Internet-based intelligent home appliance fault monitoring, acquisition and processing system and method

Abstract

The present invention discloses an Internet-based intelligent home appliance fault monitoring, acquisition and processing system and method, belonging to the technical field of intelligent home appliances. The present invention includes: monitoring the interaction behavior of intelligent home appliances, obtaining the response data and user reaction data of intelligent home appliances during interaction, uploading the data to the monitoring center after security processing; processing the received data through the monitoring center, discriminating the interaction behavior of intelligent home appliances based on the fault detection model, when it is determined that there is a fault in the interaction behavior of intelligent home appliances, the monitoring center sends a communication request to the intelligent home appliances, and collects the interaction-related data of the intelligent home appliances after obtaining user authorization to generate a machine training set; optimizing the interaction function of intelligent home appliances based on the machine training set, and processing the interaction function faults of intelligent home appliances through machine learning. It improves the accuracy of the interaction between intelligent home appliances and users, solves the interaction behavior faults of intelligent home appliances, and optimizes the user experience.

Description

Technical Field

[0001] The present invention relates to the technical field of smart home appliances, and particularly to an intelligent home appliance fault monitoring, acquisition and processing system and method based on the Internet. Background Technique

[0002] Smart home appliances refer to home appliance products with intelligent functions that are connected through the Internet or other wireless networks. These home appliance products can be remotely controlled and monitored through intelligent terminal devices and have human-computer interaction interfaces such as voice recognition and touch screens. With the development of Internet technology and the improvement of consumer demands, different smart home appliances can be interconnected and communicate with each other to achieve the linkage of home devices.

[0003] When a user interacts with a smart home appliance during actual use, the smart home appliance may have abnormal recognition of the user's instructions, resulting in no response or incorrect response. This may be because the instructions provided by the user are not clear or explicit enough, or the smart home appliance has not completed the learning of the user's preferences, resulting in problems with the settings of the smart home appliance, causing the interaction function of the smart home appliance to malfunction and affecting the user's interaction experience. At the same time, most users are difficult to diagnose the cause of interaction failure by themselves and need to use professional diagnostic tools or contact the after-sales service of the manufacturer for on-site inspection, thereby reducing the user's usage frequency of smart home appliances and being unfavorable to the development of smart home appliances. Summary of the Invention

[0004] The purpose of the present invention is to provide an intelligent home appliance fault monitoring, acquisition and processing system and method based on the Internet to solve the problems raised in the above background technique.

[0005] To solve the above technical problems, the present invention provides the following technical solution: An intelligent home appliance fault monitoring, acquisition and processing method based on the Internet, including the following steps:

[0006] Step S100: Monitor the interaction behavior of the smart home appliance, obtain the response data and user reaction data of the smart home appliance during interaction, and upload the data to the monitoring center after security processing;

[0007] Step S200: Process the received data through the monitoring center, discriminate the interaction behavior of the smart home appliance based on the fault detection model, and execute Step S300 when it is determined that the interaction behavior of the smart home appliance has a fault;

[0008] Step S300: The monitoring center sends a communication request to the smart home appliance, and after obtaining user authorization, collects the interaction-related data of the smart home appliance to generate a machine training set;

[0009] Step S400: Optimize the interaction function of the smart home appliance based on the machine training set, and process the interaction function failures of the smart home appliance through machine learning.

[0010] Further, the response data during the interaction of the smart home appliance is specifically the first instruction data sent by the user received by the smart home appliance, and the execution result data returned to the system after the instruction is executed according to the preset conditions; the user response data is used to feedback the user performance after the smart home appliance executes the instruction, including user behavior data and second instruction data; the user behavior data is specifically the operation data of the user on the smart home appliance after the smart home appliance executes the instruction;

[0011] The specific process of uploading the data after security processing is to encrypt it using an encryption algorithm before data transmission, transmit the encrypted data using a secure communication protocol such as HTTPS, and make a local backup of the uploaded data;

[0012] The monitoring center is used to detect the interaction behavior of the smart home appliance and determine whether there are any failures in the interaction behavior of the smart home appliance.

