Intelligent housekeeper system based on artificial intelligence

Abstract

The invention discloses an intelligent housekeeper system based on artificial intelligence, and relates to the technical field of intelligent housekeepers. The intelligent housekeeper system based on artificial intelligence obtains multi-modal data of a user and extracts multi-modal features; determining a most similar user set based on the multi-modal features and historical multi-modal features of other users; acquiring motion data features and health monitoring data features in real time, and determining an optimal intelligent equipment regulation and control parameter matrix based on the motion data features and the health monitoring data features; whether the health state of the user is abnormal or not is judged in real time based on the health monitoring data features, and if yes, early warning is conducted; the intelligent equipment is regulated and controlled based on the determined intelligent equipment regulation and control parameter matrix, user modification parameters are obtained in real time, the intelligent equipment regulation and control parameter matrix is updated based on the user modification parameters and then stored in a database, and the problems that an existing system has passive response and lacks collaborative optimization are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent housekeepers, and specifically to an intelligent housekeeper system based on artificial intelligence. Background Art

[0002] With the rapid development of the information technology and Internet of Things fields, more and more technological products are connected to each other through the Internet of Things, greatly facilitating people's lives. As a new industry, smart home is at the critical point between the introduction period and the growth period. At present, for the control of some household appliances and lights, the existing technologies can already achieve it, but there is still a certain distance from an intelligent housekeeper.

[0003] Chinese patent application with the publication number CN108345221A discloses a home intelligent housekeeper system, including a safety monitoring system, an intelligent replenishment system, a household appliance control system, a lighting control system, and a housekeeper host. The safety monitoring system includes an air quality detection module, a water quality detection module, a gas monitoring and management module, a power distribution monitoring and management module, a household appliance monitoring module, and a security monitoring module. The intelligent replenishment system includes a food monitoring module, a clothing monitoring module, and a daily necessities monitoring module. The household appliance control system includes a switch control module and a reservation use module. The lighting control system includes an atmosphere regulation module and an environment monitoring module. The housekeeper host includes a voice system and a central control system.

[0004] However, the existing intelligent housekeeper systems have problems of passive response and lack of collaborative optimization. Summary of the Invention

[0005] In view of the deficiencies of the prior art, the present invention provides an intelligent housekeeper system based on artificial intelligence, which solves the problems of passive response and lack of collaborative optimization existing in the existing systems.

[0006] To achieve the above objectives, the present invention is realized through the following technical solutions: An intelligent housekeeper system based on artificial intelligence includes the following steps: acquiring user multi-modal data and extracting multi-modal features; determining the most similar set of user users based on the multi-modal features and the historical multi-modal features of other using users; obtaining real-time motion data features and health monitoring data features, and determining an optimal intelligent device control parameter matrix based on the motion data features and the health monitoring data features; judging in real time whether the user's health status is abnormal based on the health monitoring data features, and giving an early warning if it is abnormal; controlling the intelligent devices based on the determined intelligent device control parameter matrix, and obtaining real-time user modified parameters, and updating the intelligent device control parameter matrix based on the user modified parameters and storing it in the database.

[0007] Furthermore, the user multimodal data includes user basic information, device usage logs, motion data, health monitoring data, and user-device historical interaction records. The multimodal features include user basic information features, device usage log features, motion data features, health monitoring data features, and user-device historical interaction record features. The historical multimodal features include historical user basic information features, historical device usage log features, historical motion data features, historical health monitoring data features, and historical user-device historical interaction record features.

[0008] Furthermore, determining the most similar set of user usage based on multimodal features and historical multimodal features of other user usage includes the following steps: obtaining the historical multimodal features of other user usage stored in the database; initially comparing the user basic information features, device usage log features, and user-device historical interaction record features of the user multimodal data with the historical user basic information features, historical device usage log features, and historical user-device historical interaction record features of the historical multimodal features of other user usage to obtain an initial similarity coefficient, where the user basic information features include user individual feature data and registered device information; marking the other user usage corresponding to the historical multimodal features with an initial similarity coefficient greater than the set initial similarity threshold as the initial most similar set of user usage; performing a secondary similarity analysis on the motion data features and health monitoring data features with the historical motion data features and historical health monitoring data features of the initial most similar set of user usage to obtain a secondary similarity coefficient, and marking the initial most similar set of user usage with a secondary similarity coefficient greater than the set secondary similarity threshold as the most similar set of user usage.

