Control method and control system of smart home system

Abstract

The invention discloses a control method and a control system of a smart home system, relates to the technical field of smart homes, and aims to solve the technical problem that a control system cannot recognize a corresponding instruction or mistakenly recognizes the corresponding instruction due to current traditional voice instruction control, and further causes unnecessary troubles. Comprising the following steps: S1, defining a multi-dimensional and refined film watching behavior mode; comprising body postures, facial expressions, eyeball movement, sound features, handheld device operation and the like, specific recognition standards and threshold values are set for each type of features, and a film watching behavior pattern library is formed; s2, behavior pattern acquisition; by integrating various high-precision sensors such as a high-definition camera, an infrared sensor, a pressure sensor and a microphone, multi-dimensional behavior data of a user in the film watching process are collected in real time and preprocessed to improve the data quality; and S3, behavior pattern recognition. The method has the advantage of improving the convenience and accuracy of intelligent home equipment control.

Description

Technical Field

[0001] The present invention relates to the field of smart home technology, and more specifically, to a control method and a control system for a smart home system. Background Art

[0002] Smart home is based on the house as a platform, using integrated wiring technology, network communication technology, security technology, automatic control technology, audio and video technology to integrate facilities related to home life, build an efficient management system for residential facilities and family schedule affairs, so as to improve home safety, convenience, comfort, artistry, and achieve an environmentally friendly and energy-saving living environment. Smart home is not a single product, but a technical means to connect various devices in the home into an organic system, so that the owner can control and manage it anytime and anywhere.

[0003] With the popularization of smart homes, users' demand for intelligent control of viewing experience is increasing, and the control systems of smart homes are becoming more and more perfect. At present, the control systems of smart homes mainly control indoor smart furniture by receiving voice commands. Users need to actively send commands to the control systems of smart homes. For users in different regions, their accents are often different, and non-standard command input will cause the control system to be unable to recognize the corresponding commands, or misrecognize the corresponding commands, which will cause unnecessary troubles. In view of this, we propose a control method and control system for a smart home system. Summary of the invention

[0004] The purpose of the present invention is to provide a control method and control system for a smart home system to solve the technical problem that the current traditional voice command control easily causes the control system to be unable to recognize the corresponding command, or misrecognizes the corresponding command, thereby causing unnecessary troubles.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: a control method of a smart home system, comprising the following steps:

[0006] S1. Multi-dimensional and refined definition of movie-watching behavior patterns;

[0007] Including body posture, facial expression, eye movement, voice characteristics and handheld device operation, etc., set specific recognition standards and thresholds for each type of feature to form a viewing behavior pattern library;

[0008] S2, behavioral pattern collection;

[0009] By integrating high-definition cameras, infrared sensors, pressure sensors, microphones and other high-precision sensors, the multi-dimensional behavior data of users during the viewing process is collected in real time, and pre-processed to improve data quality;

[0010] S3, behavioral pattern recognition;

[0011] A behavior pattern recognition system based on deep learning and hybrid machine learning models is used to receive pre-processed data input and output the judgment results of whether the user is in the movie-watching mode and the specific behavior pattern. The SVM classification algorithm is used for recognition. The specific algorithm formula is:

[0012]

[0013] Where MovieModePredicted is the predicted movie mode state, m is the mode category, i.e., movie mode or non-movie mode, n is the number of support vectors, αi and yi are the coefficients and labels of the support vectors, respectively, K(xuser, xi) is the kernel function used to calculate the similarity between the user's current state xuser and the support vector xi, and b is the bias term;

[0014] S4, intelligent furniture linkage control;

[0015] According to the results of behavioral pattern recognition, the status of smart home devices, including curtains, lights, speakers, air conditioners, etc., are automatically adjusted to create the best viewing environment;

[0016] S5, long-term learning and adaptive optimization;

[0017] Establish a long-term learning and optimization mechanism based on reinforcement learning, continuously collect user feedback, behavior data and device status information, and iteratively optimize the behavior pattern recognition model and control strategy;

[0018] Use the policy gradient algorithm to adjust the model parameters to improve the recognition accuracy and control effect. The policy gradient algorithm formula is as follows:

[0019]

[0020] Where θ is the model parameter, α is the learning rate, πθ(at|st) is the policy to take action a in state st, Rt is the immediate reward, γ is the discount factor, rt+k is the future reward, b(st) is the baseline function used to reduce variance, and T is the time step.

[0021] The present invention collects the user's behavior pattern through a behavior pattern collection module, and identifies the user's specific behavior through a behavior pattern recognition module, utilizes a behavior pattern recognition system of degree learning and hybrid machine learning models, uses an SVM classification algorithm for recognition, and automatically adjusts the state of smart home devices according to the recognition results, thereby realizing behavior pattern control of smart home devices, abandoning the traditional voice control mode, effectively reducing the occurrence of unrecognizable or misrecognized commands, and is conducive to improving the convenience and accuracy of smart home device control.