[0013] Further, the step S200 specifically includes the following steps:

[0014] S201: Process the received data through the monitoring center to generate a dataset to be tested. The dataset to be tested includes n groups of data to be tested. Denote the u-th group of data to be tested in the dataset to be tested as D u =(J u , X u , Y u , Z u ), u ∈ [1, n]; Each group of data to be tested includes four elements, namely the first instruction, the executed response code, the user behavior category score, and the second instruction; where J u represents the first instruction received by the smart home appliance in the u-th group of data to be tested, X u represents the response code executed by the smart home appliance in the u-th group of data to be tested, Y u represents the user behavior category score of the smart home appliance after executing the first instruction in the u-th group of data to be tested, and Z u represents the second instruction received by the smart home appliance in the u-th group of data to be tested;

[0015] Among them, the user behavior categories include the first behavior category, the second behavior category, and the third behavior category; the first behavior category means that the user uses the same instruction output method as when issuing the first instruction during operation, the second behavior category means that the user uses an instruction output method different from that when issuing the first instruction during operation, and the third behavior category means that the user does not output an instruction during operation; specifically, the output methods of the first instruction and the second instruction include voice input, gesture input, home appliance button input, APP input, remote control input, etc.

[0016] S202: Detect each group of data to be tested in the data set to be tested through a fault detection model, and calculate the user feedback score Score of the u-th group of data to be tested in the data set to be tested according to the following formula u :

[0017]

[0018] Among them, Score u represents the user feedback score obtained by the smart home appliance after executing the response behavior X u ;

[0019] represents the relationship function between the response behavior code executed by the smart home appliance and the first instruction. When the instruction received by the smart home appliance corresponds to the response behavior code executed, When the instruction received by the smart home appliance does not correspond to the response behavior code executed, g(Z u ) represents the relationship function between the second instruction issued by the user after the smart home appliance executes the response behavior and the first instruction. When , g(Z u ) = 0. When Z u ∈J u , g(Z u ) = 1. When , g(Z u ) = A, A > 1; is the weight coefficient;

[0020] It should be noted that when Z u ∈J u , it means that the second instruction is a subordinate instruction of the first instruction and will not change the state of the smart home appliance after executing the first instruction; When, it means that the second instruction is a peer or superior instruction of the first instruction, and the state of the smart home appliance after executing the first instruction will be overwritten after executing the second instruction.

[0021] represents the user behavior category score corresponding to the j-th behavior data of the smart home appliance by the user after the smart home appliance executes the first instruction; ωj It represents the influence coefficient corresponding to the user behavior category to which the j-th behavior data belongs; j ∈ [1, m], where m represents the number of user behaviors on the smart home appliance within a unit time after the smart home appliance executes the first instruction.

[0022] S203: According to the detection result of the fault detection model, when the user feedback scores of more than B% of the test data in the test data set exceed the preset threshold e, it is determined that there is a fault in the interaction behavior of the smart home appliance, and step S300 is executed.

[0023] Furthermore, the specific steps of the said step S300 are as follows:

[0024] S301: The monitoring center sends a communication request to the smart home appliance, and after obtaining user authorization, it collects the interaction-related data of the smart home appliance, including: the original data of the user sending the first instruction, the response code of the smart home appliance executing the first instruction, the user behavior data after the smart home appliance executes the first instruction, and the original data of the user sending the second instruction within a unit time.

[0025] S302: Process the collected interaction-related data to generate a machine training set, which contains W groups of training data. Any group of training data is denoted as (J, x, y, z); store the machine training set in the buffer. The four elements in the training data respectively correspond to the original data of the user sending the first instruction, the response code of the smart home appliance executing the first instruction, the user behavior data after the smart home appliance executes the first instruction, and the original data of the user sending the second instruction within a unit time after data normalization.

[0026] Furthermore, the specific steps of the said step S400 are as follows:

[0027] S401: Randomly select training data from the machine training set for learning, and calculate any group of the selected training data according to the following formula:

[0028]

[0029] where, P i represents the target value of the i-th group of training data, r represents the user feedback score obtained after the smart home appliance executes the response behavior y, λ represents the discount factor, θ' represents the parameters of the target network; Q(z, y'; θ') represents the target value of the optimal response y' taken under the next instruction z calculated by the target network using the parameters θ'.

