A set-top box-based home cloud computing method and system

Abstract

The application discloses a kind of based on set top box's home cloud computing method and system, it is related to the home cloud computing field of set top box, including: set top box is deployed in home network as edge node, responsible for receiving the operation data of intelligent household appliance, and processing and cloud computing management server interaction task;Establish household appliance data dynamic upload mechanism, according to server state adaptive switching data upload mode;Set dynamic sliding window and attention mechanism, dynamically capture the timing law of user behavior;Based on neural network algorithm, dynamically regulate the operation mode of intelligent household appliance;Set standby control channel, detect whether there is connection exception between intelligent household appliance and set top box, dynamically adjust main control and standby control channel.The application has the advantages that: through load perception upload, timing behavior analysis and neural network regulation, while guaranteeing appliance intelligent regulation, through double-channel redundant control ensure system robustness, significantly improve response speed and reliability.

Description

A home cloud computing method and system based on set-top box Technical Field

[0001] This invention relates to the field of home cloud computing based on set-top boxes, and more specifically to a home cloud computing method and system based on set-top boxes. Background Technology

[0002] With the booming development of IoT and AI technologies, the digitalization and intelligentization of home scenarios are accelerating. The widespread application of smart home appliances has led to an increasing demand for data interaction between home devices. Users not only expect to remotely control home appliances, but also pursue the proactive sensing functions and personalized services of the devices. At the same time, the maturity of cloud computing technology has made it possible to centrally process and store home data. By connecting home devices to the cloud, distributed computing and storage resources can be integrated to improve the intelligent analysis capabilities and dynamic sensing capabilities of home devices.

[0003] Against this backdrop, set-top boxes, as core devices connecting televisions and external networks in home networks, are gradually expanding their functions from simple audio and video decoding to smart home control. Deploying set-top boxes as edge nodes enables preliminary data processing and interaction locally in the home, reducing reliance on the cloud, lowering network latency, and improving service response speed. However, existing home cloud computing solutions still have many shortcomings in user behavior prediction, intelligent control of home appliances, and network connection stability assurance, and more comprehensive technical solutions are urgently needed to meet users' higher expectations for smart home living.

[0004] Current home cloud computing technologies rely heavily on simple rule matching or preset scenarios to predict user behavior and control smart home appliances. They can only trigger tasks at fixed times or from a single sensor, and cannot proactively predict based on long-term behavioral patterns. This makes it difficult to accurately capture complex and ever-changing user behavior patterns. Furthermore, existing technologies are insufficient in handling connection anomalies between smart home appliances and set-top boxes. When network latency or signal interruption occurs, they often cannot detect and respond effectively, leading to problems such as lost appliance control commands, out-of-sync device status, data transmission congestion, and resource waste. This reduces the stability and reliability of the home smart system. Summary of the Invention

[0005] To address the aforementioned technical problems, this paper provides a home cloud computing method and system based on a set-top box. This technical solution solves the problems mentioned in the background technology, which rely heavily on simple rule matching or preset scenarios, can only execute tasks triggered by fixed times or single sensors, and cannot actively predict based on long-term user behavior patterns, making it difficult to accurately capture complex and ever-changing user behavior patterns. Secondly, existing technologies have insufficient ability to handle connection anomalies between smart home appliances and set-top boxes. When network latency, signal interruption, or other issues occur, they often cannot quickly detect and respond effectively, leading to problems such as lost appliance control commands, unsynchronized device status, data transmission congestion, and resource waste, thereby reducing the stability and reliability of the home smart system.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] A home cloud computing method based on a set-top box includes:

[0008] The set-top box is deployed as an edge node in the home network, responsible for receiving the operation data of smart home appliances and handling the interaction tasks with the cloud computing management server;

[0009] Establish a dynamic data upload mechanism for home appliances, monitor the load status of the cloud computing management server in real time, and adaptively switch the data upload mode according to the server status.

[0010] Based on historical data of user action timing, a dynamic sliding window and attention mechanism are set up to dynamically capture the timing patterns of user behavior;

[0011] Based on the temporal patterns of user behavior and using neural network algorithms, the operating modes of smart home appliances are dynamically adjusted according to real-time user behavior.

[0012] Set up a backup control channel to detect any connection issues between smart home appliances and the set-top box. If an issue is detected, activate the backup control channel to enable localized data storage and command control. Otherwise, continue using the smart control method with the set-top box as the primary control channel.

[0013] Preferably, the establishment of a dynamic data upload mechanism for home appliances, which monitors the load status of the cloud computing management server in real time and adaptively switches the data upload mode according to the server status, specifically includes:

[0014] Based on the load status of the cloud computing management server and the interaction feedback with the set-top box, feature items reflecting the status of the cloud computing management server are extracted.

[0015] The features include: CPU utilization, memory utilization, disk I / O read / write speed, network bandwidth utilization, number of tasks currently queued for processing, average task processing time, task success rate, server temperature, number of hardware failure warnings, number of software service anomalies, and the number of times the interaction latency with the set-top box exceeds the threshold and the number of disconnections.

[0016] Based on expert analysis and historical data, a scalar score is given to the operating status of the cloud computing management server to obtain the operating status label value of the cloud computing management server.