[0009] Furthermore, the methods for obtaining the initial similarity coefficient Fx and the secondary similarity coefficient Sx are as follows:

[0010]

[0011] Sx = (e - 1) σ(yY,syY) + σ(yK, syK);

[0012] where yJ is the user basic information feature, yR is the device usage log feature, yRS is the user-device historical interaction record feature, yK is the health monitoring data feature, yY is the motion data feature, syJ is the user basic information feature, syR is the historical device usage log feature, syRS is the historical user-device historical interaction record feature, syK is the historical health monitoring data feature, syY is the historical motion data feature, σ(·) is the similarity function, and e is the natural constant.

[0013] Further, determining the intelligent device control parameter matrix includes the following steps: determining the user's current motion type and motion area based on the user's current motion feature data; combining the motion type, motion area, and the user's current health monitoring data features into intelligent device control parameter influence data; obtaining the alternative intelligent device control parameter matrices corresponding to each of the most similar user groups in the most similar user group based on the intelligent device control parameter influence data; obtaining the optimal intelligent device control parameter matrix from the alternative intelligent device control parameter matrices based on the optimization objective function, where the rows of the optimal intelligent device control parameter matrix represent intelligent devices, and the columns represent the control parameters of each intelligent device.

[0014] Further, obtaining the alternative intelligent device control parameter matrices corresponding to each of the most similar user groups in the most similar user group based on the intelligent device control parameter influence data is as follows: comparing the intelligent device control parameter influence data with the historical intelligent device control parameter influence data of each similar user in the most similar user group stored in the database to determine the comparison coefficient sets [F1, F2,..., F i ,...] j , F i =σ(Zt, SZt i ), where j is the number of the similar user and i is the number of the historical intelligent device control parameter influence data; determining the maximum comparison coefficient in each comparison coefficient set and obtaining the alternative parameter control scheme corresponding to each maximum comparison coefficient from the database.

[0015] Further, obtaining the optimal intelligent device control parameter matrix from the alternative intelligent device control parameter matrices based on the optimization objective function is as follows: determining the objective function value T corresponding to each alternative intelligent device control parameter matrix based on the objective function a :

[0016]

[0017] where Gz is the just increment after device control adjustment according to the alternative intelligent device control parameter matrix, xT is the response time when the alternative intelligent device control parameter matrix is applied, ΔJK is the positive increment of the health state, Ct is the duration of the improvement of the health state of the similar user after the historical execution of the alternative intelligent device control parameter matrix, α is the weight factor of (Fx + Sx), β is the weight factor of , γ is the weight factor of , is the weight factor of (ΔJK + Ct); denoting the alternative intelligent device control parameter matrix corresponding to the maximum objective function value as the optimal intelligent device control parameter matrix.

[0018] Furthermore, to determine in real time whether the user's health status is abnormal based on the characteristics of health monitoring data, the following steps are included: obtaining the calibrated characteristics of health monitoring data stored in the database, including physiological calibrated characteristics, micro-expression calibrated characteristics, and intonation calibrated characteristics; obtaining the current exercise type and exercise area of the user, and obtaining the regional environmental parameters of the exercise area; obtaining the allowable deviation values of health status based on the current exercise data characteristics, regional environmental parameters, and user individual characteristic data of the user, including the allowable deviation values of physiological characteristics, micro-expression characteristics, and intonation characteristics; determining the user's health status index based on the health monitoring data characteristics, calibrated characteristics of health monitoring data, and allowable deviation values of health status, where the health monitoring data characteristics include physiological characteristics, micro-expression characteristics, and intonation characteristics; if the user's health status index is greater than the user's health status threshold in the database, the user's health status is abnormal, and if it is not greater, the user's health status is normal.

[0019] Furthermore, the method for obtaining the user's health status index is as follows:

[0020]

[0021] Among them, JK is the user's health status index, St is the physiological characteristic, Wt is the micro-expression characteristic, Yt is the intonation characteristic, cSt is the physiological calibrated characteristic, cWt is the micro-expression calibrated characteristic, cYt is the intonation calibrated characteristic, ΔSt is the allowable deviation value of physiological characteristics, ΔWt is the allowable deviation value of micro-expression characteristics, and ΔYt is the allowable deviation value of intonation characteristics.