[0030] Specifically, the above-mentioned target network is implemented based on the Bellman equation;

[0031] S402: Calculate the predicted value based on the neural network, and update the parameters of the online network by minimizing the loss function. The specific formula is:

[0032] Q(J, x; θ) = f(J; θ);

[0033]

[0034] Wherein, f is a neural network, and the calculation of the neural network is implemented through matrix multiplication and vector addition; J represents the original data features of the first instruction issued by the user, x represents the response behavior encoding executed by the smart home appliance, and θ represents the parameters of the neural network; Q(J, x; θ) represents the predicted value of the smart home appliance executing the response behavior x under the instruction J; L represents the loss function, N represents the number of groups of training data extracted, and i ∈ [1, N];

[0035] Through step S402, the network weights can be adjusted to reduce the gap between the target value and the predicted value;

[0036] S403: Loop steps S401 and S402. After k times of training, copy the parameters of the online network to the target network; apply the updated interaction function to the smart home appliance. Through the above steps, the interaction recognition section of the smart home appliance can be updated according to the user's preference data, generate personalized response executions for the user's behavior, improve the accuracy of the interaction between the smart home appliance and the user, solve the interaction behavior faults of the smart home appliance, and optimize the user experience.

[0037] The intelligent home appliance fault monitoring, acquisition and processing system based on the Internet includes the following modules: a home appliance monitoring module, a fault detection module, a data acquisition module, and an intelligent processing module;

[0038] The home appliance monitoring module is used to monitor the usage data of the smart home appliance and obtain the interaction-related data between the user and the smart home appliance; by monitoring the usage data of the smart home appliance, faults in the interaction behavior can be discovered in a timely manner and the user can be notified in a timely manner, thereby reducing the problems encountered by the user during use and improving the user experience.

[0039] The fault detection module is used to detect the monitoring data output by the home appliance monitoring module and determine whether there are faults in the interaction behavior of the smart home appliance;

[0040] The data acquisition module is used to initiate a communication request to the smart home appliance, and after obtaining user authorization, collect the interaction-related data of the smart home appliance and generate a machine training set;

[0041] The intelligent processing module optimizes the interaction function of the smart home appliance based on the machine training set and processes the interaction function faults of the smart home appliance through machine learning.

[0042] Further, the home appliance monitoring module includes a monitoring and acquisition unit and a data transmission unit;

[0043] The monitoring and acquisition unit is used to acquire the response data and user reaction data during the interaction of intelligent household appliances;

[0044] The data transmission unit is used to perform security processing on the data acquired by the monitoring and acquisition unit and transmit it to the monitoring center.

[0045] Further, the fault detection module includes a data analysis unit and a fault discrimination unit;

[0046] The data analysis unit is used to process the data received by the monitoring center to generate a dataset to be tested. The dataset to be tested includes n groups of data to be tested, and each group of data to be tested includes four elements, namely the first instruction, the executed response code, the user behavior category score, and the second instruction;

[0047] The fault discrimination unit is used to detect each group of data to be tested in the dataset to be tested according to the fault detection model, and determine whether there is a fault in the interaction behavior of the intelligent household appliance by calculating the user feedback score of the data to be tested.

[0048] Further, the data acquisition module includes an authentication acquisition unit and a training set generation unit;

[0049] The authentication acquisition unit initiates a communication request to the intelligent household appliance based on the monitoring center, and acquires the interaction-related data of the intelligent household appliance after obtaining user authorization, including: the original data containing the first instruction issued by the user, the response code for the intelligent household appliance to execute the first instruction, the user behavior data after the intelligent household appliance executes the first instruction, and the original data of the second instruction issued by the user per unit time;

[0050] The training set generation unit is used to process the acquired interaction-related data to generate a machine training set and store the machine training set in the buffer.

[0051] Further, the intelligent processing module includes a machine learning unit and a function update unit;

[0052] The machine learning unit is used to randomly extract training data from the machine training set for learning, and update the parameters of the target network based on the online network, so that the interaction ability of the intelligent household appliance better meets the user preferences;

[0053] The function update unit is used to deploy the trained target network and apply the updated interaction function to the intelligent household appliance.