[0017] A machine learning algorithm based on traditional linear regression is used to establish a cloud computing management server operation status evaluation model to dynamically obtain the evaluation value of the cloud computing management server operation status.

[0018] Based on the evaluation values ​​of the cloud computing management server's operating status, a dynamic data upload mechanism for home appliances is established, and the data upload mode is adaptively switched according to the server status.

[0019] Preferably, the dynamic data upload mechanism for home appliances specifically includes:

[0020] Based on the evaluation value of the cloud computing management server's operating status, set the operating status level range of the cloud computing management server, which includes: stable status, normal status, fluctuating status and stressed status.

[0021] When the evaluation value of the cloud computing management server's running status is within the stable range, the real-time full upload mode is adopted, and the data is uploaded immediately after collection is completed.

[0022] When the evaluation value of the cloud computing management server's operating status is within the normal range, a short time delay window is set, the collected data is temporarily stored in the local cache area, and the data is uploaded according to the delay window.

[0023] When the evaluation value of the cloud computing management server's operating status is within the fluctuation range, the short-term delay window is extended, and the amount of data transmission is reduced through compression algorithms. Data is then uploaded according to the extended delay window.

[0024] When the evaluation value of the cloud computing management server's running status is within the tense state range, set the priority for data upload of each home appliance: high-priority data is uploaded in real time, medium-priority data is uploaded with a delay, and low-priority data is uploaded only after the local cache is stable.

[0025] The priority evaluation expression for the data upload of each home appliance is as follows:

[0026]

[0027] In the formula, For the first The priority evaluation value for data uploads from individual home appliances. , , The coefficients for the base weight, usage status weight, and appliance relevance weight are respectively obtained through the least squares method. For the first Mapping function for each home appliance tag , , These are the basic weight, usage status weight, and appliance relevance weight. The attenuation coefficient is... This refers to the average duration of continuous use of home appliances. A piecewise function for evaluating the functionality of home appliances. The number of related home appliances. For the first Home appliances and the first The correlation value of each household appliance For the first Each home appliance label.

[0028] Preferably, the step of setting up a dynamic sliding window and attention mechanism based on historical data of user action timing to dynamically capture the temporal patterns of user behavior specifically includes:

[0029] Based on historical data of user action time sequence, a sequence list of continuous user actions is established, in which the time sequence of user actions each day is treated as a sub-sequence;

[0030] Based on the time difference between adjacent actions, the attention weights of adjacent action pairs are calculated using a Gaussian function.

[0031] Based on the attention weights of adjacent actions and the sequence list of continuous user actions, obtain the weighted transition count of each adjacent action pair.

[0032] Calculate the probability that one action immediately follows the previous action in an adjacent action pair based on the weighted number of transitions in each adjacent action pair.

[0033] Based on the PID algorithm, an adaptive adjustment window is set according to the user's action frequency to improve the stickiness of user action data.

[0034] Preferably, the step of setting up a backup control channel, detecting whether there is a connection abnormality between the smart home appliance and the set-top box, and dynamically adjusting the main control and backup control channels specifically includes:

[0035] Based on the heartbeat detection mechanism, the abnormal frequency and duration of the heartbeat response of smart home appliances and set-top boxes are obtained to determine the abnormality index of each home appliance.

[0036] Based on the Pearson correlation formula, the correlation between the functions of each smart home appliance and the set-top box is determined, and this correlation value is used as the weight of the abnormal state of each home appliance and the set-top box.

[0037] Based on the monitoring data of the home network, determine the network latency index and assess the impact of network latency on the abnormal status of home appliances and set-top boxes;

[0038] Based on the anomaly index of home appliances, the weight of the abnormal state of home appliances and set-top boxes, and the network latency index, an anomaly feedback formula for home appliances and set-top boxes is constructed to determine the connection status of smart home appliances and set-top boxes.

[0039] A backup control channel is set up. When the set-top box is detected to be in an abnormal state, the backup control channel is activated to realize local data storage and command control. Otherwise, the intelligent control method with the set-top box as the main control channel continues to be used.

[0040] The expression for the anomaly index of the home appliance is as follows:

[0041]

[0042] In the formula, For the first Abnormality index of individual home appliances , These represent the average number of historical heartbeat failures and their duration, respectively. , These are the standard deviations of the number of historical heartbeat failures and their duration, respectively. For safety factor, is a constant term. As a balancing term, For the first The number of failed heartbeats per unit time for each household appliance For the first The duration of the current continuous heartbeat failure of each home appliance;

[0043] The expression for the network latency metric is:

[0044]

[0045] In the formula, For the first Network latency metrics at any given time For constant terms, For the first RTT value at time 10:00 This is the historical average RTT value, which is a fixed value. To preset the maximum acceptable RTT value, The standard deviation of RTT fluctuation within one hour. The standard deviation of the baseline RTT fluctuation;

[0046] The formula expression for the electrical and set-top box anomaly feedback is:

[0047]

[0048] In the formula, This is a comprehensive abnormal indicator for home appliances. The number of smart home appliances connected to the set-top box. For the first The weighting of abnormal states of individual home appliances and set-top boxes. It is the hyperbolic tangent function. , , The constant coefficients, It is the absolute value symbol. This represents the probability of an abnormal status for the set-top box. This represents the network packet loss rate.