[0022] Preferably, in S2, data preprocessing includes denoising, feature extraction, data fusion and time series analysis.

[0023] Preferably, in S3, the behavior pattern recognition also includes consideration of individual differences such as user habits, age, gender, etc., to achieve personalized recognition.

[0024] Preferably, in S4, the smart home device linkage control strategy includes dynamically adjusting the opening and closing angle of curtains, light brightness and color temperature, speaker volume and sound effects, air conditioning temperature and wind speed.

[0025] A control system of a smart home system, comprising a data acquisition unit for collecting multi-dimensional behavioral data of a user during a movie-watching process, the data acquisition unit being connected to a data processing unit for preprocessing the data, the data processing unit being connected to a behavior pattern recognition unit for identifying whether the user is in a movie-watching mode and his or her specific behavior pattern, the behavior pattern recognition unit being connected to an intelligent decision-making and recognition unit for automatically formulating and executing control strategies for smart home devices based on an output result of the behavior pattern recognition unit, the intelligent decision-making and recognition unit being connected to a long-term learning and adaptive optimization unit for performing long-term learning training and adaptive optimization adjustment on behavior pattern recognition and intelligent decision-making and recognition, and the system also comprising a user interaction and feedback unit.

[0026] Preferably, the data acquisition unit integrates a high-definition camera, an infrared sensor, a pressure sensor and a microphone to capture and collect multi-dimensional behavioral data of the user during the viewing process in real time, including but not limited to body posture, facial expressions, eye movements, sound characteristics and handheld device operations.

[0027] Preferably, the data processing unit performs preprocessing operations such as denoising, feature extraction, data fusion and time series analysis on the raw data transmitted from the data acquisition unit to improve data quality and analyzability.

[0028] Preferably, the behavior pattern recognition unit adopts deep learning and hybrid machine learning algorithms to construct a complex behavior pattern recognition model, and conducts in-depth analysis on the processed data to identify with high accuracy whether the user is in the movie-watching mode and his or her specific behavior pattern.

[0029] Preferably, the intelligent decision-making and recognition unit automatically formulates and executes control strategies for smart home devices based on the output results of the behavior pattern recognition unit, including dynamically adjusting the opening and closing angles of curtains, the brightness and color temperature of lights, the volume and sound effects of speakers, the temperature and wind speed of air conditioners, etc., so as to create an optimal viewing environment.

[0030] Preferably, the long-term learning and adaptive optimization unit has the ability of self-learning and optimization, and uses reinforcement learning algorithms to adjust and optimize behavior pattern recognition models and control strategies by continuously collecting user behavior data, environmental parameters and system feedback, so as to achieve continuous improvement of personalized services and system performance.

[0031] Compared with the prior art, the present invention has the following beneficial effects:

[0032] 1. The present invention collects the user's behavior pattern through a behavior pattern collection module, and identifies the user's specific behavior through a behavior pattern recognition module, utilizes a behavior pattern recognition system of degree learning and hybrid machine learning models, uses an SVM classification algorithm for recognition, and automatically adjusts the state of smart home devices according to the recognition results, thereby realizing behavior pattern control of smart home devices, abandoning the traditional voice control mode, and effectively reducing the occurrence of commands that cannot be recognized or misrecognized, which is beneficial to improving the convenience and accuracy of smart home device control, and solves the problem that traditional voice command control easily causes the control system to be unable to recognize the corresponding commands, or misrecognizes the corresponding commands, thereby causing unnecessary troubles.

[0033] 2. The present invention also designs a long-term learning and adaptive optimization unit, establishes a long-term learning and optimization mechanism based on reinforcement learning, uses a policy gradient algorithm to adjust model parameters, and continuously trains behavior pattern recognition to improve recognition accuracy and control effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a schematic diagram of the system framework of the present invention;

[0035] Figure 2 It is a schematic diagram of the system framework of the data acquisition unit of the present invention;

[0036] Figure 3 A schematic diagram of the system framework of a data processing unit of the present invention;

[0037] Figure 4 It is a schematic diagram of the system framework of the behavior pattern recognition unit of the present invention;

[0038] Figure 5 A schematic diagram of the system framework of the intelligent decision-making and identification unit of the present invention;

[0039] Figure 6 Schematic diagram of the system framework of the long-term learning and adaptive optimization unit of the present invention. DETAILED DESCRIPTION

[0040] In order to facilitate those skilled in the art to understand the technical solution of the present invention, the technical solution of the present invention is further described in conjunction with the accompanying drawings of the specification.

