A home computing network host application method and system

Abstract

The application relates to the field of artificial intelligence and smart home technology, and specifically provides a home algorithm network host application method and system. After the home algorithm network host is powered on, a voice module, a semantic understanding module, an AI monitoring module, a system control module and a communication module are started in sequence; the voice module loads a wake-up word model and VAD parameters, and is ready to receive user voice input; the semantic understanding module realizes semantic analysis and context understanding; the AI monitoring module starts a camera and an environment sensor, realizes real-time collection of video streams and environment data; the system control module initializes a local device interface, and provides interface support for home appliance control; the communication module establishes an internal message queue and an external network interface, and provides guarantee for data interaction among modules and external device communication. Compared with the prior art, the application can construct a high-performance, low-delay, safe and reliable home smart algorithm network host, and provides an integrated solution for home environment control and home security.

Description

Technical Field

[0001] This invention relates to the fields of artificial intelligence and smart home technology, specifically providing a method and system for applying a home computing host. Background Technology

[0002] With the rapid development of smart home and artificial intelligence technologies, voice assistants and home security systems have been widely used in daily life. However, existing smart home devices generally suffer from problems such as fragmented functions, high response latency, and insufficient intelligence. On the one hand, most common voice assistants on the market rely on cloud computing for voice recognition and semantic understanding. When the network is unstable or the latency is high, the voice response speed drops significantly, severely affecting the interactive experience. At the same time, uploading voice data to the cloud also brings potential privacy risks. On the other hand, traditional home monitoring systems can only record video and perform simple motion detection, lacking the ability to analyze images at the semantic level. They cannot intelligently identify the behavioral patterns of family members or sudden events, such as falls, prolonged stillness, or abnormal sounds, resulting in a low level of security intelligence.

[0003] Furthermore, voice assistants and monitoring systems are currently typically two independent modules, lacking a unified computing and communication architecture, leading to resource redundancy and control chaos. For example, data cannot be shared between the voice control module and the video monitoring module, hindering cross-modal integrated judgment and automatic decision-making. This fragmented system architecture not only increases equipment costs and power consumption but also limits the further development of home intelligence.

[0004] Therefore, there is an urgent need for a home computing host system that integrates speech recognition, semantic understanding, and visual intelligent analysis capabilities, enabling it to have localized AI computing power and multimodal data processing capabilities, and to achieve integrated functions such as voice control, environmental perception, behavior recognition, and anomaly detection, thereby improving the response speed and reliability of home intelligent systems while ensuring privacy and security. Summary of the Invention

[0005] This invention addresses the shortcomings of the prior art by providing a practical method for using a home network host.

[0006] A further technical objective of this invention is to provide a reasonably designed, secure, and applicable home computing host application system.

[0007] The technical solution adopted by this invention to solve its technical problem is: A method for applying a home computing network host, wherein after the home computing network host is powered on, the voice module, semantic understanding module, AI monitoring module, system control module and communication module are started sequentially; The voice module loads the wake word model and VAD parameters, preparing to receive user voice input; The semantic understanding module implements semantic parsing and contextual understanding; The AI ​​monitoring module activates the camera and environmental sensors to achieve real-time acquisition of video streams and environmental data. The system control module initializes the local device interface to provide interface support for the control of home appliances. The communication module establishes an internal message queue and an external network interface to ensure data interaction between modules and communication with external devices.

[0008] Furthermore, in the speech module, VAD detection is used to determine whether the input audio stream contains a valid speech signal, based on an energy threshold + spectral entropy analysis method. ; Where E(t) is the energy of the current frame, H(t) is the spectral entropy, and α and β are empirical parameters; The wake word model uses a combination of small model inference and keyword fuzzy matching algorithm to perform fuzzy matching and recognition of user wake words: ; x is the input speech sequence, y is the wake-up word sequence, LCS(x, y) is the longest common subsequence length, and S is the threshold. The wake-up sensitivity is changed by adjusting the threshold.

[0009] Furthermore, the semantic understanding module recognizes the intent of the voice command, identifies the specific behavioral instructions, and implements each operation through HTTP interface interaction.

[0010] Furthermore, the video input terminal of the AI ​​monitoring module uses a camera to collect data. After image preprocessing, the data is input into the Qwen multimodal model for behavior recognition. In sleep detection, the detection of key points on the human body determines whether the person is in a static or lying position. In reading detection, the detection of the hand and book area comprehensively determines the reading state. In fall detection, the vertical displacement velocity change rate Δy of key points and the aspect ratio AR of the human body are used to determine abnormalities. ; , These represent the height and width of the human body, respectively, and Δy represents the vertical displacement of the key point. The threshold is used to determine a fall event when AR drops sharply and Δy exceeds the threshold.