[0054] Compared with the prior art, the beneficial effects achieved by the present invention are:

[0055] By monitoring the usage data of intelligent household appliances, faults in the interaction behavior can be discovered in a timely manner and users can be notified in a timely manner, thereby reducing the problems encountered by users during use and improving the user experience.

[0056] Through monitoring data and machine learning algorithms, potential faults in smart home appliances can be predicted, reducing the need for repairs and avoiding a decrease in the usage frequency of smart home appliances due to faults.

[0057] By analyzing user interaction data and analyzing the interaction data between users and smart home appliances, the interaction recognition section of smart home appliances can be updated based on user preference data, generating personalized responses for user behavior execution, improving the accuracy of the interaction between smart home appliances and users, solving the interaction behavior faults of smart home appliances, and improving the performance and user satisfaction of smart home appliances. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] The drawings are used to provide a further understanding of the present invention and form a part of the specification. They are used together with the embodiments of the present invention to explain the present invention and do not constitute a limitation to the present invention. In the drawings:

[0059] Figure 1 is a schematic structural diagram of the Internet-based smart home appliance fault monitoring, acquisition and processing system of the present invention;

[0060] Figure 2 is a schematic flow diagram of the Internet-based smart home appliance fault monitoring, acquisition and processing method of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0061] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0062] Please refer to Figure 1 , in the first embodiment: Provide an Internet-based smart home appliance fault monitoring, acquisition and processing system, including the following modules: a home appliance monitoring module, a fault detection module, a data acquisition module, and an intelligent processing module;

[0063] The home appliance monitoring module is used to monitor the usage data of smart home appliances and obtain data related to the interaction between users and smart home appliances; it includes a monitoring and acquisition unit and a data transmission unit; the monitoring and acquisition unit is used to obtain the response data and user reaction data during the interaction of smart home appliances; the data transmission unit is used to perform security processing on the data obtained by the monitoring and acquisition unit and transmit it to the monitoring center.

[0064] The fault detection module is used to detect the monitoring data output by the home appliance monitoring module and determine whether there are faults in the interaction behavior of smart home appliances; it includes a data analysis unit and a fault discrimination unit;

[0065] The data analysis unit is used to process the data received by the monitoring center to generate a dataset to be measured. The dataset to be measured includes n groups of data to be measured, and each group of data to be measured includes four elements, namely the first instruction, the executed response code, the user behavior category score, and the second instruction; the u-th group of data to be measured in the dataset to be measured is denoted as D u =(J u , X u , Y u , Z u ), u ∈ [1, n];

[0066] The fault discrimination unit is used to detect each group of data to be measured in the dataset to be measured according to the fault detection model, and determine whether there is a fault in the interaction behavior of the smart home appliance by calculating the user feedback score of the data to be measured. When the user feedback scores of more than 30% of the data to be measured in the dataset to be measured exceed the preset threshold, it is determined that there is a fault in the interaction behavior of the smart home appliance.

[0067] Among them, when the fault detection model detects each group of data to be measured in the dataset to be measured, the user feedback score Score of the u-th group of data to be measured in the dataset to be measured is calculated according to the following formula u :

[0068]

[0069] Among them, Score u represents the user feedback score obtained by the smart home appliance after executing the response behavior X u ;

[0070] represents the relationship function between the response behavior code executed by the smart home appliance and the first instruction. When the instruction received by the smart home appliance corresponds to the executed response behavior code, When the instruction received by the smart home appliance does not correspond to the executed response behavior code, g(Z u ) represents the relationship function between the second instruction issued by the user after the smart home appliance executes the response behavior and the first instruction. When , g(Z u ) = 0. When Z u ∈J u , g(Z u ) = 1. When , g(Z u ) = 2; is the weight coefficient;

[0071] Indicates the user behavior category score corresponding to the j-th behavior data of the smart home appliance after the smart home appliance executes the first instruction; ω j Indicates the influence coefficient corresponding to the user behavior category to which the j-th behavior data belongs; j ∈ [1, m], where m represents the number of user behaviors on the smart home appliance within a unit time after the smart home appliance executes the first instruction.