[0041] A control method for a smart home system comprises the following steps:

[0042] S1. Multi-dimensional and refined definition of movie-watching behavior patterns;

[0043] Including body posture, facial expression, eye movement, voice characteristics and handheld device operation, etc., set specific recognition standards and thresholds for each type of feature to form a viewing behavior pattern library;

[0044] S2, behavioral pattern collection;

[0045] By integrating high-definition cameras, infrared sensors, pressure sensors, microphones and other high-precision sensors, the multi-dimensional behavior data of users during the viewing process is collected in real time, and pre-processed to improve data quality;

[0046] S3, behavioral pattern recognition;

[0047] A behavior pattern recognition system based on deep learning and hybrid machine learning models is used to receive pre-processed data input and output the judgment results of whether the user is in the movie-watching mode and the specific behavior pattern. The SVM classification algorithm is used for recognition. The specific algorithm formula is:

[0048]

[0049] Where MovieModePredicted is the predicted movie mode state, m is the mode category, i.e., movie mode or non-movie mode, n is the number of support vectors, αi and yi are the coefficients and labels of the support vectors, respectively, K(xuser, xi) is the kernel function used to calculate the similarity between the user's current state xuser and the support vector xi, and b is the bias term;

[0050] S4, intelligent furniture linkage control;

[0051] According to the results of behavioral pattern recognition, the status of smart home devices, including curtains, lights, speakers, air conditioners, etc., are automatically adjusted to create the best viewing environment;

[0052] S5, long-term learning and adaptive optimization;

[0053] Establish a long-term learning and optimization mechanism based on reinforcement learning, continuously collect user feedback, behavior data and device status information, and iteratively optimize the behavior pattern recognition model and control strategy;

[0054] Use the policy gradient algorithm to adjust the model parameters to improve the recognition accuracy and control effect. The policy gradient algorithm formula is as follows:

[0055]

[0056] Where θ is the model parameter, α is the learning rate, πθ(at|st) is the policy to take action a in state st, Rt is the immediate reward, γ is the discount factor, rt+k is the future reward, b(st) is the baseline function used to reduce variance, and T is the time step.

[0057] In S2, data preprocessing includes denoising, feature extraction, data fusion and time series analysis. In S3, behavioral pattern recognition also includes consideration of individual differences such as user habits, age, gender, etc. to achieve personalized recognition. In S4, the linkage control strategy of smart home devices includes dynamic adjustment of curtain opening and closing angles, light brightness and color temperature, speaker volume and sound effects, air conditioning temperature and wind speed.

[0058] A control system for a smart home system, comprising a data acquisition unit for collecting multi-dimensional behavioral data of a user during a movie-watching process, the data acquisition unit being connected to a data processing unit for preprocessing the data, the data processing unit being connected to a behavior pattern recognition unit for recognizing whether the user is in a movie-watching mode and his or her specific behavior pattern, the behavior pattern recognition unit being connected to an intelligent decision-making and recognition unit for automatically formulating and executing control strategies for smart home devices based on output results of the behavior pattern recognition unit, the intelligent decision-making and recognition unit being connected to a long-term learning and adaptive optimization unit for performing long-term learning training and adaptive optimization adjustment on behavior pattern recognition and intelligent decision-making and recognition, and the system also comprising a user interaction and feedback unit.

[0059] like Figure 1-Figure 6As shown, the data acquisition unit integrates a high-definition camera, an infrared sensor, a pressure sensor and a microphone to capture and collect multi-dimensional behavioral data of users in the process of watching movies in real time, including but not limited to body posture, facial expression, eye movement, sound characteristics and handheld device operation. The data processing unit performs pre-processing operations such as denoising, feature extraction, data fusion and time series analysis on the original data transmitted by the data acquisition unit to improve data quality and analyzability. The behavior pattern recognition unit adopts deep learning and hybrid machine learning algorithms to build a complex behavior pattern recognition model, and conducts in-depth analysis on the processed data to identify with high accuracy whether the user is in the movie-watching mode and his specific behavior pattern. The intelligent decision-making and recognition unit automatically formulates and executes the control strategy of the smart home device according to the output result of the behavior pattern recognition unit, including dynamically adjusting the opening and closing angle of the curtains, the brightness and color temperature of the lights, the volume and sound effects of the speakers, the temperature and wind speed of the air conditioner, etc., to create the best viewing environment. The long-term learning and adaptive optimization unit has the ability of self-learning and optimization. By continuously collecting user behavior data, environmental parameters and system feedback, the reinforcement learning algorithm is used to adjust and optimize the behavior pattern recognition model and control strategy to achieve personalized services and continuous improvement of system performance.

[0060] The embodiments of the present invention disclose preferred embodiments, but are not limited thereto. A person skilled in the art can easily understand the spirit of the present invention based on the above embodiments and make different extensions and changes. However, as long as they do not deviate from the spirit of the present invention, they are all within the protection scope of the present invention.

