parameter range for each time period specifically includes:

[0034] Step 1: Based on the predicted return time of each family member, determine the family members who will be indoors at each future time period;

[0035] Step 2: Initialize the proportional coefficient p and the amplification coefficient of each family member;

[0036] Step 3: Determine whether the suitable environmental parameter ranges of the family members determined in step 1 have an intersection. If so, use the intersection as the target home environmental parameter range; otherwise, proceed to step 4.

[0037] Step 4: p=p+0.01, amplify each of the suitable environmental parameter ranges according to the proportional coefficient and the amplification coefficient, and return to step 3.

[0038] Furthermore, the range of each suitable environmental parameter is enlarged, and the formula is:

[0039]

[0040] in, denote the upper and lower bounds of the jth environmental parameter of the i-th family member, respectively; j = 1, 2, 3, 4 denote air temperature, humidity, carbon dioxide concentration, and particulate matter concentration, respectively; q i is the amplification coefficient of the i-th family member.

[0041] Furthermore, the amplification factor is calculated based on the variance of each environmental parameter in each set of environmental parameter points, and the formula is:

[0042]

[0043] in, represents the jth environmental parameter value of the nth environmental parameter point in the i-th environmental parameter point set, where N is the total number of environmental parameter points; is the mean variance of the environmental parameters of the ith family member, and e is a natural constant.

[0044] Through the above technical solutions, it can be seen that compared with the existing technology, the present invention provides a smart home environment management system based on data analysis. By analyzing the travel behavior data of each family member, it is possible to predict their return time based on their current travel displacement and historical travel behavior data, and automatically adjust the home environment before returning home, thereby ensuring that they have a comfortable environment after returning home; and based on the predicted historical environmental data of each family member, it is possible to analyze the corresponding appropriate environmental parameter range, and intelligently adjust electrical equipment according to the situation of family members in the room at each time period, thereby realizing intelligent adjustment of the home environment and improving the comfort of each family member. The present invention can more intelligently manage the home environment and comprehensively improve the experience of each family member. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0046] Figure 1 Schematic diagram of the overall structure of an embodiment of the present invention. DETAILED DESCRIPTION

[0047] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0048] The embodiment of the present invention discloses a smart home environment management system based on data analysis, such as Figure 1 As shown, it includes: a data acquisition module, a data analysis module and a control module;

[0049] The data acquisition module acquires the travel behavior data of each family member and the home environment data of each family member at home at all times;

[0050] The data analysis module predicts the predicted return time of each family member based on travel behavior data; and analyzes the suitable environmental parameter range for each family member based on home environment data;

[0051] The control module determines the target home environment parameter range for each time period based on the predicted return time of each family member and the suitable environmental parameter range, controls the operation of each household appliance, and adjusts the indoor environment.

[0052] In a specific embodiment, the data acquisition module obtains the travel and displacement data of each family member through the mobile terminal (such as a mobile phone, smart watch, etc.) worn by each family member; determines the time when each family member leaves and returns home through the face recognition device installed at the door and the door opening and closing sensor on the door, and compiles the time series of each family member's departure and return time;

[0053] Travel behavior data include: travel displacement data, departure time series, and return time series;

[0054] The data acquisition module detects the temperature, humidity, carbon dioxide concentration and particulate matter concentration of the indoor air at each moment through temperature sensors, humidity sensors, carbon dioxide concentration sensors and laser particle counters to obtain home environment data.

[0055] In a specific embodiment, the data analysis module includes a home time prediction unit, which trains a corresponding home time prediction model based on each member's historical travel displacement data and each member's historical departure time series and return time series;

[0056] The home time prediction unit predicts the home time of each family member based on the travel displacement data of each family member after the current trip, as well as the departure time series and home time series of the set time length before the current moment, using the trained home time prediction model of each family member.

[0057] In a specific embodiment, the home time prediction model is a TreeCNN-LSTM network model, including a TreeCNN network, a first LSTM network, a second LSTM network, and a fully connected network;

[0058] The TreeCNN network performs convolution and feature extraction on the trajectory tree in the input sample, and outputs the feature sequence to the first LSTM network in the time order of the features;

[0059] The first LSTM network outputs the first prediction time based on the input feature sequence;

[0060] The second LSTM network outputs the second predicted time based on the input time series of going out and returning home;

[0061] The fully connected network performs weighted calculation on the first prediction time and the second prediction time and outputs the final prediction time.