[0072] The data acquisition module is used to initiate a communication request to the smart home appliance, collect the interaction-related data of the smart home appliance after obtaining user authorization, and generate a machine training set; the data acquisition module includes an authentication acquisition unit and a training set generation unit;

[0073] Based on the monitoring center, the authentication acquisition unit initiates a communication request to the smart home appliance, and after obtaining user authorization, collects the interaction-related data of the smart home appliance, including: the original data containing the first instruction issued by the user, the response code of the smart home appliance executing the first instruction, the user behavior data of the smart home appliance after executing the first instruction, and the original data of the second instruction issued by the user within a unit time;

[0074] The training set generation unit is used to process the collected interaction-related data, generate a machine training set, and store the machine training set in the buffer.

[0075] The intelligent processing module optimizes the interaction function of the smart home appliance based on the machine training set, and processes the interaction function faults of the smart home appliance through machine learning; it includes a machine learning unit and a function update unit;

[0076] The machine learning unit is used to randomly extract training data from the machine training set for learning, and update the parameters of the target network based on the online network, so that the interaction ability of the smart home appliance better meets user preferences;

[0077] The function update unit is used to deploy the trained target network and apply the updated interaction function to the smart home appliance.

[0078] Please refer to Figure 2 , in the second embodiment, a method for monitoring, collecting and processing faults of an Internet-based smart home appliance is provided, including the following steps:

[0079] Step S100: Monitor the interaction behavior of the smart home appliance, obtain the first instruction data issued by the user received by the smart home appliance, and the execution result data returned to the system after executing the instruction according to the preset conditions; obtain the user performance after the smart home appliance executes the instruction, including the operation data and the second instruction data of the user on the smart home appliance after the smart home appliance executes the instruction;

[0080] Exemplarily, the operation data of the smart home appliance can be to adjust the mode or parameters of the smart home appliance through the APP or the remote control;

[0081] The acquired data is encrypted using encryption algorithms such as AES, RSA, etc., and the encrypted data is transmitted using secure communication protocols such as HTTPS, TLS, etc., and local backups are made for the uploaded data.

[0082] Step S200: Process the received data through the monitoring center, and discriminate the interaction behavior of the smart home appliance based on the fault detection model;

[0083] S201: Process the received data through the monitoring center to generate a dataset to be tested, where the dataset to be tested includes n groups of data to be tested, and the u-th group of data to be tested in the dataset to be tested is denoted as D u =(J u , X u , Y u , Z u ), u ∈ [1, n]; Each group of data to be tested includes four elements, namely the first instruction, the executed response code, the user behavior category score, and the second instruction; among them, the user behavior categories include the first behavior category, the second behavior category, and the third behavior category; the first behavior category means that the user uses the same instruction output method as the first instruction when operating, the second behavior category means that the user uses a different instruction output method from the first instruction when operating, and the third behavior category means that the user does not output an instruction when operating; specifically, the output methods of the first instruction and the second instruction include voice input, gesture input, home appliance button input, APP input, remote control input, etc.;

[0084] S202: Detect each group of data to be tested in the dataset to be tested through the fault detection model, and calculate the user feedback score Score u :

[0085]

[0086] where Score u represents the user feedback score obtained by the smart home appliance after executing the response behavior X u ;

[0087] represents the relationship function between the response behavior code executed by the smart home appliance and the first instruction. When the instruction received by the smart home appliance corresponds to the executed response behavior code, When the instruction received by the smart home appliance does not correspond to the executed response behavior code, g(Z u ) represents the relationship function between the second instruction issued by the user after the smart home appliance executes the response behavior and the first instruction. When , g(Z u ) = 0, when Zu ∈ J u When, g(Z u ) = 1, when When, g(Z u ) = 2; is the weight coefficient;

[0088] It should be noted that when Z u ∈ J u , it means that the second instruction is a subordinate instruction of the first instruction and will not change the state of the smart home appliance after executing the first instruction; When, it means that the second instruction is a parallel or superior instruction of the first instruction, and after executing the second instruction, the state of the smart home appliance after executing the first instruction will be overwritten.