[0049] Furthermore, this solution proposes a set-top box-based home cloud computing system to implement the set-top box-based home cloud computing method described above, including:

[0050] The data interaction module is used to deploy the set-top box as an edge node in the home network, and is responsible for receiving the operation data of smart home appliances and processing the interaction tasks with the cloud computing management server.

[0051] The intelligent upload module is used to establish a dynamic data upload mechanism for home appliances, monitor the load status of the cloud computing management server in real time, and adaptively switch the data upload mode according to the server status.

[0052] The dynamic control module is used to set up a dynamic sliding window and attention mechanism based on historical data of user action time sequence to dynamically capture the time sequence pattern of user behavior; and based on the time sequence pattern of user behavior, and based on neural network algorithm, dynamically control the operation mode of smart home appliances according to real-time user behavior.

[0053] A backup control module is used to set up a backup control channel and detect whether there is a connection abnormality between the smart home appliance and the set-top box. If so, the backup control channel is activated to realize local data storage and command control. If not, the smart control method with the set-top box as the main control channel continues to be used.

[0054] Preferably, the dynamic control module includes:

[0055] A behavior capture unit is used to dynamically capture the temporal patterns of user behavior based on historical data of user action timing, by setting up a dynamic sliding window and an attention mechanism.

[0056] The intelligent control unit is used to dynamically adjust the operating mode of smart home appliances based on the temporal patterns of user behavior and a neural network algorithm, according to the real-time behavior of the user.

[0057] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0058] This invention provides a home cloud computing method based on a set-top box. By using the set-top box as an edge node, it intelligently controls the interaction between smart home appliance operation data and the cloud computing management server. This reduces the burden on the cloud computing server while improving the real-time performance and efficiency of data processing. Based on the load status of the cloud computing management server and the interaction feedback with the set-top box, it establishes a cloud computing management server operation status evaluation model and a dynamic data upload mechanism for home appliances. The data upload mode is adaptively switched according to the server status to ensure the timeliness and effectiveness of data upload while avoiding server overload. Furthermore, based on historical data of user action sequences, it obtains the probability of adjacent actions relative to the next action immediately following the previous one, and based on a PID algorithm, it calculates the probability of user action frequency. To improve user engagement, an adaptive adjustment window is set to increase the stickiness of user action data. This allows for real-time analysis of user behavior using an LSTM neural network algorithm. Based on actual user needs and environmental changes, the operating status of home appliances is automatically adjusted, enhancing user experience and energy efficiency. Finally, to ensure the stability of data transmission between the set-top box and smart home appliances, preventing data loss due to network fluctuations or set-top box malfunctions, a backup control channel is implemented. Based on the connection status between the smart home appliances and the set-top box, the main and backup control channels are dynamically adjusted. This achieves load-sensing uploading, time-series behavior analysis, and neural network regulation. While ensuring intelligent control of home appliances, dual-channel redundancy control ensures system robustness, significantly improving response speed and reliability. Attached Figure Description

[0059] Figure 1 is a flowchart of a home cloud computing method based on a set-top box according to the present invention;

[0060] Figure 2 is a flowchart of the present invention for establishing a dynamic data upload mechanism for home appliances, monitoring the load status of the cloud computing management server in real time, and adaptively switching the data upload mode according to the server status.

[0061] Figure 3 is a flowchart of the present invention, which uses historical data based on user action time sequence, sets up a dynamic sliding window and attention mechanism to dynamically capture the time sequence pattern of user behavior.

[0062] Figure 4 is a flowchart of the present invention for setting up a backup control channel, detecting whether there is a connection abnormality between smart home appliances and set-top boxes, and dynamically adjusting the main control and backup control channels. Detailed Implementation

[0063] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0064] Referring to Figure 1, a home cloud computing method based on a set-top box includes:

[0065] The set-top box is deployed as an edge node in the home network, responsible for receiving the operation data of smart home appliances and handling the interaction tasks with the cloud computing management server;

[0066] Establish a dynamic data upload mechanism for home appliances, monitor the load status of the cloud computing management server in real time, and adaptively switch the data upload mode according to the server status.

[0067] Based on historical data of user action timing, a dynamic sliding window and attention mechanism are set up to dynamically capture the timing patterns of user behavior;

[0068] Based on the temporal patterns of user behavior and using neural network algorithms, the operating modes of smart home appliances are dynamically adjusted according to real-time user behavior.

[0069] Set up a backup control channel to detect any connection issues between smart home appliances and the set-top box. If an issue is detected, activate the backup control channel to enable localized data storage and command control. Otherwise, continue using the smart control method with the set-top box as the primary control channel.