[0022] Furthermore, the process of obtaining the allowable deviation values of health status based on the current exercise data characteristics, regional environmental parameters, and user individual characteristic data of the user is as follows: combining the current exercise data characteristics, regional environmental parameters, and user individual characteristic data of the user and recording it as the current deviation determination data Sp; comparing the current deviation determination data with each deviation determination matching data cSp stored in the database b one by one to obtain the determination coefficient σ(Sp, cSp b ); determining the deviation determination matching data corresponding to the maximum determination coefficient, and obtaining the allowable deviation value of health status corresponding to the deviation determination matching data from the database.

[0023] The present invention has the following beneficial effects:

[0024] The intelligent housekeeper system based on artificial intelligence integrates multi-modal data of users, uses historical data to match similar user groups, analyzes motion characteristics and health status in real time, dynamically generates an optimal device control parameter matrix, and realizes proactive health management through a health warning mechanism. At the same time, it continuously optimizes the control strategy in combination with user feedback, thereby improving the intelligence level of device control, meeting the personalized needs of users in different scenarios, ensuring the health status, and solving the defects of passive response and lack of collaborative optimization existing in the existing system.

[0025] Of course, it is not necessary for any product implementing the present invention to achieve all the above-mentioned advantages simultaneously. Brief Description of the Drawings

[0026] Figure 1 It is a flow block diagram of the intelligent housekeeper system based on artificial intelligence of the present invention. Detailed Embodiments

[0027] Please refer to Figure 1 , an embodiment of the present invention provides a technical solution: an intelligent housekeeper system based on artificial intelligence, including the following steps: obtaining multi-modal data of users and extracting multi-modal features;

[0028] User basic information: Obtain user basic information through the personal information actively provided by the user during system registration, such as age, gender, height, weight, self-evaluation of health status, etc., and the information of intelligent devices associated with the user during registration, such as the models, IDs, etc. of intelligent bracelets, intelligent weighing scales, intelligent air conditioners, etc. Device usage logs: The system automatically records the operation behaviors of the user using various intelligent devices, including the opening / closing time of the device, adjustment parameters (such as temperature, wind speed, etc.), usage duration, etc., to form device usage logs.

[0029] Motion data: With the help of intelligent sports devices worn by the user (such as intelligent bracelets, intelligent watches, etc.) or sensors installed indoors (such as accelerometers, gyroscopes, etc.), real-time collect data such as the user's number of steps, movement distance, calories consumed, movement trajectory, movement speed, etc., and capture the user's activity images indoors through an intelligent camera to identify the user's actions and postures, so as to obtain motion data. At the same time, housework activities such as cooking and cleaning can also be included in the scope of motion data, and the above-mentioned devices and sensors are used to capture the action and behavior data of the user during these activities.

[0030] Health monitoring data: Use professional health monitoring devices (such as intelligent sphygmomanometers, intelligent blood glucose meters, intelligent electrocardiogram monitors, etc.) or intelligent devices with health monitoring functions to collect users' physiological index data, such as heart rate, blood pressure, blood oxygen saturation, body temperature, sleep quality, etc., to obtain health monitoring data. User-device historical interaction records: The system records the interaction history between users and intelligent devices, including users' control instructions for devices (such as voice instructions, manual adjustment instructions, etc.), the response results of devices to users' instructions, the time and frequency of interactions, etc., to form user-device historical interaction records.

[0031] User basic information characteristics: Extract users' individual characteristic data from user basic information, such as age range, gender category, body mass index (BMI), etc., and registered device information, such as device type, number of devices, etc., as user basic information characteristics. Device usage log characteristics: Analyze the device usage logs to extract characteristics such as device usage frequency, usage time period distribution, usage correlation between different devices, etc., such as the average daily usage duration of an intelligent air conditioner by a user, the peak time period of using intelligent lighting devices, etc., as device usage log characteristics. Process and analyze the exercise data to extract characteristics such as exercise type (such as walking, running, yoga, cooking, cleaning, etc.), exercise intensity, exercise duration, exercise frequency, etc., such as the number of times of running per week, the average distance and duration of each run, and the duration and frequency of cooking and cleaning, etc., as exercise data characteristics.

