A device control method, apparatus, electronic device, and storage medium

By constructing a cognitive model of the physical characteristics of building space, the problem of lack of perception of the physical characteristics of building space in the distributed intelligent device control scheme is solved, and precise device control is achieved, improving control accuracy and stability.

CN123067243APending Publication Date: 2026-11-10GREE ELECTRIC APPLIANCE INC OF ZHUHAI +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611156247.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-07-31
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

Existing distributed intelligent device control solutions lack perception and understanding of the physical characteristics of buildings, resulting in a disconnect between control actions and the real physical environment, reducing the accuracy and stability of regulation, and affecting the efficiency of multi-device collaborative management.

Method used

By constructing an initial spatial framework model that integrates the location, material, and volume of walls, a cognitive model of the physical characteristics of building space is trained and generated. This model accurately analyzes the impact of physical barriers and thermal properties of walls, and generates refined control commands to regulate edge nodes.

Benefits of technology

It enables targeted control based on the actual physical characteristics of buildings, eliminates business anomalies caused by wall obstruction or thermal characteristics, and improves the control accuracy and collaborative stability of distributed smart home networks in complex environments.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123067243A_ABST
    Figure CN123067243A_ABST
Patent Text Reader

Abstract

This invention discloses a device control method, apparatus, electronic device, and storage medium. It constructs an initial spatial framework model integrating wall location, material, and volume, and calibrates parameters such as wireless signal attenuation, thermal performance, and thermal inertia level of the walls to train and generate a cognitive model of the building's spatial physical characteristics. When abnormal events related to the physical characteristics of the building space occur in the service data of edge nodes, the central control terminal calls this cognitive model to analyze the impact of the walls on signal or heat conduction, generates and issues control commands matching the current physical environment, and achieves targeted regulation based on the actual physical characteristics of the building. This effectively eliminates service anomalies caused by wall obstruction or thermal characteristics, significantly improving control accuracy and collaborative stability.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of equipment control technology, and in particular to an equipment control method, an equipment control device, an electronic device, and a readable storage medium. Background Technology

[0002] In the fields of smart home and smart building technology, the distributed intelligent device control schemes of related technologies adopt a linear response mechanism in which a central control terminal issues commands and terminal devices passively execute them. The central control terminal typically responds passively based on a single metric; for example, when the temperature sensor reading exceeds a preset threshold, the system turns on the air conditioner; when the light sensor detects that the environment has darkened, the system turns on the lights. While this approach is feasible for handling isolated and simple control needs, it lacks the ability to perceive and understand the control results and the physical spatial characteristics of the building. This makes it difficult for the system to adaptively adjust to real-time intervention in the business operation status based on the physical environment, which can easily cause the control actions to become disconnected from and interfere with the real physical space environment. This not only reduces the accuracy and stability of terminal device control, but also seriously affects the efficiency of collaborative management of multiple devices and the actual user experience. Summary of the Invention The present invention provides a device control method, apparatus, electronic device, and readable storage medium to overcome or at least partially solve the above-mentioned problems.

[0003] To solve the above-mentioned technical problems, the present invention is implemented as follows: In a first aspect, embodiments of the present invention provide a device control method, the method being applied to a central control terminal located within a building, the building being divided into multiple functional spaces by multiple walls, and a distributed smart home network in which the central control terminal is located, comprising one or more edge nodes placed within the functional spaces, the method including: The location information and material identification data of the target wall are determined, and the volume of the functional space is calculated; the central control terminal is used to transmit and receive radio signals through the target wall and the edge nodes. An initial spatial framework model is constructed using the location information, the material identification data, and the volume size data. Determine the parameters of the target wall's blocking and attenuation of the radio signal, as well as the thermal performance parameters and thermal inertia level parameters of the target wall; The initial spatial frame model is trained based on the barrier attenuation parameter, the thermal performance parameter, and the thermal inertia level parameter to generate a cognitive model of the physical characteristics of the building space. The system monitors the service data of the edge node. When it is determined from the service data that an abnormal event related to the physical characteristics of the building space occurs in the current service of the edge node, the system calls the building space physical characteristics cognitive model to generate control instructions for the edge node and sends the control instructions to the edge node. The control instructions are used to control the edge node to execute the current service based on the adjustment parameters of the control instructions.