[0070] This can be explained by the fact that by analyzing user behavior patterns and determining the temporal regularity of user behavior, the operating status of home appliances can be automatically adjusted according to the actual needs of users and changes in the environment, thereby improving user experience and energy efficiency. While receiving real-time operating data from smart home appliances through the set-top box and conducting in-depth analysis through a cloud computing management server, it is necessary to ensure stable connections between the smart home appliances and the set-top box, and between the set-top box and the cloud computing management server, and to prevent data loss. Secondly, when analyzing the temporal regularity of user behavior, the influence of time windows and action windows, as well as the attention mechanisms of behavioral habits, must be considered simultaneously. Therefore, This solution dynamically adjusts the operating mode of smart home appliances based on the input, combining adaptive window adjustment and attention mechanisms, and utilizing the LSTM neural network algorithm. It also constructs anomaly feedback formulas for appliances and the set-top box based on appliance anomaly indicators, the weights of appliance and set-top box anomalies, and network latency. This determines the connection status between the smart appliances and the set-top box, thereby dynamically adjusting the main and backup control channels. Through load sensing uploads, timing behavior analysis, and neural network control, this solution ensures intelligent appliance control while maintaining system robustness through dual-channel redundant control, significantly improving response speed and reliability.

[0071] The deployment of the set-top box as an edge node in the home network, responsible for receiving operational data from smart home appliances and handling interaction tasks with the cloud computing management server, specifically includes:

[0072] The set-top box scans for smart home appliances in the home network that can be used for cloud computing, obtains the resource information of the smart home appliances, encapsulates it into a registration request, and sends it to the home cloud computing management server to complete the home appliance registration.

[0073] The home cloud computing management server assigns a unique identifier to each registered smart home appliance based on the registration request, marks it as registered, and establishes a smart home appliance management database.

[0074] The set-top box is used as an edge node, responsible for receiving the operating data of smart home appliances and handling the interaction tasks with the cloud computing management server.

[0075] Based on the smart home appliance management database, it is used to store and classify the operation data of smart home appliances as well as the interaction data between the set-top box and the cloud computing management server.

[0076] This can be explained by the fact that the set-top box, as a crucial edge node for receiving operational data from smart home appliances and processing interactions with the cloud computing management server, plays a vital role in establishing the connection and data interaction process between the set-top box, smart home appliances, and the cloud computing management server. By scanning the home network, the set-top box can automatically discover and identify smart home appliances that can access the cloud, acquire their resource information, encapsulate it into a registration request, and send it to the home cloud computing management server to complete the registration process for the home appliances. This ensures that the home appliances can quickly and conveniently access the system and establish a smart home appliance management database, thereby efficiently storing and classifying this data, providing strong support for subsequent data analysis and intelligent control.

[0077] Referring to Figure 2, the establishment of a dynamic data upload mechanism for home appliances, which monitors the load status of the cloud computing management server in real time and adaptively switches the data upload mode according to the server status, specifically includes:

[0078] Based on the load status of the cloud computing management server and the interaction feedback with the set-top box, feature items reflecting the status of the cloud computing management server are extracted.

[0079] The features include: CPU utilization, memory utilization, disk I / O read / write speed, network bandwidth utilization, number of tasks currently queued for processing, average task processing time, task success rate, server temperature, number of hardware failure warnings, number of software service anomalies, and the number of times the interaction latency with the set-top box exceeds the threshold and the number of disconnections.

[0080] Based on expert analysis and historical data, a scalar score is given to the operating status of the cloud computing management server to obtain the operating status label value of the cloud computing management server.

[0081] A machine learning algorithm based on traditional linear regression is used to establish a cloud computing management server operation status evaluation model to dynamically obtain the evaluation value of the cloud computing management server operation status.

[0082] Based on the evaluation values ​​of the cloud computing management server's operating status, a dynamic data upload mechanism for home appliances is established, and the data upload mode is adaptively switched according to the server status.

[0083] This can be explained by accurately assessing the operational status of the cloud computing management server, enabling adaptive adjustment of the home appliance data upload mode, and improving the overall efficiency and stability of the home cloud computing system. Specifically, it extracts multi-dimensional features such as CPU utilization and task success rate from server load status and interaction feedback with the set-top box to comprehensively reflect the server's operational status; it uses expert analysis combined with historical data to score the server status scalar and assign it a quantitative label value, making the status assessment more objective; it constructs an operational status assessment model based on traditional linear regression machine learning algorithms to dynamically obtain assessment values, ensuring real-time monitoring and analysis of the server status; and finally, it establishes a dynamic upload mechanism based on the assessment values, adaptively switching the data upload mode according to different server statuses, avoiding data transmission congestion and task backlog caused by excessive server load, or resource waste caused by inappropriate upload modes. While reducing server pressure, it ensures the timeliness and reliability of home appliance data upload, optimizes the resource allocation of the home cloud computing system, enhances the system's stability and reliability, and provides users with a smoother and more efficient home cloud computing service experience.

[0084] The dynamic data upload mechanism for home appliances specifically includes:

[0085] Based on the evaluation value of the cloud computing management server's operating status, set the operating status level range of the cloud computing management server, which includes: stable status, normal status, fluctuating status and stressed status.

[0086] When the evaluation value of the cloud computing management server's running status is within the stable range, the real-time full upload mode is adopted, and the data is uploaded immediately after collection is completed.