[0032] Health monitoring data characteristics: Extract characteristics such as the numerical range, fluctuation trend, and abnormal detection results of physiological indicators from health monitoring data, such as the average value, maximum value, and minimum value of heart rate, the systolic and diastolic blood pressure levels and their stability, etc., as health monitoring data characteristics. User-device historical interaction record characteristics: Analyze the user-device historical interaction records to extract characteristics such as users' preferred operations for different devices, interaction patterns (such as timed operations, condition-triggered operations, etc.), the accuracy and complexity of users' instructions, etc., such as users often opening intelligent curtains at a specific time, or the success rate of users' voice instructions for intelligent speakers, etc., as user-device historical interaction record characteristics.

[0033] Determine the most similar set of user users based on multi-modal characteristics and other historical multi-modal characteristics of the user; User multi-modal data includes user basic information, device usage logs, exercise data, health monitoring data, and user-device historical interaction records, and multi-modal characteristics include user basic information characteristics, device usage log characteristics, exercise data characteristics, health monitoring data characteristics, and user-device historical interaction record characteristics; Historical multi-modal characteristics include historical user basic information characteristics, historical device usage log characteristics, historical exercise data characteristics, historical health monitoring data characteristics, and historical user-device historical interaction record characteristics.

[0034] Obtain the historical multi-modal features of other users stored in the database; conduct an initial similarity comparison between the user basic information features, device usage log features, and user-device historical interaction record features of the user multi-modal data and the historical user basic information features, historical device usage log features, and historical user-device historical interaction record features of the historical multi-modal features of other users to obtain an initial similarity coefficient, where the user basic information features include user individual feature data (including but not limited to age, height, gender, weight) and registered device information (including but not limited to intelligent air conditioning systems and intelligent lighting systems); mark the other users corresponding to the historical multi-modal features with an initial similarity coefficient greater than the set initial similarity threshold as the initial most similar user set.

[0035] By conducting an initial similarity comparison between specific features in the user multi-modal data and the corresponding features of other users, and setting an initial similarity threshold to screen the initial most similar user set, the scope can be quickly narrowed down, focusing on those user groups that are relatively similar to the target user in terms of basic information, device usage, and interaction records, improving the search efficiency and reducing unnecessary calculations.

[0036] Conduct a secondary similarity analysis between the motion data features and health monitoring data features and the historical motion data features and historical health monitoring data features of the initial most similar user set to obtain a secondary similarity coefficient, and mark the initial most similar users with a secondary similarity coefficient greater than the set secondary similarity threshold as the most similar user set. Considering the motion data features and health monitoring data features, conducting a secondary similarity analysis on the initial most similar user set, and setting a secondary similarity threshold to finally determine the most similar user set further improves the accuracy of similar user matching. This ensures that the selected users are similar to the target user in a more comprehensive feature dimension, making the subsequent data analysis based on similar users and the formulation of intelligent device control schemes more targeted and effective, and better meeting the personalized needs of users.

[0037] The methods for obtaining the initial similarity coefficient Fx and the secondary similarity coefficient Sx are as follows:

[0038]

[0039] Sx = (e - 1) σ(yY,syY) + σ(yK, syK);

[0040] Among them, yJ is the user basic information feature, yR is the device usage log feature, yRS is the user-device historical interaction record feature, yK is the health monitoring data feature, yY is the exercise data feature, syJ is the historical user basic information feature, syR is the historical device usage log feature, syRS is the historical user-device historical interaction record feature, syK is the historical health monitoring data feature, syY is the historical exercise data feature, σ(·) is the similarity function, which can specifically be the cosine similarity function, and e is the natural constant.

[0041] By defining specific similarity functions and calculation formulas, the basic information, device usage logs, and user-device historical interaction record features in the user's multimodal data are quantitatively compared with the historical data, enabling the accurate measurement of the similarity between users. This quantitative method makes the evaluation of similarity more objective and accurate, avoiding the deviation caused by subjective judgment and providing a reliable basis for screening similar users subsequently.

[0042] On the basis of the initial similarity coefficient, the exercise data feature and the health monitoring data feature are further incorporated into the similarity calculation, making the screening of the similar user set more comprehensive and accurate. By setting the secondary similarity threshold, it is ensured that the finally determined most similar user set is similar to the target user in a wider feature dimension, which helps to better understand the user's living habits, health status, etc., and lays a foundation for providing personalized services.