[0004] Optionally, the service data is the network transmission data of the edge node, and the step of calling the building space physical characteristic cognitive model to generate control instructions for the edge node includes: Determine the transmission rate, retransmission rate, and packet loss rate of the network data; When the building space physical characteristics cognitive model determines the event type of the abnormal event as network congestion caused by signal attenuation due to the radio signal passing through the target wall, based on the transmission rate, retransmission rate and packet loss rate, the building space physical characteristics cognitive model is invoked to retrieve other services of other edge nodes; When the building space physical characteristics cognitive model determines that the other services are high-bandwidth services, the building space physical characteristics cognitive model is invoked to determine the network channel of the other edge nodes. The system invokes the building space physical characteristics cognitive model to find other network channels besides the network channel and generates channel switching instructions for the other network channels; the channel switching instructions are used to control the edge node to switch to the other network channel to execute the current service.

[0005] Optionally, the edge node is a smart air conditioner, the business data is temperature sensing data, and the user's posture and micro-motion characteristic data, and the step of calling the building space physical characteristic cognitive model to generate control commands for the edge node includes: When the building space physical characteristics cognitive model determines the event type of the abnormal event as uneven body comfort caused by the building space physical characteristics based on the temperature sensing data and the posture and micro-motion characteristic data, the building space physical characteristics cognitive model is called to query the air conditioner installation coordinates, furniture layout coordinates and load-bearing wall external corner geometric parameters in the functional space corresponding to the smart air conditioner. The building space physical characteristics cognitive model is invoked to generate the air outlet temperature and swing angle for the smart air conditioner based on the air conditioner installation coordinates, the furniture layout coordinates, and the geometric structural parameters of the load-bearing wall corner. The control command for the smart air conditioner is generated using the outlet air temperature and the swing angle; the control command is used to control the smart air conditioner to perform air supply service based on the outlet air temperature and the swing angle.

[0006] Optionally, the edge node is a refrigeration device equipped with a compressor, the service data is the current operating frequency of the compressor, and the floor vibration spectrum data of the functional space corresponding to the refrigeration device, and the step of calling the building space physical characteristic cognitive model to generate control commands for the edge node includes: When the building space physical characteristics cognitive model determines the event type of the abnormal event as a structural resonance-type abnormality caused by the vibration characteristics of the compressor and the floor slab of the functional space based on the current operating frequency and the floor vibration spectrum data, the building space physical characteristics cognitive model is invoked to extract the main peak frequency in the floor vibration spectrum data, and a target safe operating frequency that avoids the main peak frequency is determined based on the main peak frequency and the current operating frequency. The target safe operating frequency is used to generate control commands for the refrigeration equipment; the control commands are used to control the refrigeration equipment to adjust the operating frequency of the compressor to the target safe operating frequency.

[0007] Optionally, the step of determining the location information and material identification data of the target wall, and calculating the volume of the functional space includes: Transmit broadband pulsed microwave signals to the edge nodes and receive the reflected echoes of the broadband pulsed microwave signals; Determine the time difference between the transmitted and received signals, the echo intensity, and the spectral characteristics of the reflected echo; Based on the time difference between the transmitted and received signals, the echo intensity, and the spectral characteristics, the location information and material identification data of the target wall are calculated. Collect ambient acoustic signals within the functional space; Determine the reverberation decay time corresponding to the environmental acoustic signal; The volume of the functional space is calculated based on the reverberation decay time.