[0087] When the evaluation value of the cloud computing management server's operating status is within the normal range, a short time delay window is set, the collected data is temporarily stored in the local cache area, and the data is uploaded according to the delay window.

[0088] When the evaluation value of the cloud computing management server's operating status is within the fluctuation range, the short-term delay window is extended, and the amount of data transmission is reduced through compression algorithms. Data is then uploaded according to the extended delay window.

[0089] When the evaluation value of the cloud computing management server's running status is within the tense state range, set the priority for data upload of each home appliance: high-priority data is uploaded in real time, medium-priority data is uploaded with a delay, and low-priority data is uploaded only after the local cache is stable.

[0090] The priority evaluation expression for the data upload of each home appliance is as follows:

[0091]

[0092] In the formula, For the first The priority evaluation value for data uploads from individual home appliances. , , The coefficients for the base weight, usage status weight, and appliance relevance weight are respectively obtained through the least squares method. For the first Mapping function for each home appliance tag , , These are the basic weight, usage status weight, and appliance relevance weight. The attenuation coefficient is... This refers to the average duration of continuous use of home appliances. A piecewise function for evaluating the functionality of home appliances. The number of related home appliances. For the first Home appliances and the first The correlation value of each household appliance For the first Each home appliance label.

[0093] This can be explained by the fact that when the cloud computing management server's operational status assessment value is within the "stressed" range, it indicates that the server's acceptable data volume and transmission rate are extremely limited. Performing a real-time full upload under this condition can easily lead to data congestion and data loss, and in more severe cases, cause the cloud computing management server to crash. Therefore, to ensure timely upload of important data while reducing the upload of non-critical data, this solution reduces the operational pressure on the cloud computing management server and improves data transmission stability and integrity by first storing non-critical data locally. Thus, this solution establishes a dynamic data upload mechanism for home appliances, intelligently uploading data according to the cloud computing management server's operational status range. Finally, by setting the upload priority for each home appliance, high-priority data is uploaded in real-time, medium-priority data is uploaded with a delay, and low-priority data is uploaded only after local caching has stabilized, reducing data transmission problems during the stressed period. For the first Home appliances and the first The correlation value of each household appliance can be obtained using the Pearson correlation formula. , , The solution can be obtained using the least squares method. Here, "usage status" refers to the state of the appliance under human use or intervention. For example, the usage status of a refrigerator includes dynamic data related to human use, such as refrigerator door opening times and function adjustments. For the first The mapping function for each appliance label aims to uniformly map the original label to 1 (representing a valid label) while retaining the type information of the original label.

[0094] Referring to Figure 3, the specific steps of setting up a dynamic sliding window and attention mechanism based on historical data of user action timing to dynamically capture the temporal patterns of user behavior include:

[0095] Based on historical data of user action time sequence, a sequence list of continuous user actions is established, in which the time sequence of user actions each day is treated as a sub-sequence;

[0096] Based on the time difference between adjacent actions, the attention weights of adjacent action pairs are calculated using a Gaussian function.

[0097] Based on the attention weights of adjacent actions and the sequence list of continuous user actions, obtain the weighted transition count of each adjacent action pair.

[0098] Calculate the probability that one action immediately follows the previous action in an adjacent action pair based on the weighted number of transitions in each adjacent action pair.

[0099] Based on the PID algorithm, an adaptive adjustment window is set according to the user's action frequency to improve the stickiness of user action data.

[0100] This can be explained by calculating the attention weights of adjacent action pairs using a Gaussian function, highlighting action pairs with short time intervals and strong correlations, and obtaining the weighted transition count and action occurrence probability to further quantify the transition relationship between actions, clearly presenting user behavior patterns, thereby obtaining an evaluation of the user's behavior from one action to the next. Simultaneously, the PID algorithm is used to adaptively adjust the window based on the frequency of user actions, ensuring the data collection range closely matches changes in user behavior, improving the alignment between data and actual behavior. This allows the home smart system to proactively learn user habits, anticipate user needs, and adjust the operating mode of smart home appliances before actual user operation, providing personalized services such as pre-activating air conditioning and automatically adjusting lights, improving the intelligence and convenience of home life, and enhancing user experience and satisfaction with the smart home system. The expression for calculating the attention weights of adjacent action pairs is:

[0101]

[0102] In the formula, The attention weights for adjacent action pairs, The time difference between adjacent action pairs. The standard deviation of time sensitivity;

[0103] The expression for obtaining the weighted transition count of each adjacent action pair is:

[0104]

[0105] In the formula, For adjacent action pairs The number of weighted transfers, The total length of the action sequence. For the first Attention weights for next-next action pairs For indicator functions, when The value is 1 if it is true, and 0 otherwise. Set labels for adjacent action pairs;

[0106] The probability expression for the adjacent action to the next action immediately following the previous action is:

[0107]

[0108] In the formula, Let be the probability that an adjacent action is followed immediately by another adjacent action. For action tags Total number of times it appears in the sequence list.

[0109] The method of dynamically adjusting the operating mode of smart home appliances based on the temporal patterns of user behavior and using neural network algorithms specifically includes:

[0110] A synchronous data acquisition method is adopted to ensure the consistency of smart home appliance operation data and user real-time behavior data in terms of data bits;

[0111] The system acquires real-time operating data of smart home appliances and real-time user behavior data through a set-top box, and then filters and normalizes the data before using it as input to a neural network algorithm.