[0043] Obtain the exercise data feature and the health monitoring data feature in real time. Based on the exercise data feature and the health monitoring data feature, determine the optimal intelligent device control parameter matrix; based on the user's current exercise feature data, determine the user's current exercise type and exercise area; based on the user's current exercise feature data, determine the exercise type and exercise area, which can accurately understand the user's current activity status and the environment where the user is located, providing a targeted basis for subsequent intelligent device control.

[0044] Combine the exercise type, exercise area, and the user's current health monitoring data feature into the intelligent device control parameter influence data; combine the exercise type, exercise area, and the user's current health monitoring data feature into the intelligent device control parameter influence data, realizing the fusion of multi-dimensional data, fully considering the user's needs in different exercise and health states, and making the intelligent device control more comprehensive and personalized. The exercise type judgment can use machine learning algorithms (such as decision trees, support vector machines, etc.) to classify the extracted exercise features and identify the exercise type the user is performing, such as walking, running, yoga, cooking, cleaning, etc. The exercise area is matched with the indoor area division map through the user's geographical location information to determine the precise area where the user is located.

[0045] Based on the data of the intelligent device control parameter influence, obtain the alternative intelligent device control parameter matrices corresponding to each of the most similar user in the set of the most similar users; obtain the optimal intelligent device control parameter matrix from the alternative intelligent device control parameter matrices based on the optimization objective function, where the rows of the optimal intelligent device control parameter matrix represent intelligent devices and the columns represent the control parameters of each intelligent device.

[0046] Compare the data of the intelligent device control parameter influence with the historical data of the intelligent device control parameter influence of each similar user in the set of the most similar users stored in the database, and determine the set of comparison coefficients [F1, F2,..., F i ,...] j , F i =σ(Zt, SZt i ), where j is the number of the similar user and i is the number of the historical data of the intelligent device control parameter influence; determine the maximum comparison coefficient in each set of comparison coefficients, and obtain the alternative parameter control scheme corresponding to each maximum comparison coefficient from the database.

[0047] By comparing the data of the intelligent device control parameter influence with the historical data of the intelligent device control parameter influence of each similar user in the set of the most similar users stored in the database, it is possible to accurately find historical control cases similar to the current user's situation, providing a reference for the intelligent device control of the current user. Determine the set of comparison coefficients corresponding to each similar user, and evaluate the similarity degree between the current user and the similar users in a quantitative way, making the subsequent selection process more objective and accurate.

[0048] Determine the maximum comparison coefficient in each set of comparison coefficients, and obtain the alternative parameter control scheme corresponding to each maximum comparison coefficient from the database, which can efficiently screen out the intelligent device control scheme most suitable for the current user's situation, improving the response speed and control efficiency of the system. Through comparison and screening, multiple alternative intelligent device control parameter matrices are provided for the current user, increasing the diversity and flexibility of selection, avoiding the risks that may be brought by a single scheme, and at the same time providing a basis for subsequent optimal selection.

[0049] Based on the data of the intelligent device control parameter influence, obtain the alternative intelligent device control parameter matrices corresponding to each of the most similar user in the set of the most similar users, and use the control experience of similar users to provide the current user with a variety of control scheme selections, avoiding the mismatch problem that may be brought by a single control strategy.

[0050] Based on the objective function, determine the objective function value T corresponding to each alternative intelligent device control parameter matrix a :

[0051]

[0052] Among them, Gz is the just-increase amount after adjusting the device control according to the alternative intelligent device regulation parameter matrix, xT is the response time when the alternative intelligent device regulation parameter matrix is applied, and ΔJK is the positive increment of the health state. Specifically, first determine the initial health state level of the user (compare the health monitoring data characteristics of the user with each set of health monitoring data characteristics in the database one by one, determine the closest set of health monitoring data characteristics, and obtain the health state level corresponding to the set of health monitoring data characteristics stored in the database). When the alternative intelligent device regulation parameter matrix is applied, obtain the final health state level of the user, and take the difference between the final health state level and the initial health state level, which is recorded as the positive increment of the health state. Ct is the duration of the improvement of the health state of similar users after the historical execution of the alternative intelligent device regulation parameter matrix, α is the weight factor of (Fx + Sx), β is the weight factor of and γ is the weight factor of (ΔJK + Ct); The alternative intelligent device regulation parameter matrix corresponding to the maximum objective function value is recorded as the optimal intelligent device regulation parameter matrix.