[0008] Optionally, the step of determining the blocking attenuation parameter of the target wall to the radio signal, as well as the thermal performance parameters and thermal inertia level parameters of the target wall, includes: Test data packets are sent to the edge nodes of each of the functional spaces at a preset fixed transmission power, and the received signal strength indication data fed back by the edge nodes are obtained. The difference between the received signal strength indication data and the preset theoretical value of free space propagation loss is calculated to obtain the instantaneous blocking attenuation data. Obtain historical barrier attenuation data, and perform a weighted average calculation on the instant barrier attenuation data and the historical barrier attenuation data to generate the barrier attenuation parameter. Acquire outdoor meteorological data and indoor temperature change rate data within the functional space; Based on the outdoor meteorological data and the indoor temperature change rate data, the thermal insulation and heat storage capacity of the functional space is calculated. The thermal performance parameters and thermal inertia level parameters of the target wall are determined based on the heat insulation and heat storage capacity.

[0009] Optionally, the central control terminal is configured with a corresponding cloud server, and after the step of sending the control command to the edge node, it further includes: The service data is re-collected and updated according to the preset sampling period, and the updated service data is used to generate the performance evaluation result of the edge node executing the current service based on the control command. When the performance evaluation result is determined to meet the preset conditions, control experience sample data is generated based on the business data, the control instructions, and the performance evaluation result. The model data of the cognitive model of physical characteristics of building space is adjusted using the aforementioned regulatory experience sample data to obtain an updated cognitive model of physical characteristics of building space, and the updated cognitive model of physical characteristics of building space is confirmed as the first updated model. Extract the model gradient information generated during the training process of the first updated model; The model gradient information is encrypted using a preset encryption algorithm to generate ciphertext gradient data. The encrypted gradient data is uploaded to the cloud server; the cloud server is used to aggregate the encrypted gradient data from different central control terminals using a federated averaging algorithm, generate global model parameters, and send the global model parameters to the central control terminal. The global model parameters are received, and the first updated model is updated again using the global model parameters to generate the second updated model.

[0010] This invention provides a device control method applied to an edge node. In a distributed smart home network where the edge node is located, a corresponding central control terminal is configured. The central control terminal is placed inside a building, which is divided into multiple functional spaces by multiple walls. One or more edge nodes are deployed within each functional space. The central control terminal is used to: determine the location information and material identification data of a target wall, and calculate the volume of the functional space; transmit and receive radio signals with the edge nodes through the target wall; and construct an initial space using the location information, material identification data, and volume data. The method includes: establishing a spatial frame model; determining the attenuation parameters of the target wall for the radio signal, as well as the thermal performance parameters and thermal inertia level parameters of the target wall; training the initial spatial frame model based on the attenuation parameters, the thermal performance parameters, and the thermal inertia level parameters to generate a building space physical characteristic cognitive model; monitoring the service data of the edge node; when the service data determines that the current service of the edge node has an abnormal event related to the building space physical characteristics of the building, invoking the building space physical characteristic cognitive model to generate control commands for the edge node, and sending the control commands to the edge node. Receive the control command; The current service is executed based on the control parameters specified in the control command.

[0011] This invention provides a device control apparatus. The method is applied to a central control terminal located within a building, which is divided into multiple functional spaces by multiple walls. In a distributed smart home network where the central control terminal is located, one or more edge nodes are configured within each of the functional spaces. The apparatus includes: The data calculation module is used to determine the location information and material identification data of the target wall, and to calculate the volume of the functional space; the central control terminal is used to transmit and receive radio signals through the target wall and the edge nodes. An initial spatial framework model construction module is used to construct an initial spatial framework model using the location information, the material identification data, and the volume size data. The building space physical characteristics determination module is used to determine the blocking attenuation parameters of the target wall to the radio signal, as well as the thermal performance parameters and thermal inertia level parameters of the target wall. The building space physical characteristics cognitive model generation module is used to train the initial space frame model based on the barrier attenuation parameter, the thermal performance parameter and the thermal inertia level parameter, and generate the building space physical characteristics cognitive model. The control instruction generation module is used to monitor the service data of the edge node. When it is determined from the service data that an abnormal event related to the physical characteristics of the building space occurs in the current service of the edge node, the module calls the building space physical characteristics cognitive model to generate control instructions for the edge node and sends the control instructions to the edge node. The control instructions are used to control the edge node to execute the current service based on the adjustment parameters of the control instructions.