[0112] The probability of an adjacent action following the previous action is used as the attention mechanism of a neural network algorithm.

[0113] Based on the input, and combining adaptive window adjustment and attention mechanisms, the operating mode of smart home appliances is dynamically controlled using the LSTM neural network algorithm.

[0114] This can be explained by the fact that by acquiring real-time operating data of smart home appliances and real-time user behavior data through a set-top box, and combining adaptive window adjustment and attention mechanisms with the powerful time-series data processing capabilities of the LSTM neural network algorithm, it is possible to effectively learn the historical patterns of user behavior and make dynamic predictions based on real-time behavior data. This allows for precise control of the smart home appliance operating mode, overcoming the lack of flexibility and personalization in existing intelligent control technologies. This enables home appliances to proactively adapt to user habits and needs, creating a personalized experience.

[0115] Referring to Figure 4, the specific steps of setting up a backup control channel and dynamically adjusting the main and backup control channels to detect connection abnormalities between smart home appliances and the set-top box include:

[0116] Based on the heartbeat detection mechanism, the abnormal frequency and duration of the heartbeat response of smart home appliances and set-top boxes are obtained to determine the abnormality index of each home appliance.

[0117] Based on the Pearson correlation formula, the correlation between the functions of each smart home appliance and the set-top box is determined, and this correlation value is used as the weight of the abnormal state of each home appliance and the set-top box.

[0118] Based on the monitoring data of the home network, determine the network latency index and assess the impact of network latency on the abnormal status of home appliances and set-top boxes;

[0119] Based on the anomaly index of home appliances, the weight of the abnormal state of home appliances and set-top boxes, and the network latency index, an anomaly feedback formula for home appliances and set-top boxes is constructed to determine the connection status of smart home appliances and set-top boxes.

[0120] A backup control channel is set up. When the set-top box is detected to be in an abnormal state, the backup control channel is activated to realize local data storage and command control. Otherwise, the intelligent control method with the set-top box as the main control channel continues to be used.

[0121] The expression for the anomaly index of the home appliance is as follows:

[0122]

[0123] In the formula, For the first Abnormality index of individual home appliances , These represent the average number of historical heartbeat failures and their duration, respectively. , These are the standard deviations of the number of historical heartbeat failures and their duration, respectively. For safety factor, is a constant term. As a balancing term, For the first The number of failed heartbeats per unit time for each household appliance For the first The duration of the current continuous heartbeat failure of each home appliance;

[0124] The expression for the network latency metric is:

[0125]

[0126] In the formula, For the first Network latency metrics at any given time For constant terms, For the first RTT value at time 10:00 This is the historical average RTT value, which is a fixed value. To preset the maximum acceptable RTT value, The standard deviation of RTT fluctuation within one hour. The standard deviation of the baseline RTT fluctuation;

[0127] The formula expression for the electrical and set-top box anomaly feedback is:

[0128]

[0129] In the formula, This is a comprehensive abnormal indicator for home appliances. The number of smart home appliances connected to the set-top box. For the first The weighting of abnormal states of individual home appliances and set-top boxes. It is the hyperbolic tangent function. , , The constant coefficients, It is the absolute value symbol. This represents the probability of an abnormal status for the set-top box. This represents the network packet loss rate.

[0130] This can be explained by the fact that, to reduce the memory space occupied by local data storage, data is usually uploaded to a cloud computing management server first, thus reducing the need for local data storage. Therefore, when transmitting home appliance data, it is necessary to prioritize ensuring the stability of the connection between the home appliance and the set-top box. This is to avoid connection abnormalities caused by network anomalies or set-top box damage, which could lead to control failure or disorder of the set-top box over the home appliance, resulting in wasted resources. Therefore, this solution judges the connection status between the smart home appliance and the set-top box. When an abnormal connection status is detected, the home appliance data is stored locally through a backup control channel, and the home appliance is controlled by the backup control channel. The backup control channel refers to the set-top box releasing its main control authority and transferring control authority to each smart home appliance control unit or a local edge controller that does not require network access.

[0131] Furthermore, this solution proposes a set-top box-based home cloud computing system to implement the set-top box-based home cloud computing method described above, including:

[0132] The data interaction module is used to deploy the set-top box as an edge node in the home network, and is responsible for receiving the operation data of smart home appliances and processing the interaction tasks with the cloud computing management server.

[0133] The intelligent upload module is used to establish a dynamic data upload mechanism for home appliances, monitor the load status of the cloud computing management server in real time, and adaptively switch the data upload mode according to the server status.

[0134] The dynamic control module is used to set up a dynamic sliding window and attention mechanism based on historical data of user action time sequence to dynamically capture the time sequence pattern of user behavior; and based on the time sequence pattern of user behavior, and based on neural network algorithm, dynamically control the operation mode of smart home appliances according to real-time user behavior.