[0062] In a specific embodiment, the TreeCNN network includes an input layer, a tree convolution layer, and a fully connected layer connected in sequence; the input layer is used to input a trajectory tree of a family member; the tree convolution layer performs convolution on each data node in the trajectory tree to obtain a feature representation of each data node in the trajectory tree; the fully connected layer generates a feature sequence based on the time information of the feature representation;

[0063] The first and second LSTM networks consist of an input layer, an LSTM layer, a fully connected layer, and an output layer, connected in sequence. The input layer is used to input a feature sequence, or a sequence of departure and return time. The LSTM layer processes the input sequence and extracts features. The fully connected layer maps the extracted features to the final predicted return time, which is then output to the fully connected network by the output layer.

[0064] The fully connected network maps the predicted home time of the first LSTM network and the second LSTM network to obtain the final predicted home time.

[0065] In a specific embodiment, the trajectory tree is obtained by:

[0066] The travel displacement data of a family member's trip is sampled to obtain three-dimensional tensor data, where the first dimension represents the sampling time, the second dimension represents the two-dimensional plane coordinates of the family member at the sampling time, and the third dimension is the speed of the family member at the sampling time;

[0067] Gridding plane coordinate system, dividing each slice into grids according to the second dimension of each slice in the three-dimensional tensor data, and counting the slice density of each grid;

[0068] According to the set density threshold, key grids are selected from all grids as child nodes of the trajectory tree;

[0069] Taking home as the root node of the trajectory tree, the connection relationship of each node is determined based on whether there is a slice with a sampling time difference less than a set threshold between two nodes, and the nodes are connected in time sequence to obtain the trajectory tree.

[0070] Specifically, after sampling the displacement data, it is gridded into a plane coordinate system and key grids are identified. This effectively simplifies travel trajectories and highlights the differences in trajectories for different travel purposes, facilitating subsequent analysis. Connectivity is then determined based on time nodes, and a trajectory tree is constructed. This allows the use of a TreeCNN network to extract feature representations of position, time, and speed at each key location, thereby outputting a feature time series. The first LSTM network then analyzes the context of the feature time series to generate a first predicted time. The second LSTM network predicts the current homecoming time based on a series of consecutive departure and return times for family members over a period of time (e.g., a month or a week), generating a second predicted time. Because travel behavior varies among family members, some family members have more cyclical departure and return times, resulting in a more accurate second predicted time. However, for others, departure and return times are more variable and random, resulting in a more accurate first predicted time. By using a fully connected network to assign weights between the first and second predicted times for each member, the final predicted homecoming time is more accurate.

[0071] When training the TreeCNN-LSTM network model, the travel displacement data of a family member is first segmented. A trajectory tree is constructed using a complete day's travel and return displacement data as the trajectory features of a sample. A sliding window is used to sample fixed-length door and return time series from the departure and return time series before the current day as the time series features for this sample. The return time of the same day is used as the sample label to generate a training sample. Multiple training samples form the training sample set for each family member, and training is performed to obtain a prediction model for their return time. Through training, an independent home time prediction model is established for each family member, accurately predicting their return time.

[0072] In a specific embodiment, the data analysis module further includes a suitable environment parameter analysis unit, which samples the home environment data according to the time period of each family member at home and the set step length to obtain a plurality of environmental parameter point data of each family member, and then obtain the environmental parameter point set of each member;

[0073] Eliminate points in the set of environmental parameter points that have manual environmental control within the set time period before the sampling time;

[0074] Perform Kmeans clustering on the set of environmental parameter points after the elimination operation, use the k cluster centers as the key points of the environmental parameter point set, and use the maximum and minimum values ​​of air temperature, humidity, carbon dioxide concentration, and particulate matter concentration represented by all key points to determine the appropriate environmental parameter range for the family members corresponding to the environmental parameter point set;

[0075] Eliminate and cluster each set of environmental parameter points one by one to determine the appropriate environmental parameter range for each family member.

[0076] In a specific embodiment, determining the target home environment parameter range for each time period specifically includes:

[0077] Step 1: Based on the predicted return time of each family member, determine the family members who will be indoors at each time period in the future;

[0078] Step 2: Initialize the proportional coefficient p and the amplification coefficient of each family member;

[0079] Step 3: Determine whether the suitable environmental parameter ranges of the family members determined in step 1 have an intersection. If so, use the intersection as the target home environmental parameter range; otherwise, proceed to step 4.