[0012] This invention provides a device control apparatus. The method is applied to an edge node. In a distributed smart home network where the edge node is located, a corresponding central control terminal is set up. The central control terminal is placed inside a building, which is divided into multiple functional spaces by multiple walls. One or more edge nodes are deployed within each functional space. The central control terminal is used to: determine the location information and material identification data of a target wall, and calculate the volume of the functional space; the central control terminal is used to transmit and receive radio signals through the target wall and the edge nodes; and to construct an initial space using the location information, the material identification data, and the volume data. The device includes: a spatial frame model; determining the blocking attenuation parameters of the target wall for the radio signal, as well as the thermal performance parameters and thermal inertia level parameters of the target wall; training the initial spatial frame model based on the blocking attenuation parameters, the thermal performance parameters, and the thermal inertia level parameters to generate a building space physical characteristic cognitive model; monitoring the service data of the edge node, and when it is determined through the service data that the current service of the edge node has an abnormal event related to the building space physical characteristics of the building, calling the building space physical characteristic cognitive model to generate control commands for the edge node, and sending the control commands to the edge node. The device includes: A control command receiving module is used to receive the control commands; The current service execution module is used to execute the current service based on the control parameters of the control command.

[0013] This invention provides an electronic device including a processor, a memory, and a program or instructions stored in the memory and executable on the processor. When the program or instructions are executed by the processor, they implement the steps of the method described in the first aspect.

[0014] This invention provides a readable storage medium on which a program or instructions are stored, which, when executed by a processor, implement the steps of the method described in the first aspect.

[0015] Fifthly, embodiments of the present invention provide a chip, the chip including a processor and a communication interface, the communication interface being coupled to the processor, the processor being used to run programs or instructions to implement the method as described in the first aspect.

[0016] The embodiments of the present invention have the following advantages: This invention first constructs an initial spatial framework model integrating wall location, material, and volume. Then, it calibrates the wireless signal attenuation parameters, thermal performance parameters, and thermal inertia level parameters of the walls, training a multi-dimensional cognitive model of building space physical characteristics. This overcomes the shortcomings of traditional systems that treat buildings as homogeneous containers and ignore the physical properties of the medium. When abnormal events related to the physical characteristics of the building space occur in the service data of edge nodes, the central control terminal can call this cognitive model to accurately analyze the impact of wall physical resistance on signal or heat conduction. This generates refined control commands matching the current physical environment and sends them to the edge nodes, achieving targeted regulation based on real building physical characteristics. This effectively eliminates service anomalies caused by wall obstruction or thermal characteristics, significantly improving the control accuracy and collaborative stability of distributed smart home networks in complex building environments. Attached Figure Description

[0017] Figure 1 This is a flowchart of the steps of a device control method provided in an embodiment of the present invention; Figure 2 This is a flowchart illustrating a device control method provided in an embodiment of the present invention; Figure 3 This is a structural block diagram of a device control apparatus provided in an embodiment of the present invention; Figure 4 This is a hardware structure block diagram of an electronic device provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of a computer-readable medium provided in an embodiment of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. 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.