Claims

1. A control method for a smart home system, characterized in that: The following steps are involved: S1. Multi-dimensional and refined definition of movie-watching behavior patterns; Including body posture, facial expression, eye movement, voice characteristics and handheld device operation, etc., set specific recognition standards and thresholds for each type of feature to form a viewing behavior pattern library; S2, behavioral pattern collection; By integrating high-definition cameras, infrared sensors, pressure sensors, microphones and other high-precision sensors, the multi-dimensional behavior data of users during the viewing process is collected in real time, and pre-processed to improve data quality; S3, behavioral pattern recognition; A behavior pattern recognition system based on deep learning and hybrid machine learning models is used to receive pre-processed data input and output the judgment results of whether the user is in the movie-watching mode and the specific behavior pattern. The SVM classification algorithm is used for recognition. The specific algorithm formula is: Where MovieModePredicted is the predicted movie mode state, m is the mode category, i.e., movie mode or non-movie mode, n is the number of support vectors, αi and yi are the coefficients and labels of the support vectors, respectively, K(xuser, xi) is the kernel function used to calculate the similarity between the user's current state xuser and the support vector xi, and b is the bias term; S4, intelligent furniture linkage control; According to the results of behavioral pattern recognition, the status of smart home devices, including curtains, lights, speakers, air conditioners, etc., are automatically adjusted to create the best viewing environment; S5, long-term learning and adaptive optimization; Establish a long-term learning and optimization mechanism based on reinforcement learning, continuously collect user feedback, behavior data and device status information, and iteratively optimize the behavior pattern recognition model and control strategy; Use the policy gradient algorithm to adjust the model parameters to improve the recognition accuracy and control effect. The policy gradient algorithm formula is as follows: Where θ is the model parameter, α is the learning rate, πθ(at|st) is the policy to take action a in state st, Rt is the immediate reward, γ is the discount factor, rt+k is the future reward, b(st) is the baseline function used to reduce variance, and T is the time step.

2. A control method for a smart home system according to claim 1, characterized in that: In S2, data preprocessing includes denoising, feature extraction, data fusion and time series analysis.

3. The control method of a smart home system according to claim 1, characterized in that: In S3, the behavior pattern recognition also includes consideration of individual differences such as user habits, age, gender, etc., to achieve personalized recognition.

4. The control method of a smart home system according to claim 1, characterized in that: In S4, the smart home device linkage control strategy includes dynamically adjusting the opening and closing angle of curtains, light brightness and color temperature, speaker volume and sound effects, air conditioning temperature and wind speed.

5. A control system for a smart home system, characterized in that: The system comprises a data collection unit for collecting multi-dimensional behavior data of users in the process of watching a movie, the data collection unit is connected to a data processing unit for preprocessing the data, the data processing unit is connected to a behavior pattern recognition unit for identifying whether the user is in a movie-watching mode and his specific behavior pattern, the behavior pattern recognition unit is connected to an intelligent decision-making and recognition unit for automatically formulating and executing control strategies for smart home devices according to an output result of the behavior pattern recognition unit, the intelligent decision-making and recognition unit is connected to a long-term learning and adaptive optimization unit for long-term learning and training and adaptive optimization adjustment of behavior pattern recognition and intelligent decision-making and recognition, and the system also comprises a user interaction and feedback unit.

6. The control system of a smart home system according to claim 5, characterized in that: The data acquisition unit integrates a high-definition camera, an infrared sensor, a pressure sensor and a microphone, and is used to capture and collect multi-dimensional behavioral data of users during the viewing process in real time, including but not limited to body posture, facial expressions, eye movements, sound characteristics and handheld device operations.

7. The control system of a smart home system according to claim 6, characterized in that: The data processing unit performs preprocessing operations such as denoising, feature extraction, data fusion and time series analysis on the original data transmitted by the data acquisition unit to improve data quality and analyzability.

8. The control system of a smart home system according to claim 7, characterized in that: The behavior pattern recognition unit adopts deep learning and hybrid machine learning algorithms to build a complex behavior pattern recognition model, conducts in-depth analysis of the processed data, and identifies with high accuracy whether the user is in the movie-watching mode and his or her specific behavior pattern.

9. The control system of a smart home system according to claim 8, characterized in that: The intelligent decision-making and recognition unit automatically formulates and executes control strategies for smart home devices based on the output results of the behavior pattern recognition unit, including dynamically adjusting the opening and closing angles of curtains, the brightness and color temperature of lights, the volume and sound effects of speakers, the temperature and wind speed of air conditioners, etc., to create an optimal viewing environment.

10. The control system of a smart home system according to claim 9, characterized in that: The long-term learning and adaptive optimization unit has the ability of self-learning and optimization. By continuously collecting user behavior data, environmental parameters and system feedback, it uses reinforcement learning algorithms to adjust and optimize behavior pattern recognition models and control strategies to achieve continuous improvement in personalized services and system performance.

Images