[0135] A backup control module is used to set up a backup control channel and detect whether there is a connection abnormality between the smart home appliance and the set-top box. If so, the backup control channel is activated to realize local data storage and command control. If not, the smart control method with the set-top box as the main control channel is continued to be used.

[0136] The dynamic control module includes:

[0137] A behavior capture unit is used to dynamically capture the temporal patterns of user behavior based on historical data of user action timing, by setting up a dynamic sliding window and an attention mechanism.

[0138] The intelligent control unit is used to dynamically adjust the operating mode of smart home appliances based on the temporal patterns of user behavior and a neural network algorithm, according to the real-time behavior of the user.

[0139] In summary, the advantages of this invention are: by using load sensing upload, time-series behavior analysis, and neural network regulation, it ensures intelligent control of home appliances while ensuring system robustness through dual-channel redundant control, significantly improving response speed and reliability.

[0140] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A home cloud computing method based on a set-top box, characterized in that, include: The set-top box is deployed as an edge node in the home network, responsible for receiving the operation data of smart home appliances and handling the interaction tasks with the cloud computing management server; Establish a dynamic data upload mechanism for home appliances, monitor the load status of the cloud computing management server in real time, and adaptively switch the data upload mode according to the server status. Based on historical data of user action sequences, a dynamic sliding window and attention mechanism are set up to dynamically capture the temporal patterns of user behavior. According to the temporal patterns of user behavior, based on neural network algorithms, the operating mode of smart home appliances is dynamically adjusted according to real-time user behavior. A backup control channel is set up to detect whether there is a connection abnormality between smart home appliances and set-top boxes. If so, the backup control channel is activated to realize local data storage and command control. If not, the smart control method with set-top boxes as the main control channel continues to be used. The specific steps of setting up a backup control channel to detect connection anomalies between smart home appliances and the set-top box, and dynamically adjusting the main and backup control channels, include: 1) Obtaining the abnormal frequency and duration of the heartbeat responses between the smart home appliances and the set-top box based on a heartbeat detection mechanism, and determining the anomaly index for each appliance; 2) Determining the correlation between the functions of each smart home appliance and the set-top box based on the Pearson correlation formula, and using this correlation value as the weight of the abnormal state of each appliance and set-top box; 3) Determining the network latency index based on home network monitoring data, and judging the impact of network latency on the abnormal state of the appliances and set-top box; 4) Constructing an anomaly feedback formula for the appliances and set-top box based on the appliance anomaly index, the weight of the abnormal state of the appliances and set-top box, and the network latency index, to judge the connection status of the smart home appliances and the set-top box; 5) Setting up a backup control channel, which is activated when the set-top box is detected to achieve localized data storage and command control; otherwise, continuing to use the smart control method with the set-top box as the main control channel; The expression for the appliance anomaly index is: In the formula, For the first Abnormality index of individual home appliances 、 These represent the average number of historical heartbeat failures and their duration, respectively. 、 These are the standard deviations of the number of historical heartbeat failures and their duration, respectively. For safety factor, is a constant term. As a balancing term, For the first The number of failed heartbeats per unit time for each household appliance For the first The duration of consecutive heartbeat failures for each household appliance; the expression for the network latency index is: In the formula, For the first Network latency metrics at any given time For constant terms, For the first RTT value at time 10:00 This is the historical average RTT value, which is a fixed value. To preset the maximum acceptable RTT value, The standard deviation of RTT fluctuation within one hour. The standard deviation of the baseline RTT fluctuation is used; the formula expression for the electrical and set-top box anomaly feedback is: In the formula, This is a comprehensive abnormal indicator for home appliances. The number of smart home appliances connected to the set-top box. For the first The weighting of abnormal states of individual home appliances and set-top boxes. It is the hyperbolic tangent function. 、 、 、 The constant coefficients, It is the absolute value symbol. This represents the probability of an abnormal status for the set-top box. This represents the network packet loss rate.

2. The home cloud computing method based on a set-top box according to claim 1, characterized in that, The deployment of the set-top box as an edge node in the home network, responsible for receiving operational data from smart home appliances and handling interaction tasks with the cloud computing management server, specifically includes: scanning the home network for smart home appliances available for cloud computing using the set-top box, obtaining resource information of the smart home appliances, encapsulating it into a registration request, and sending it to the home cloud computing management server to complete the appliance registration; the home cloud computing management server assigning a unique identifier to each registered smart home appliance based on the registration request, marking it as registered, and establishing a smart home appliance management database; the set-top box as an edge node responsible for receiving operational data from smart home appliances and handling interaction tasks with the cloud computing management server; and the smart home appliance management database used to store and classify the operational data of smart home appliances and the interaction data between the set-top box and the cloud computing management server.