[0019] The terms "first," "second," etc., used in the specification and claims of this invention are used to distinguish similar objects and are not used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention can be implemented in orders other than those illustrated or described herein. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0020] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0021] Reference Figure 1 The diagram illustrates a flowchart of a device control method provided in an embodiment of the present invention, which may specifically include the following steps: Step 101: Determine the location information and material identification data of the target wall, and calculate the volume of the functional space; the central control terminal is used to transmit and receive radio signals through the target wall and the edge nodes; Step 102: Construct an initial spatial framework model using the location information, the material identification data, and the volume size data; Step 103: Determine the blocking attenuation parameter of the target wall to the radio signal, as well as the thermal performance parameters and thermal inertia level parameters of the target wall; Step 104: Train the initial spatial frame model based on the barrier attenuation parameter, the thermal performance parameter, and the thermal inertia level parameter to generate a cognitive model of the physical characteristics of the building space. Step 105: Monitor the service data of the edge node. When it is determined from the service data that an abnormal event related to the physical characteristics of the building space occurs in the current service of the edge node, call the building space physical characteristics cognitive model to generate control instructions for the edge node and send the control instructions to the edge node. The control instructions are used to control the edge node to execute the current service based on the adjustment parameters of the control instructions.

[0022] In this embodiment of the invention, the location information and material identification data of the target wall can be determined, and the volume of the functional space can be calculated. The central control terminal is used to transmit and receive radio signals through the target wall and the edge nodes to obtain the physical three-dimensional location, structural material, and three-dimensional spatial volume of the interior walls of the building, providing a basic data source for establishing a spatial model describing the geometric topological characteristics of the building. The target wall refers to the enclosing structure that is distributed inside a building, used to divide different functional spaces, and provides physical resistance to the penetration of radio signals and the conduction of heat. Location information refers to data used to indicate the geometric coordinates and boundary range of a target wall in the three-dimensional space inside a building. Material identification data refers to data that characterizes the compositional properties of the target wall medium (such as reinforced concrete or lightweight partition walls). Functional space refers to a partial area of ​​a building that is divided by walls and has a specific function (such as a living room, bedroom, or study). Volume size data refers to numerical values ​​or parameters that characterize the three-dimensional volume of each functional space. A central control terminal refers to a local smart device deployed within a building, serving as the core node of a distributed smart home network, responsible for data processing, model inference, and issuing control commands. Edge nodes refer to terminal intelligent devices deployed within a functional space that establish wireless communication connections with the central control terminal and perform specific services (such as network data transmission and environmental regulation), such as smart air conditioners, smart refrigerators, smart TVs, tablets, etc. Radio signals refer to electromagnetic wave signals used to transmit communication data and commands between the central control terminal and edge nodes. In this embodiment of the invention, an initial spatial framework model can be constructed using the location information, the material identification data, and the volume size data. This model integrates spatial geometric topology, wall material, and room volume data to create a digital three-dimensional framework that reflects the basic physical structure of a building. The initial spatial framework model refers to a three-dimensional spatial digital representation model that integrates static physical information such as the geometric position of the target wall, the material of the medium, and the volume of the functional space. In this embodiment of the invention, the blocking and attenuation parameters of the target wall to the radio signal, as well as the thermal performance parameters and thermal inertia level parameters of the target wall, can be determined to quantify the physical attenuation degree of the target wall to the penetration of electromagnetic wave signals, and the thermal insulation and temperature hysteresis response capabilities of the wall in the thermal dimension, thereby endowing the spatial model with multi-dimensional physical characteristic parameters. The attenuation parameter refers to the numerical value that characterizes the degree of attenuation of electromagnetic wave energy caused by the physical medium blocking radio signals when they penetrate a target wall. Thermal performance parameters refer to physical parameters that characterize the ability of a target wall to prevent heat penetration and conduction (such as thermal conductivity or heat transfer coefficient). Thermal inertia rating parameters are parameters that characterize the heat absorption and storage properties of a target wall and its resistance to temperature fluctuation delays. In this embodiment of the invention, the initial spatial frame model can be trained based on the barrier attenuation parameter, the thermal performance parameter, and the thermal inertia level parameter to generate a cognitive model of the physical characteristics of the building space. Multidimensional physical characteristic parameters are then injected into the initial geometric frame model for training and updating, generating an intelligent cognitive model that can comprehensively perceive and characterize the building's geometric topology, radio propagation characteristics, and thermal characteristics. The building space physical characteristics cognitive model refers to a physical perception model that integrates spatial geometric framework, electromagnetic barrier characteristics, and thermal performance parameters to support subsequent anomaly diagnosis and command generation. It can be divided into the following three core structural modules: 1. Static spatial geometry topology layer, responsible for providing the digital geometric skeleton of the building.