[0011] Furthermore, the home computing host adopts an inter-thread message queue communication mechanism, cross-module logical interaction is carried out through standardized API calls, external network interfaces use the ZigBee protocol for communication, and support local HTTP REST interfaces for easy access by third parties.

[0012] A home computing host application system, wherein after the home computing host is powered on, the voice module, semantic understanding module, AI monitoring module, system control module and communication module are started sequentially; The voice module loads the wake word model and VAD parameters, preparing to receive user voice input; The semantic understanding module implements semantic parsing and contextual understanding; The AI ​​monitoring module activates the camera and environmental sensors to achieve real-time acquisition of video streams and environmental data. The system control module initializes the local device interface to provide interface support for the control of home appliances. The communication module establishes an internal message queue and an external network interface to ensure data interaction between modules and communication with external devices.

[0013] Furthermore, in the speech module, VAD detection is used to determine whether the input audio stream contains a valid speech signal, based on an energy threshold + spectral entropy analysis method. ; Where E(t) is the energy of the current frame, H(t) is the spectral entropy, and α and β are empirical parameters; The wake word model uses a combination of small model inference and keyword fuzzy matching algorithm to perform fuzzy matching and recognition of user wake words: ; x is the input speech sequence, y is the wake-up word sequence, LCS(x, y) is the longest common subsequence length, and S is the threshold. The wake-up sensitivity is changed by adjusting the threshold.

[0014] Furthermore, the semantic understanding module recognizes the intent of the voice command, identifies the specific behavioral instructions, and implements each operation through HTTP interface interaction.

[0015] Furthermore, the video input terminal of the AI ​​monitoring module uses a camera to collect data. After image preprocessing, the data is input into the Qwen multimodal model for behavior recognition. In sleep detection, the detection of key points on the human body determines whether the person is in a static or lying position. In reading detection, the detection of the hand and book area comprehensively determines the reading state. In fall detection, the vertical displacement velocity change rate Δy of key points and the aspect ratio AR of the human body are used to determine abnormalities. ; , These are the height and width of the human body, respectively. This refers to the vertical displacement change at key points. The threshold is used to determine a fall event when AR drops sharply and Δy exceeds the threshold.

[0016] Furthermore, the home computing host adopts an inter-thread message queue communication mechanism, cross-module logical interaction is carried out through standardized API calls, external network interfaces use the ZigBee protocol for communication, and support local HTTP REST interfaces for easy access by third parties.

[0017] Compared with existing technologies, the home computing host application method and system of the present invention have the following outstanding advantages: This invention endows the system with localized AI computing power and multimodal data processing capabilities, enabling integrated functions such as voice control, environmental perception, behavior recognition, and anomaly detection. This improves the response speed and reliability of the home smart system while ensuring privacy and security. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating a method for using a home computer network host. Detailed Implementation

[0020] To enable those skilled in the art to better understand the present invention, the present invention will be further described in detail below with reference to specific embodiments. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] The following is a preferred embodiment: Example 1:

[0022] like Figure 1 As shown in this embodiment, a home computing host application method is described. After the home computing host is powered on, the voice module, semantic understanding module, AI monitoring module, system control module, and communication module are started in sequence. The voice module loads the wake word model and VAD parameters, preparing to receive user voice input; The semantic understanding module implements semantic parsing and contextual understanding; The AI ​​monitoring module activates the camera and environmental sensors to achieve real-time acquisition of video streams and environmental data. The system control module initializes the local device interface to provide interface support for the control of home appliances. The communication module establishes an internal message queue and an external network interface to ensure data interaction between modules and communication with external devices.

[0023] In the speech module, VAD detection is used to determine whether the input audio stream contains a valid speech signal, based on an energy threshold + spectral entropy analysis method. ; Where E(t) is the energy of the current frame, H(t) is the spectral entropy, and α and β are empirical parameters.

[0024] The wake word model uses a combination of small model inference and keyword fuzzy matching algorithm to perform fuzzy matching and recognition of user wake words: ; x is the input speech sequence, y is the wake-up word sequence, LCS(x, y) is the longest common subsequence length, and S is the threshold. The wake-up sensitivity is changed by adjusting the threshold.

[0025] The semantic understanding module integrates the Qwen multimodal large model to achieve semantic parsing and contextual understanding. This module can recognize the intent of voice commands, such as "turn on the light," engage in dialogue with the user, check the weather, etc., identify specific behavioral instructions, and implement various operations through HTTP interface interaction.