[0089] represents the user behavior category score corresponding to the j-th behavior data of the user for the smart home appliance after the smart home appliance executes the first instruction; ω j represents the influence coefficient corresponding to the user behavior category to which the j-th behavior data belongs; j ∈ [1, m], where m represents the number of user behaviors for the smart home appliance within a unit time after the smart home appliance executes the first instruction;

[0090] S203: According to the detection result of the fault detection model, if the user feedback scores of more than 30% of the test data in the test data set exceed the preset threshold e, it is determined that there is a fault in the interaction behavior of the smart home appliance, and step S300 is executed.

[0091] Step S300: The monitoring center sends a communication request to the smart home appliance, and after obtaining user authorization, collects the interaction-related data of the smart home appliance to generate a machine training set;

[0092] S301: The monitoring center sends a communication request to the smart home appliance, and after obtaining user authorization, collects the interaction-related data of the smart home appliance, including: the original data of the user sending the first instruction, the response code of the smart home appliance executing the first instruction, the user behavior data of the smart home appliance after executing the first instruction, and the original data of the user sending the second instruction within a unit time;

[0093] S302: Process the collected interaction-related data to generate a machine training set, which contains W groups of training data, and any group of training data is denoted as (J, x, y, z); store the machine training set in the buffer. The four elements in the training data respectively correspond to the original data of the user sending the first instruction, the response code of the smart home appliance executing the first instruction, the user behavior data of the smart home appliance after executing the first instruction, and the original data of the user sending the second instruction within a unit time after data normalization.

[0094] Step S400: Optimize the interaction function of the smart home appliance based on the machine training set, and process the interaction function failure of the smart home appliance by means of machine learning;

[0095] S401: Randomly extract training data from the machine training set for learning, and calculate any group of the extracted training data according to the following formula:

[0096]

[0097] where, P i represents the target value of the i-th group of training data, r represents the user feedback score obtained after the smart home appliance executes the response behavior y, λ represents the discount factor, and θ' represents the parameters of the target network; Q(z, y'; θ') represents the target value of taking the optimal response y' under the next instruction z calculated by the target network using the parameters θ';

[0098] Specifically, the above target network is implemented based on the Bellman equation;

[0099] S402: Calculate the predicted value based on the neural network, and update the parameters of the online network by minimizing the loss function. The specific formula is:

[0100] Q(J, x; θ) = f(J; θ);

[0101]

[0102] where, f is the neural network, and the calculation of the neural network is realized through matrix multiplication and vector addition, and is automatically completed by PyTorch in this embodiment; J represents the original data feature of the user issuing the first instruction, x represents the response behavior encoding executed by the smart home appliance, and θ represents the parameters of the neural network; Q(J, x; θ) represents the predicted value of the smart home appliance executing the response behavior x under the instruction J represented by the original data feature of the user issuing the first instruction; L represents the loss function, N represents the number of groups of the extracted training data, and i ∈ [1, N];

[0103] By analyzing the relationship between the instruction represented by the original data feature of the user issuing the first instruction and the response behavior encoding executed by the smart home appliance, the meaning of the instruction issued by the user is parsed, the accuracy of the smart home appliance for extracting instructions is improved, the invalid operation caused by incorrect instruction recognition is avoided, and the effective response efficiency of the smart home appliance to the user instruction is enhanced.

[0104] S403: Loop through steps S401 and S402. When the training reaches 10,000 times or the stop condition is met, copy the parameters of the online network to the target network; apply the updated interaction function to the smart home appliances. Through the above steps, the interaction recognition section of the smart home appliances can be updated according to the user's preference data, generate personalized response executions for the user's behavior, improve the accuracy of the interaction between the smart home appliances and the user, solve the interaction behavior faults of the smart home appliances, and optimize the user experience.

[0105] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device.