[0023] Input nodes represent individual functional spaces (such as living room, master bedroom, and study). Each node contains the physical geometric characteristics of that space, such as its volume.

[0024] The connecting edge represents the target wall that divides different functional spaces. Each edge contains the wall's location information and material identification data (such as reinforced concrete, lightweight brick partition wall, etc.).

[0025] 2. The multidimensional physical property representation layer assigns the influence of the physical world on signals, heat, and acoustics / vibration to the "nodes" and "edges" of the above geometric topology framework.

[0026] Electromagnetic propagation attenuation characterizes the attenuation parameter of a target wall to radio signals, used to describe the energy attenuation characteristics of electromagnetic waves when they penetrate the wall.

[0027] Thermal response characterization calibrates the thermal performance parameters (such as heat transfer coefficient) and thermal inertia level parameters (such as heat storage and temperature delay characteristics) of the target wall, which, together with the room volume, characterize the space's heat insulation and heat storage capacity.

[0028] Structural and acoustic characterization maps geometric structural parameters such as the external corner of the load-bearing wall and the inherent vibration spectrum characteristics of the floor structure.

[0029] 3. The neural network mapping and decision reasoning layer typically uses graph neural networks (GNNs) or deep multilayer perceptrons (MLPs) as the main model.

[0030] Feature fusion and embedding encodes the input of "geometric topology and multidimensional physical properties" into a high-dimensional "spatial physical feature vector".

[0031] State and business correlation reasoning can be achieved by inputting real-time business data from edge nodes (such as network transmission data, temperature and humidity / human posture data, and equipment operating frequency data), and then using model reasoning to determine whether business anomalies are caused by physical factors such as wall obstruction, thermal characteristics, or structural resonance.

[0032] The process of outputting control parameters can be based on the model's inference results to calculate and output targeted control parameters for edge nodes (such as channel switching frequency band, sweep angle / temperature setting, frequency avoidance target safe frequency, etc.). In this embodiment of the invention, by monitoring the business data of the edge node, when it is determined from the business data that an abnormal event related to the physical characteristics of the building's space occurs in the current business of the edge node, the building space physical characteristics cognitive model is invoked to generate control instructions for the edge node, and the control instructions are sent to the edge node. The control instructions are used to control the edge node to execute the current business based on the adjustment parameters of the control instructions, so as to monitor the business operation status of the edge node in real time. Once a business abnormality related to the physical characteristics of the building is detected, the cognitive model is used to perform accurate diagnosis and adjustment parameter calculation, and targeted control instructions are issued to restore the normal and efficient operation of the business. Business data refers to the status data generated or collected during the execution of the current business (such as network transmission rate, temperature and humidity sensor data, user body sensation data, or compressor operating parameters). Current business refers to the specific functional service that the edge node is currently performing (such as video data transmission, air conditioning supply, or cooling operation). Abnormal events refer to the service status of edge nodes deviating from the predetermined normal range, and the deviation is caused by factors such as wall obstruction attenuation or thermal physical characteristics, resulting in a fault or adverse state. In the embodiments of the present invention, events that are directly related to the physical characteristics of the building space can be characterized. For example, radio signals are prone to significant electromagnetic energy attenuation during the process of penetrating the physical medium, resulting in an increase in the bit error rate of data packets and a surge in the frequency of retransmissions, which in turn leads to local area network channel blockage. Or, the transportation of hot and cold media with fixed angles and temperatures is prone to collision reflection and airflow dead angles at local walls and corners, resulting in a serious imbalance in the temperature and humidity field distribution in the space. Or, the vibration energy of frequency conversion drive equipment is prone to mechanical wave interference with the building envelope, which in turn leads to low-frequency noise and structural resonance. Of course, the above examples are merely illustrative. Those skilled in the art can also identify other events that are directly related to the physical characteristics of building space as abnormal events. The embodiments of the present invention do not limit this.