[0080] Step 4: p=p+0.01, amplify the range of each suitable environmental parameter according to the proportional coefficient and the amplification coefficient, and return to step 3.

[0081] In a specific embodiment, the range of each suitable environmental parameter is enlarged, and the formula is:

[0082]

[0083] in, denote the upper and lower bounds of the jth environmental parameter of the i-th family member, respectively; j = 1, 2, 3, 4 denote air temperature, humidity, carbon dioxide concentration, and particulate matter concentration, respectively; q i is the amplification coefficient of the i-th family member.

[0084] In a specific embodiment, the amplification factor is calculated based on the variance of each environmental parameter in each environmental parameter point set, and the formula is:

[0085]

[0086] in, represents the jth environmental parameter value of the nth environmental parameter point in the i-th environmental parameter point set, where N is the total number of environmental parameter points; is the mean variance of the environmental parameters of the ith family member, and e is a natural constant.

[0087] Specifically, the proportional coefficient p is used to control the speed of amplification. The initial value is usually 0.1, and it is gradually increased in each iteration to increase the ratio of the environmental parameter range amplification so that the comfortable environmental parameter ranges of each family member have an intersection. Each family member has different sensitivities to comfortable environmental conditions. Some people can adapt to a wider range of environmental parameters, so the environmental parameter point data in their corresponding environmental parameter point set often have a large degree of discreteness. Some people are more sensitive to comfortable environments. When they feel uncomfortable, they often manually adjust various electrical equipment. In the environmental parameter point set, the environmental parameter point data within a period of time after manual adjustment are eliminated, so that the comfortable environmental parameter range to which they adapt can be determined more accurately. Since the comfortable range to which they adapt is smaller, the environmental parameter point data in the corresponding environmental parameter point set have a low degree of discreteness. By setting the amplification coefficient q of each family member i By adjusting the ratio between the amplification speeds of each family member, it can better balance the sensitivity of each family member to a comfortable environment and improve the overall feeling of each family member indoors.

[0088] After the control module determines the temperature control range, it controls the air conditioner to adjust the temperature; after determining the humidity control range, it controls the air conditioner to dehumidify or controls the humidifier to humidify; after determining the carbon dioxide concentration control range, it controls the fresh air fan to operate for ventilation; after determining the particulate matter concentration range, it controls the air purifier to operate to reduce the particulate matter concentration in the air to a predetermined range.

[0089] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Reference can be made to the common and similar parts between the various embodiments. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the method description.

[0090] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not limited to the embodiments shown herein but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A smart home environment management system based on data analysis, characterized in that: include: Data acquisition module, data analysis module and control module; The data acquisition module acquires the travel behavior data of each family member; acquires the home environment data of each family member at each time when they are at home; The data analysis module predicts the predicted return home time of each family member based on the travel behavior data; and analyzes the suitable environmental parameter range of each family member based on the home environment data; The control module determines the target home environment parameter range for each time period based on the predicted home time of each family member and the suitable environment parameter range, controls the operation of each household appliance, and adjusts the indoor environment: The data acquisition module acquires the travel and displacement data of each family member through the mobile terminal worn by each family member; determines the time of departure and return of each family member through the face recognition device installed at the door and the door opening and closing sensor on the door, and compiles the time series of departure and return of each family member; The travel behavior data includes: the travel displacement data, the outgoing time series and the returning home time series; The data acquisition module detects the temperature, humidity, carbon dioxide concentration and particulate matter concentration of the indoor air at each moment through a temperature sensor, a humidity sensor, a carbon dioxide concentration sensor and a laser particle counter to obtain the home environment data; The data analysis module further includes a suitable environment parameter analysis unit, which samples the home environment data according to the time period of each family member at home and the set step length to obtain a plurality of environmental parameter point data of each family member, and then obtain an environmental parameter point set for each member; Eliminate points in the set of environmental parameter points that have been manually controlled within a set time period before the sampling moment; Perform Kmeans clustering on the set of environmental parameter points after the elimination operation, use the k cluster centers as the key points of the environmental parameter point set, and use the maximum and minimum values ​​of air temperature, humidity, carbon dioxide concentration, and particulate matter concentration represented by all key points to determine the appropriate environmental parameter range for the family members corresponding to the environmental parameter point set; Eliminate and cluster each set of environmental parameter points one by one to determine the appropriate environmental parameter range for each family member; Determining the target home environment parameter range for each time period specifically includes: Step 1: Based on the predicted return time of each family member, determine the family members who will be indoors at each future time period; Step 2: Initialize the proportional coefficient p and the amplification coefficient of each family member; Step 3: Determine whether the suitable environmental parameter ranges of the family members determined in step 1 have an intersection. If so, use the intersection as the target home environmental parameter range; otherwise, proceed to step 4. Step 4: p=p+0.01, amplify each of the suitable environmental parameter ranges according to the proportional coefficient and the amplification coefficient, and return to step 3.