[0053] By optimizing the objective function to obtain the optimal intelligent device regulation parameter matrix from the alternative intelligent device regulation parameter matrix, multiple factors such as the increased comfort after regulation, response time, and health state improvement are comprehensively considered, ensuring that the finally selected regulation scheme can achieve the best balance in all aspects and improve the user experience and health benefits.

[0054] The whole process from user state perception to data fusion analysis, and then to scheme selection and optimization based on similar user data reflects the intelligent characteristics of the system, enabling the intelligent butler system to automatically adjust intelligent devices according to the user's real-time state and personalized needs without manual intervention by the user, improving the convenience and practicality of the system. When determining the optimal intelligent device regulation parameter matrix, factors such as the positive increment of the health state and the duration of the improvement of the health state are fully considered, ensuring that the regulation of intelligent devices can not only improve the user's immediate comfort but also contribute to the long-term health maintenance of the user, realizing the organic combination of health and comfort.

[0055] Based on the characteristics of health monitoring data, it is determined in real time whether the user's health status is abnormal. If it is abnormal, a warning is issued; the reference characteristics of health monitoring data stored in the database are obtained, including physiological reference characteristics, micro-expression reference characteristics, and voice intonation reference characteristics; the current exercise type and exercise area of the user are obtained, and the regional environmental parameters of the exercise area (including but not limited to temperature, humidity, brightness, and noise decibels) are obtained; based on the current exercise data characteristics of the user, regional environmental parameters, and user individual characteristic data, the allowable deviation values of health status are obtained, including physiological characteristic allowable deviation values, micro-expression characteristic allowable deviation values, and voice intonation characteristic allowable deviation values; based on the current exercise data characteristics, regional environmental parameters, and user individual characteristic data of the user, obtaining the allowable deviation values of health status can accurately determine the reasonable fluctuation range of health status for different users, different environments, and activity situations.

[0056] Based on the health monitoring data characteristics, health monitoring data reference characteristics, and allowable deviation values of health status, the user's health status index is determined, where the health monitoring data characteristics include physiological characteristics, micro-expression characteristics, and voice intonation characteristics; if the user's health status index is greater than the user's health status threshold in the database, the user's health status is abnormal, and if it is not greater, the user's health status is normal.

[0057] The method for obtaining the user's health status index is as follows:

[0058]

[0059] Among them, JK is the user's health status index, St is the physiological characteristic, Wt is the micro-expression characteristic, Yt is the voice intonation characteristic, cSt is the physiological reference characteristic, cWt is the micro-expression reference characteristic, cYt is the voice intonation reference characteristic, ΔSt is the physiological characteristic allowable deviation value, ΔWt is the micro-expression characteristic allowable deviation value, and ΔYt is the voice intonation characteristic allowable deviation value.

[0060] Based on the current exercise data characteristics of the user, regional environmental parameters, and user individual characteristic data, the allowable deviation values of health status are obtained as follows: The current exercise data characteristics, regional environmental parameters, and user individual characteristic data of the user are combined and recorded as the current deviation determination data Sp; the current deviation determination data is compared one by one with each deviation determination matching data cSp b in the database to obtain the determination coefficient σ(Sp, cSp b ); the deviation determination matching data corresponding to the maximum determination coefficient is determined, and the allowable deviation value of health status corresponding to the deviation determination matching data is obtained from the database. Comparing the current deviation determination data with each deviation determination matching data stored in the database can accurately find historical data similar to the current user situation, and measure the matching degree by obtaining the determination coefficient, providing a reliable basis for obtaining the allowable deviation values of health status in the future.

[0061] Determine the deviation corresponding to the maximum determination coefficient to determine the matching data, and obtain the allowable deviation value of the health status corresponding to this data from the database, which can provide personalized and dynamic allowable deviation values of the health status for each user, fully considering factors such as the user's current exercise, environment, and individual characteristics, making the assessment of the health status more in line with the actual situation of the user.

[0062] Regulate the intelligent device based on the determined intelligent device regulation parameter matrix, and obtain the user-modified parameters in real time. After updating the intelligent device regulation parameter matrix based on the user-modified parameters, store it in the database.

[0063] An electronic device includes: a processor; and a memory in which computer program instructions are stored, and when the computer program instructions are run by the processor, the processor executes the intelligent housekeeper system based on artificial intelligence as described above.