[0033] Control parameters refer to specific values ​​or settings used to guide edge nodes in adjusting their operating status. Control commands refer to data packets or commands issued by the central control terminal to the edge nodes to drive the edge nodes to adjust their operating status according to the control parameters. This invention first constructs an initial spatial framework model integrating wall location, material, and volume. Then, it calibrates the wireless signal attenuation parameters, thermal performance parameters, and thermal inertia level parameters of the walls, training a multi-dimensional cognitive model of building space physical characteristics. This overcomes the shortcomings of traditional systems that treat buildings as homogeneous containers and ignore the physical properties of the medium. When abnormal events related to the physical characteristics of the building space occur in the service data of edge nodes, the central control terminal can call this cognitive model to accurately analyze the impact of wall physical resistance on signal or heat conduction. This generates refined control commands matching the current physical environment and sends them to the edge nodes, achieving targeted regulation based on real building physical characteristics. This effectively eliminates service anomalies caused by wall obstruction or thermal characteristics, significantly improving the control accuracy and collaborative stability of distributed smart home networks in complex building environments. For example, in a residence comprising a living room and a study, a central control terminal is placed in the living room, and a laptop computer serving as an edge node is placed in the study. A reinforced concrete target wall separates the living room and study. The central control terminal first determines the location information and material identification data of the target wall, and calculates the volume data of the study and living room to construct an initial spatial framework model. Next, the central control terminal determines the radio signal attenuation parameters of the target wall, as well as the thermal performance parameters and thermal inertia level parameters of the target wall, and uses these parameters to train the initial spatial framework model to generate a cognitive model of the physical characteristics of the building space. When a laptop in the study is conducting a video conference, the central control terminal monitors its data (e.g., a sudden drop in transmission rate). Using a cognitive model of the building's physical characteristics, the central control terminal determines that the sudden drop is caused by an anomaly resulting from radio signals penetrating a high-barrier reinforced concrete wall. Subsequently, the central control terminal uses the cognitive model to calculate channel allocation and control parameters that can overcome the wall's attenuation, generates a channel switching control command for the laptop, and sends it to the laptop. This controls the laptop to switch to a channel with strong anti-interference capabilities and matching penetration power to perform the video conference, thereby eliminating lag and restoring smooth communication.

[0034] Based on the above embodiments, modified embodiments of the above embodiments are proposed. It should be noted that, in order to keep the description brief, only the differences from the above embodiments are described in the modified embodiments.

[0035] In an optional embodiment of the present invention, the service data is the network transmission data of the edge node, and the step of calling the building space physical characteristics cognitive model to generate control instructions for the edge node includes: Determine the transmission rate, retransmission rate, and packet loss rate of the network data; When the building space physical characteristics cognitive model determines the event type of the abnormal event as network congestion caused by signal attenuation due to the radio signal passing through the target wall, based on the transmission rate, retransmission rate and packet loss rate, the building space physical characteristics cognitive model is invoked to retrieve other services of other edge nodes; When the building space physical characteristics cognitive model determines that the other services are high-bandwidth services, the building space physical characteristics cognitive model is invoked to determine the network channel of the other edge nodes. The system invokes the building space physical characteristics cognitive model to find other network channels besides the network channel and generates channel switching instructions for the other network channels; the channel switching instructions are used to control the edge node to switch to the other network channel to execute the current service.

[0036] In this embodiment of the invention, the physical root cause of network congestion is accurately identified by calling the building space physical characteristics cognitive model. Combined with the high bandwidth service distribution and channel occupancy of other edge nodes in the local area network, other isolated network channels are automatically optimized and allocated for the affected edge nodes, thereby eliminating channel contention and congestion and restoring the smoothness and stability of network communication.