3. The home cloud computing method based on a set-top box according to claim 2, characterized in that, The establishment of a dynamic data upload mechanism for home appliances, which monitors the load status of the cloud computing management server in real time and adaptively switches the data upload mode based on the server status, specifically includes: extracting feature items reflecting the status of the cloud computing management server based on its load status and interaction feedback with the set-top box; these feature items include: CPU utilization, memory utilization, disk I / O read / write speed, network bandwidth utilization, number of tasks currently queued for processing, average task processing time, task success rate, server temperature, number of hardware fault warnings, number of software service anomalies, and the number of times the interaction latency with the set-top box exceeds the threshold and the number of disconnections; based on expert analysis and combined with historical data, scalar scoring is performed on the operating status of the cloud computing management server to obtain a label value for the operating status; based on a machine learning algorithm using traditional linear regression, an evaluation model for the operating status of the cloud computing management server is established to dynamically obtain the evaluation value of the operating status; based on the evaluation value of the operating status of the cloud computing management server, a dynamic data upload mechanism for home appliances is established, adaptively switching the data upload mode according to the server status.

4. The home cloud computing method based on a set-top box according to claim 3, characterized in that, The dynamic data upload mechanism for home appliances specifically includes: setting a range of operating status levels for the cloud computing management server based on its evaluation value, where each level range includes: stable state, normal state, fluctuating state, and stressful state; when the evaluation value of the cloud computing management server's operating status is within the stable state range, a real-time full upload mode is adopted, and data is uploaded immediately after collection; when the evaluation value of the cloud computing management server's operating status is within the normal state range, a short-term delay window is set, the collected data is temporarily stored in the local cache area, and the data is uploaded according to the delay window; when the evaluation value of the cloud computing management server's operating status is within the fluctuating state range, the short-term delay window is extended, and a compression algorithm is used to reduce the amount of data transmitted, and the data is uploaded according to the extended delay window; when the evaluation value of the cloud computing management server's operating status is within the stressful state range, a priority is set for uploading data from each home appliance, with high-priority data uploaded in real time, medium-priority data uploaded with a delay, and low-priority data uploaded only after the local cache has stabilized; the priority evaluation expression for uploading data from each home appliance is: In the formula, For the first The priority evaluation value for data uploads from individual home appliances. 、 、 The coefficients for the base weight, usage status weight, and appliance relevance weight are respectively obtained through the least squares method. For the first Mapping function for each home appliance tag 、 、 These are the basic weight, usage status weight, and appliance relevance weight. The attenuation coefficient is... This refers to the average duration of continuous use of home appliances. A piecewise function for evaluating the functionality of home appliances. The number of related home appliances. For the first Home appliances and the first The correlation value of each household appliance For the first Each home appliance label.

5. A home cloud computing method based on a set-top box according to claim 4, characterized in that, The method of dynamically capturing the temporal patterns of user behavior by setting up a dynamic sliding window and attention mechanism based on historical user action time-series data specifically includes: establishing a sequence list of continuous user actions based on historical user action time-series data, wherein the daily user action time-series is treated as a subsequence; calculating the attention weight of adjacent action pairs based on the time difference between adjacent actions using a Gaussian function; obtaining the weighted transition count of each adjacent action pair based on the sequence list of continuous user actions and the attention weight of adjacent actions; calculating the probability that another action immediately follows the previous action of an adjacent action pair based on the weighted transition count of each adjacent action pair; and setting an adaptive adjustment window based on the user's action frequency using a PID algorithm to improve the stickiness of user action data.

6. The home cloud computing method based on a set-top box according to claim 5, characterized in that, The specific steps of dynamically adjusting the operating mode of smart home appliances based on the temporal patterns of user behavior and using a neural network algorithm include: employing a synchronous acquisition method to ensure consistency between the operating data of smart home appliances and the real-time user behavior data in terms of data bits; acquiring the operating data of smart home appliances and the real-time user behavior data in real time through a set-top box, and using the filtered and normalized data as input to the neural network algorithm; using the probability of one action following another adjacent action as the attention mechanism of the neural network algorithm; and dynamically adjusting the operating mode of smart home appliances based on the LSTM neural network algorithm, combining the adaptive window adjustment and attention mechanism according to the input.

7. A home cloud computing system based on a set-top box, characterized in that, The method for implementing a home cloud computing system based on a set-top box as described in any one of claims 1-6 comprises: a data interaction module, which deploys the set-top box as an edge node in the home network, is responsible for receiving operating data from smart home appliances, and handling interaction tasks with the cloud computing management server; an intelligent upload module, which establishes a dynamic data upload mechanism for home appliances, monitors the load status of the cloud computing management server in real time, and adaptively switches the data upload mode according to the server status; a dynamic control module, which sets up a dynamic sliding window and attention mechanism based on historical data of user action timing, dynamically captures the timing patterns of user behavior, and dynamically controls the operating mode of smart home appliances based on neural network algorithms according to real-time user behavior according to the timing patterns of user behavior; and a backup control module, which sets up a backup control channel, detects whether there is a connection abnormality between the smart home appliance and the set-top box, and if so, activates the backup control channel to realize localized data storage and command control; otherwise, continues to use the intelligent control method with the set-top box as the main control channel.

8. A home cloud computing system based on a set-top box according to claim 7, characterized in that, The dynamic control module includes: a behavior capture unit, which is used to dynamically capture the temporal patterns of user behavior by setting up a dynamic sliding window and an attention mechanism based on historical data of user action timing; and an intelligent control unit, which is used to dynamically control the operating mode of smart home appliances based on the temporal patterns of user behavior and a neural network algorithm, according to the real-time behavior of the user.

Images