[0026] The AI ​​monitoring module uses a camera to collect video data. After image preprocessing, the data is input into the Qwen multimodal model for behavior recognition. In sleep detection, it uses key point detection to determine whether the person is in a static or lying position. In reading detection, it uses hand and book area detection to comprehensively determine the reading status. In fall detection, it uses the vertical displacement velocity change rate Δy of key points and the aspect ratio AR of the human body to determine abnormalities. ; , These are the height and width of the human body, respectively. This refers to the vertical displacement change at key points. The threshold is used to determine a fall event when AR drops sharply and Δy exceeds the threshold.

[0027] The home network host adopts an inter-thread message queue communication mechanism, cross-module logical interaction is carried out through standardized API calls, external network interfaces use ZigBee protocol communication, and support local HTTP REST interfaces for easy third-party access.

[0028] Example 2: After the home computing host is powered on, the voice module, semantic understanding module, AI monitoring module, system control module and communication module are started in sequence. The voice module loads the wake word model and VAD parameters, preparing to receive user voice input; The semantic understanding module implements semantic parsing and contextual understanding; The AI ​​monitoring module activates the camera and environmental sensors to achieve real-time acquisition of video streams and environmental data. The system control module initializes the local device interface to provide interface support for the control of home appliances. The communication module establishes an internal message queue and an external network interface to ensure data interaction between modules and communication with external devices.

[0029] After the user utters a wake word, the VAD (Voice Awareness Detection) system detects in real time whether there is a valid speech signal in the audio. If it is determined to be valid speech, the wake word model performs fuzzy matching to determine whether to trigger wake-up. After successful wake-up, the system enters the voice interaction activation state: the ASR (Automatic Speech Recognition) module converts the speech signal into text information and passes it to the Qwen multimodal big data model for semantic parsing; the semantic understanding module generates structured control commands, and the system control module sends operation commands to home devices or executes other commands, such as dialogue or checking the weather, through the API interface.

[0030] The AI ​​monitoring module continuously collects video streams and environmental sensor data, analyzing family member behavior and environmental conditions through multimodal fusion algorithms. The system can identify behaviors such as sleeping, reading, and falling; for abnormal events, such as falls, it makes real-time judgments, and the event judgment module sends the results to the system control module; the control module triggers voice prompts, automatically dials family contacts, or performs alarm operations according to preset rules, thus ensuring home safety.

[0031] The host has a built-in log management and health monitoring mechanism to record the running status and abnormal events of each module; it uses inter-thread message queues and API interfaces to decouple modules, improve concurrent processing capabilities and system maintainability; and it maintains stable system operation under high load conditions, achieving reliable smart home management.

[0032] Example 3: A home computing host application system, wherein after the home computing host is powered on, the voice module, semantic understanding module, AI monitoring module, system control module and communication module are started sequentially; The voice module loads the wake word model and VAD parameters, preparing to receive user voice input; The semantic understanding module implements semantic parsing and contextual understanding; The AI ​​monitoring module activates the camera and environmental sensors to achieve real-time acquisition of video streams and environmental data. The system control module initializes the local device interface to provide interface support for the control of home appliances. The communication module establishes an internal message queue and an external network interface to ensure data interaction between modules and communication with external devices.

[0033] In the speech module, VAD detection is used to determine whether the input audio stream contains a valid speech signal, based on an energy threshold + spectral entropy analysis method. ; Where E(t) is the energy of the current frame, H(t) is the spectral entropy, and α and β are empirical parameters.

[0034] The wake word model uses a combination of small model inference and keyword fuzzy matching algorithm to perform fuzzy matching and recognition of user wake words: ; x is the input speech sequence, y is the wake-up word sequence, LCS(x, y) is the longest common subsequence length, and S is the threshold. The wake-up sensitivity is changed by adjusting the threshold.

[0035] The semantic understanding module recognizes the intent of voice commands and identifies specific behavioral instructions, and then implements various operations through HTTP interface interaction.

[0036] The AI ​​monitoring module uses a camera to collect video data. After image preprocessing, the data is input into the Qwen multimodal model for behavior recognition. In sleep detection, it uses key point detection to determine whether the person is in a static or lying position. In reading detection, it uses hand and book area detection to comprehensively determine the reading status. In fall detection, it uses the vertical displacement velocity change rate Δy of key points and the aspect ratio AR of the human body to determine abnormalities. ; , These are the height and width of the human body, respectively. This refers to the vertical displacement change at key points. The threshold is used to determine a fall event when AR drops sharply and Δy exceeds the threshold.

[0037] The home network host adopts an inter-thread message queue communication mechanism, cross-module logical interaction is carried out through standardized API calls, external network interfaces use ZigBee protocol communication, and support local HTTP REST interfaces for easy third-party access.