[0106] Finally, it should be noted that the above are only the preferred embodiments of the present invention and are not used to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. An Internet-based intelligent home appliance fault monitoring, acquisition and processing method, characterized in that, It includes the following steps: Step S100: Monitor the interaction behavior of smart home appliances, obtain the response data and user reaction data during the interaction of smart home appliances, and upload the data to the monitoring center after security processing; Step S200: Process the received data through the monitoring center, and discriminate the interaction behavior of smart home appliances based on the fault detection model. When it is determined that there is a fault in the interaction behavior of smart home appliances, execute Step S300; Step S300: The monitoring center sends a communication request to the smart home appliance, and after obtaining user authorization, collects the interaction-related data of the smart home appliance to generate a machine training set; Step S400: Optimize the interaction function of the smart home appliance based on the machine training set, and process the interaction function faults of the smart home appliance through machine learning; The specific steps of Step S200 include the following steps: S201: The monitoring center processes the received data to generate a dataset to be tested, where the dataset to be tested includes n groups of data to be tested. Denote the u-th group of data to be tested in the dataset to be tested as D u =(J u , X u , Y u , Z u ), where u ∈ [1, n]; among them, J u represents the first instruction received by the smart home appliance in the u-th group of data to be tested, X u represents the response code executed by the smart home appliance in the u-th group of data to be tested, Y u represents the user behavior category score after the smart home appliance executes the first instruction in the u-th group of data to be tested, and Z u represents the second instruction received by the smart home appliance in the u-th group of data to be tested; Among them, the user behavior categories include the first behavior category, the second behavior category, and the third behavior category; the first behavior category means that the user uses the same instruction output method as when issuing the first instruction during the operation, the second behavior category means that the user uses a different instruction output method from when issuing the first instruction during the operation, and the third behavior category means that the user does not output an instruction during the operation; S202: Detect each group of data to be tested in the dataset to be tested through the fault detection model, and calculate the user feedback score Score of the u-th group of data to be tested in the dataset to be tested according to the following formula u : Among them, Score u represents the user feedback score obtained by the smart home appliance after executing the response behavior X u ; Represents the relationship function between the response behavior code executed by the smart home appliance and the first instruction. When the instruction received by the smart home appliance corresponds to the executed response behavior code, When the instruction received by the smart home appliance does not correspond to the executed response behavior code, g(Z u ) represents the relationship function between the second instruction issued by the user after the smart home appliance executes the response behavior and the first instruction. When , g(Z u ) = 0. When Z u ∈J u , g(Z u ) = 1. When , g(Z u ) = A, where A > 1; is the weight coefficient; represents the user behavior category score corresponding to the j-th behavior data of the smart home appliance after the smart home appliance executes the first instruction; ω j represents the influence coefficient corresponding to the user behavior category to which the j-th behavior data belongs; j ∈ [1, m], where m represents the number of behaviors of the user on the smart home appliance within a unit time after the smart home appliance executes the first instruction; S203: According to the detection result of the fault detection model, when the user feedback scores of more than B% of the test data in the test data set exceed the preset threshold e, it is determined that there is a fault in the interaction behavior of the smart home appliance, and Step S300 is executed.

2. The method for monitoring, collecting and processing faults of intelligent household appliances based on the Internet according to claim 1, characterized in that The response data during the interaction of the smart home appliance is specifically the first instruction data sent by the user received by the smart home appliance, and the execution result data returned to the system after the instruction is executed according to the preset conditions; the user reaction data is used to feedback the user performance after the smart home appliance executes the instruction, including user behavior data and second instruction data; The specific process of uploading the data after security processing is to encrypt it using an encryption algorithm before data transmission, transmit the encrypted data using a secure communication protocol, and make a local backup of the uploaded data; The monitoring center is used to detect the interaction behavior of smart home appliances and determine whether there is a fault in the interaction behavior of smart home appliances.

3. The method for fault monitoring, acquisition and processing of intelligent household appliances based on the Internet according to claim 1, wherein, The specific steps of Step S300 include the following steps: S301: The monitoring center sends a communication request to the smart home appliance, and after obtaining user authorization, collects the interaction-related data of the smart home appliance, including: the original data of the first instruction issued by the user, the response code of the smart home appliance executing the first instruction, the user behavior data of the smart home appliance after executing the first instruction, and the original data of the second instruction issued by the user per unit time; S302: Process the collected interaction-related data to generate a machine training set, which contains W groups of training data, and any group of training data is denoted as (J, x, y, z); the four elements in the training data correspond in sequence to the original data of the first instruction issued by the user after data normalization, the response code of the smart home appliance executing the first instruction, the user behavior data of the smart home appliance after executing the first instruction, and the original data of the second instruction issued by the user per unit time; store the machine training set in the buffer.