2. The smart home environment management system based on data analysis according to claim 1, characterized in that: The data analysis module includes a home time prediction unit, which trains a corresponding home time prediction model based on each member's historical travel displacement data and each member's historical departure time series and return time series; The home time prediction unit predicts the home time of each family member based on the travel displacement data after the current family member's trip, as well as the departure time sequence and home time sequence of the set time length before the current moment, using the trained home time prediction model of each family member.

3. The smart home environment management system based on data analysis according to claim 2, characterized in that: The home time prediction model is a TreeCNN-LSTM network model, including a TreeCNN network, a first LSTM network, a second LSTM network, and a fully connected network; The TreeCNN network performs convolution and feature extraction on the trajectory tree in the input sample, and outputs a feature sequence to the first LSTM network in the time order of the features; The first LSTM network outputs a first predicted time according to the input feature sequence; The second LSTM network outputs a second predicted time based on the input going out time sequence and returning home time sequence; The fully connected network performs weighted calculation on the first prediction time and the second prediction time, and outputs a final prediction time.

4. The smart home environment management system based on data analysis according to claim 3, characterized in that: The TreeCNN network includes an input layer, a tree convolution layer, and a fully connected layer connected in sequence; the input layer is used to input the trajectory tree of a family member; The tree convolution layer performs convolution on each data node in the trajectory tree to obtain the feature representation of each data node in the trajectory tree; A fully connected layer generates the feature sequence according to the time information represented by the feature; The first LSTM network and the second LSTM network include an input layer, an LSTM layer, a fully connected layer, and an output layer connected in sequence; the input layer is used to input the feature sequence, or the outgoing time sequence and the returning time sequence; the LSTM layer processes the input sequence and extracts features; The fully connected layer maps the extracted features to the final predicted return time, and the output layer outputs it to the fully connected network; The fully connected network maps the predicted home time of the first LSTM network and the second LSTM network to obtain a final predicted home time.

5. The smart home environment management system based on data analysis according to claim 3, characterized in that: The trajectory tree is obtained in the following way: The travel displacement data of a family member's trip is sampled to obtain three-dimensional tensor data, where the first dimension represents the sampling time, the second dimension represents the two-dimensional plane coordinates of the family member at the sampling time, and the third dimension is the speed of the family member at the sampling time; A gridded plane coordinate system is used to divide each slice into grids according to the second dimension of each slice in the three-dimensional tensor data, and the slice density of each grid is calculated; According to the set density threshold, key grids are selected from all grids as child nodes of the trajectory tree; Taking home as the root node of the trajectory tree, the connection relationship of each node is determined based on whether there is a slice with a sampling time difference less than a set threshold between two nodes, and the nodes are connected in time sequence to obtain the trajectory tree.

6. The smart home environment management system based on data analysis according to claim 1, characterized in that: The formula for enlarging the range of each suitable environmental parameter is: in, denote the upper and lower bounds of the jth environmental parameter of the i-th family member, respectively; j = 1, 2, 3, 4 denote air temperature, humidity, carbon dioxide concentration, and particulate matter concentration, respectively; q i is the amplification coefficient of the i-th family member.

7. The smart home environment management system based on data analysis according to claim 6, characterized in that: The amplification factor is calculated based on the variance of each environmental parameter in each set of environmental parameter points, and the formula is: in, represents the jth environmental parameter value of the nth environmental parameter point in the i-th environmental parameter point set, where N is the total number of environmental parameter points; is the mean variance of the environmental parameters of the ith family member, and e is a natural constant.

Images