[0038] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for applying a home computer network host, characterized in that, After the home computing host is powered on, the voice module, semantic understanding module, AI monitoring module, system control module and communication module are started in sequence. The voice module loads the wake word model and VAD parameters, preparing to receive user voice input; The semantic understanding module implements semantic parsing and contextual understanding; The AI ​​monitoring module activates the camera and environmental sensors to achieve real-time acquisition of video streams and environmental data. The system control module initializes the local device interface to provide interface support for the control of home appliances. The communication module establishes an internal message queue and an external network interface to ensure data interaction between modules and communication with external devices.

2. The method for applying a home computing network host according to claim 1, characterized in that, In the speech module, VAD detection is used to determine whether the input audio stream contains a valid speech signal, based on an energy threshold + spectral entropy analysis method. ； Where E(t) is the energy of the current frame, H(t) is the spectral entropy, and α and β are empirical parameters; The wake word model uses a combination of small model inference and keyword fuzzy matching algorithm to perform fuzzy matching and recognition of user wake words: ； x is the input speech sequence, y is the wake-up word sequence, LCS(x, y) is the longest common subsequence length, and S is the threshold. The wake-up sensitivity is changed by adjusting the threshold.

3. The method for applying a home computing network host according to claim 2, characterized in that, The semantic understanding module recognizes the intent of voice commands and identifies specific behavioral instructions, and then implements various operations through HTTP interface interaction.

4. The method for applying a home computing network host according to claim 3, characterized in that, The AI ​​monitoring module uses a camera to collect video data. After image preprocessing, the data is input into the Qwen multimodal model for behavior recognition. In sleep detection, the system uses key point detection to determine whether the user is in a static, lying position. In reading detection, the system uses hand and book area detection to comprehensively determine the reading status. In fall detection, the system uses the vertical displacement velocity change rate Δy of key points and the aspect ratio AR of the human body to determine abnormalities. ； , These represent the height and width of the human body, respectively, and Δy represents the vertical displacement of the key point. The threshold is used to determine a fall event when AR drops sharply and Δy exceeds the threshold.

5. The method for applying a home computing network host according to claim 4, characterized in that, The home network host adopts an inter-thread message queue communication mechanism, cross-module logical interaction is carried out through standardized API calls, external network interfaces use ZigBee protocol communication, and support local HTTP REST interfaces for easy third-party access.

6. A home computing network host application system, characterized in that, After the home computing host is powered on, the voice module, semantic understanding module, AI monitoring module, system control module and communication module are started in sequence. The voice module loads the wake word model and VAD parameters, preparing to receive user voice input; The semantic understanding module implements semantic parsing and contextual understanding; The AI ​​monitoring module activates the camera and environmental sensors to achieve real-time acquisition of video streams and environmental data. The system control module initializes the local device interface to provide interface support for the control of home appliances. The communication module establishes an internal message queue and an external network interface to ensure data interaction between modules and communication with external devices.

7. A home computing network host application system according to claim 6, characterized in that, In the speech module, VAD detection is used to determine whether the input audio stream contains a valid speech signal, based on an energy threshold + spectral entropy analysis method. ； Where E(t) is the energy of the current frame, H(t) is the spectral entropy, and α and β are empirical parameters; The wake word model uses a combination of small model inference and keyword fuzzy matching algorithm to perform fuzzy matching and recognition of user wake words: ； x is the input speech sequence, y is the wake-up word sequence, LCS(x, y) is the longest common subsequence length, and S is the threshold. The wake-up sensitivity is changed by adjusting the threshold.

8. A home computing network host application system according to claim 7, characterized in that, The semantic understanding module recognizes the intent of voice commands and identifies specific behavioral instructions, and then implements various operations through HTTP interface interaction.

9. A home computing network host application system according to claim 8, characterized in that, The AI ​​monitoring module uses a camera to collect video data. After image preprocessing, the data is input into the Qwen multimodal model for behavior recognition. In sleep detection, the system uses key point detection to determine whether the user is in a static, lying position. In reading detection, the system uses hand and book area detection to comprehensively determine the reading status. In fall detection, the system uses the vertical displacement velocity change rate Δy of key points and the aspect ratio AR of the human body to determine abnormalities. ； , These are the height and width of the human body, respectively. This refers to the vertical displacement change of key points. The threshold is used to determine a fall event when AR drops sharply and Δy exceeds the threshold.

10. A home computing network host application system according to claim 9, characterized in that, The home network host adopts an inter-thread message queue communication mechanism, cross-module logical interaction is carried out through standardized API calls, external network interfaces use ZigBee protocol communication, and support local HTTP REST interfaces for easy third-party access.

Images