[0064] A computer-readable storage medium is used to store a program, and when the program is executed by a processor, it implements the intelligent housekeeper system based on artificial intelligence as described above.

[0065] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program codes.

[0066] The present invention is described with reference to the flowcharts and / or block diagrams of systems, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of the processes and / or blocks in the flowchart and / or block diagram can also be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in one process Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0067] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to work in a particular manner, such that the instructions stored in the computer-readable memory produce a manufacture including an instruction device that implements the functions specified in one or more of the processes Figure 1 or blocks Figure 1 specified in one or more of the processes and / or blocks.

[0068] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, such that a series of operational steps are performed on the computer or other programmable apparatus to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable apparatus provide steps for implementing the functions specified in one or more of the processes Figure 1 or blocks Figure 1 specified in one or more of the processes and / or blocks.

[0069] Although the preferred embodiments of the present invention have been described, additional changes and modifications can be made by those skilled in the art once they learn the basic inventive concept. Therefore, the appended claims are intended to be construed to include the preferred embodiments as well as all changes and modifications that fall within the scope of the present invention.

[0070] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations.

Claims

1. An intelligent housekeeper system based on artificial intelligence, characterized in that, It includes a feature acquisition module, a most similar user determination module, an optimal regulation parameter acquisition module, a health status monitoring module, and a data storage module, where: The feature acquisition module is used to acquire user multimodal data and extract multimodal features; The most similar user determination module is used to determine the set of most similar users based on multimodal features and the historical multimodal features of other users; The optimal regulation parameter acquisition module is used to acquire the motion data features and health monitoring data features in real time, and determine the optimal intelligent device regulation parameter matrix based on the motion data features and health monitoring data features; The health status monitoring module is used to judge in real time whether the user's health status is abnormal based on the health monitoring data features, and issue a warning if it is abnormal; The data storage module is used to regulate the intelligent device based on the determined intelligent device regulation parameter matrix, and acquire the user-modified parameters in real time. After updating the intelligent device regulation parameter matrix based on the user-modified parameters, it is stored in the database.

2. The intelligent butler system based on artificial intelligence according to claim 1, wherein The user multimodal data includes user basic information, device usage logs, motion data, health monitoring data, and user-device historical interaction records. The multimodal features include user basic information features, device usage log features, motion data features, health monitoring data features, and user-device historical interaction record features; the historical multimodal features include historical user basic information features, historical device usage log features, historical motion data features, historical health monitoring data features, and historical user-device historical interaction record features.

3. The intelligent butler system based on artificial intelligence according to claim 2, characterized in that, Determining the set of most similar users based on multimodal features and the historical multimodal features of other users includes the following steps: Acquire the historical multimodal features of other users stored in the database; Perform an initial similarity comparison between the user basic information features, device usage log features, and user-device historical interaction record features of the user multimodal data and the historical user basic information features, historical device usage log features, and historical user-device historical interaction record features of the historical multimodal features of other users to obtain an initial similarity coefficient, where the user basic information features include user individual feature data and registered device information; Record the other users corresponding to the historical multimodal features with the initial similarity coefficient greater than the set initial similarity threshold as the initial set of most similar users; Perform a secondary similarity analysis on the motion data features and health monitoring data features and the historical motion data features and historical health monitoring data features of the initial set of most similar users, obtain a secondary similarity coefficient, and record the initial most similar users with the secondary similarity coefficient greater than the set secondary similarity threshold as the set of most similar users.

4. An intelligent butler system based on artificial intelligence according to claim 3, characterized in that, The methods for obtaining the initial similarity coefficient Fx and the secondary similarity coefficient Sx are as follows: Sx = (e - 1) σ(yY,syY) + σ(yK, syK); Among them, yJ is the user basic information feature, yR is the device usage log feature, yRS is the user-device historical interaction record feature, yK is the health monitoring data feature, yY is the exercise data feature, syJ is the historical user basic information feature, syR is the historical device usage log feature, syRS is the historical user-device historical interaction record feature, syK is the historical health monitoring data feature, syY is the historical exercise data feature, σ(·) is the similarity function, and e is the natural constant.