4. The method for monitoring, collecting and processing faults of intelligent household appliances based on the Internet according to claim 1, wherein, The specific steps of Step S400 include the following steps: S401: Randomly select training data from the machine training set for learning, and calculate any set of the selected training data according to the following formula: Among them, P i represents the target value of the i-th group of training data, r represents the user feedback score obtained after the smart home appliance executes the response behavior y, λ represents the discount factor, and θ' represents the parameters of the target network; Q(z, y ’ ; θ') represents the target value of the optimal response y ’ adopted under the next instruction z calculated by the target network using the parameter θ'; S402: Calculate the predicted value based on the neural network, and update the parameters of the online network by minimizing the loss function. The specific formula is: Q(J, x; θ) = f(J; θ); where f is the neural network, J represents the original data features of the first instruction issued by the user, x represents the response behavior encoding executed by the smart home appliance, θ represents the parameters of the neural network; Q(J, x; θ) represents the predicted value of the smart home appliance executing the response behavior x under the instruction J; L represents the loss function, N represents the number of groups of training data extracted, and i ∈ [1, N]; S403: Loop steps S401 and S402. After k times of training, copy the parameters of the online network to the target network; apply the updated interaction function to the smart home appliance.

5. An Internet-based intelligent home appliance fault monitoring, acquisition and processing system, which executes the Internet-based intelligent home appliance fault monitoring, acquisition and processing method according to any one of claims 1-4, characterized in that, It includes the following modules: home appliance monitoring module, fault detection module, data acquisition module, and intelligent processing module; The home appliance monitoring module is used to monitor the usage data of the smart home appliance and obtain the data related to the interaction between the user and the smart home appliance; The fault detection module is used to detect the monitoring data output by the home appliance monitoring module and judge whether there is a fault in the interaction behavior of the smart home appliance; The data acquisition module is used to send a communication request to the smart home appliance, and after obtaining user authorization, collect the data related to the interaction of the smart home appliance and generate a machine training set; The intelligent processing module optimizes the interaction function of the smart home appliance based on the machine training set and processes the interaction function faults of the smart home appliance through machine learning.

6. The internet-based intelligent home appliance fault monitoring, acquisition and processing system according to claim 5, characterized in that: The home appliance monitoring module includes a monitoring and acquisition unit and a data transmission unit; The monitoring and acquisition unit is used to obtain the response data and user reaction data of the smart home appliance during interaction; The data transmission unit is used to securely process the data obtained by the monitoring and acquisition unit and transmit it to the monitoring center.

7. The Internet-based intelligent home appliance fault monitoring, acquisition and processing system according to claim 5, characterized in that: The fault detection module includes a data analysis unit and a fault discrimination unit; The data analysis unit is used to process the data received by the monitoring center and generate a dataset to be tested. The dataset to be tested includes n groups of data to be tested, and each group of data to be tested includes four elements, namely the first instruction, the executed response encoding, the user behavior category score, and the second instruction; The fault discrimination unit is used to detect each group of data to be tested in the dataset to be tested according to the fault detection model, and judge whether there is a fault in the interaction behavior of the smart home appliance by calculating the user feedback score of the data to be tested.

8. The internet-based intelligent home appliance fault monitoring, acquisition and processing system according to claim 5, characterized in that: The data acquisition module includes an authentication and acquisition unit and a training set generation unit; The authentication and acquisition unit initiates a communication request to the smart home appliance based on the monitoring center, and after obtaining user authorization, collects the data related to the interaction of the smart home appliance, including: the original data containing the first instruction issued by the user, the response encoding of the smart home appliance executing the first instruction, the user behavior data after the smart home appliance executes the first instruction, and the original data of the second instruction issued by the user per unit time; The training set generation unit is used to process the collected data related to the interaction, generate a machine training set, and store the machine training set in the buffer.

9. The internet-based intelligent home appliance fault monitoring, acquisition and processing system according to claim 5, characterized in that: The intelligent processing module includes a machine learning unit and a function update unit; The machine learning unit is used to randomly extract training data from the machine training set for learning and update the parameters of the target network based on the online network; The function update unit is used to deploy the trained target network and apply the updated interaction function to intelligent household appliances.

Images