5. An intelligent housekeeper system based on artificial intelligence according to claim 3, characterized in that, Determine the intelligent device regulation parameter matrix, including the following steps: Determine the current exercise type and exercise area of the user based on the user's current exercise feature data; Combine the exercise type, exercise area, and the user's current health monitoring data feature into the intelligent device regulation parameter influence data; Based on the intelligent device regulation parameter influence data, obtain the alternative intelligent device regulation parameter matrices corresponding to each most similar user in the most similar user set; Based on the optimization objective function, obtain the optimal intelligent device regulation parameter matrix from the alternative intelligent device regulation parameter matrices, where the rows of the optimal intelligent device regulation parameter matrix represent intelligent devices, and the columns represent the control parameters of each intelligent device.

6. The intelligent housekeeper system based on artificial intelligence according to claim 5, characterized in that, Based on the intelligent device regulation parameter influence data, obtain the alternative intelligent device regulation parameter matrices corresponding to each most similar user in the most similar user set. The process is as follows: Compare the intelligent device regulation parameter impact data with the historical intelligent device regulation parameter impact data of each similar user in the most similar user set stored in the database to determine the comparison coefficient set [F1, F2,..., F i ,...] j , F i = σ(Zt, SZt i ), where j is the number of the similar user and i is the number of the historical intelligent device regulation parameter impact data; Determine the maximum comparison coefficient in each comparison coefficient set, and obtain the alternative parameter control scheme corresponding to each maximum comparison coefficient from the database.

7. An intelligent butler system based on artificial intelligence according to claim 5, characterized in that, Based on the optimization objective function, obtain the optimal intelligent device regulation parameter matrix from the alternative intelligent device regulation parameter matrices. The process is as follows: Determine the objective function value T corresponding to each alternative intelligent device regulation parameter matrix based on the objective function a : Among them, Gz is the just-increased amount after adjusting device control according to the alternative intelligent device regulation parameter matrix, xT is the response time when the alternative intelligent device regulation parameter matrix is applied, ΔJK is the positive increment of the health state, Ct is the duration of the improvement of the health state of similar users after the historical execution of the alternative intelligent device regulation parameter matrix, α is the weight factor of (Fx + Sx), β is 's weight factor, γ is 's weight factor, is the weight factor of (ΔJK + Ct); Record the alternative intelligent device regulation parameter matrix corresponding to the maximum objective function value as the optimal intelligent device regulation parameter matrix.

8. An intelligent housekeeper system based on artificial intelligence according to claim 5, characterized in that, Based on the health monitoring data feature, determine in real time whether the user's health status is abnormal, including the following steps: Obtain the health monitoring data calibration features stored in the database, including physiological calibration features, micro-expression calibration features, and intonation calibration features; Obtain the user's current exercise type and exercise area, and obtain the regional environment parameters of the exercise area; Based on the user's current exercise data feature, regional environment parameters, and user individual feature data, obtain the health status allowable deviation values, including physiological feature allowable deviation values, micro-expression feature allowable deviation values, and intonation feature allowable deviation values; Based on the health monitoring data feature, health monitoring data calibration feature, and health status allowable deviation value, determine the user health status index, where the health monitoring data feature includes physiological features, micro-expression features, and intonation features; If the user health status index is greater than the user health status threshold in the database, the user's health status is abnormal; if not, the user's health status is normal.

9. An intelligent housekeeper system based on artificial intelligence according to claim 8, characterized in that, The method for obtaining the user health status index is as follows: Among them, JK is the user health status index, St is the physiological feature, Wt is the micro-expression feature, Yt is the speech intonation feature, cSt is the physiological calibration feature, cWt is the micro-expression calibration feature, cYt is the speech intonation calibration feature, ΔSt is the allowable deviation value of the physiological feature, ΔWt is the allowable deviation value of the micro-expression feature, and ΔYt is the allowable deviation value of the speech intonation feature.

10. An intelligent housekeeper system based on artificial intelligence according to claim 8, characterized in that, Based on the user's current motion data characteristics, regional environmental parameters, and user individual characteristic data, obtain the allowable deviation value of the health status. The process is as follows: Combine the user's current motion data characteristics, regional environmental parameters, and user individual characteristic data and denote it as the current deviation determination data Sp; Compare the current deviation determination data with each deviation determination matching data cSp stored in the database b one by one to obtain the determination coefficient σ(Sp, cSp b ); Determine the deviation determination matching data corresponding to the maximum determination coefficient, and obtain the allowable deviation value of the health status corresponding to this deviation determination matching data from the database.

Images