[0037] Network transmission data refers to the status monitoring data generated when edge nodes conduct wireless data communication with the central control terminal or network gateway during the execution of current business processes.

[0038] Transmission rate refers to the number of data packets or data bits that an edge node successfully receives or sends per unit of time.

[0039] The retransmission rate refers to the proportion of times an edge node retransmits data packets due to data packet corruption or loss during wireless communication, out of the total number of transmissions.

[0040] Packet loss rate refers to the percentage of data packets that fail to reach the receiver during data transmission out of the total number of data packets sent by the sender.

[0041] Signal attenuation refers to the reduction in electromagnetic wave energy when radio signals pass through physical obstacles such as walls, due to physical effects such as absorption, reflection, and scattering by the medium.

[0042] Network congestion types refer to abnormal event categories caused by severe signal attenuation leading to a large number of data packet errors and retransmissions, which in turn occupy wireless channel bandwidth resources, resulting in service delays or communication interruptions.

[0043] Other edge nodes refer to other smart terminal devices deployed in the same or different functional spaces in the distributed smart home network, excluding the edge node currently experiencing network lag.

[0044] Other services refer to network communication or data transmission tasks that other edge nodes are currently performing.

[0045] High-bandwidth services refer to service types that consume a large amount of network transmission bandwidth resources and have the characteristics of continuous large-volume data transmission (such as 4K high-definition video playback or large file download).

[0046] A network channel refers to a specific frequency band or logical channel used in a wireless communication system to transmit radio signals (such as a specific channel in the 2.4 GHz band or a channel in the 5 GHz band).

[0047] Other network channels refer to backup wireless communication channels that are not currently occupied by high-bandwidth services and have good anti-interference capabilities or independent transmission frequency bands.

[0048] Channel switching commands are control commands generated by the central control terminal and sent to edge nodes to drive the edge nodes to adjust the operating frequency band or channel of their wireless communication modules to a specified other network channel.

[0049] For example, when a laptop in the study is performing a real-time video conferencing service, the central control terminal first obtains the transmission rate, retransmission rate, and packet loss rate of the laptop's network transmission data in real time. When it detects a sudden drop in transmission rate and a significant increase in retransmission rate and packet loss rate, the central control terminal calls the building space physical characteristics cognitive model. Combining this with the high-obstruction attenuation parameters of the target wall in the study recorded in the model, it determines that the lag is caused by network congestion due to the attenuation of radio signals through the wall. Subsequently, the central control terminal calls the cognitive model to retrieve high-bandwidth services such as 4K high-definition video playback being performed by the smart TV in the living room and the network channels they occupy. It further finds other network channels that are not occupied by the high-bandwidth service and have matching penetration and anti-interference capabilities. Based on this, it generates a channel switching command and sends it to the laptop, driving the laptop to seamlessly switch to the other network channel to continue performing the video conferencing service.

[0050] This invention, by combining multi-dimensional network metrics such as transmission rate, retransmission rate, and packet loss rate, utilizes a building space physical characteristics cognitive model to accurately pinpoint the nature of network congestion caused by physical wall obstruction, avoiding the traditional crude control approach of blindly increasing transmission power or ineffective reconnection. Simultaneously, by retrieving channel occupancy information for high-bandwidth services across devices, it achieves frequency band logical isolation and refined traffic splitting based on spatial physical constraints and service characteristics. Without interfering with the normal high-bandwidth services of other devices, it instantly relieves channel contention and significantly reduces packet loss and retransmission rates, ensuring the communication quality of critical real-time services and the overall stability of the network.

[0051] In an optional embodiment of the present invention, the edge node is a smart air conditioner, the service data is temperature sensing data, and the user's posture and micro-motion characteristic data, and the step of calling the building space physical characteristic cognitive model to generate control commands for the edge node includes: When the building space physical characteristics cognitive model determines the event type of the abnormal event as uneven body comfort caused by the building space physical characteristics based on the temperature sensing data and the posture and micro-motion characteristic data, the building space physical characteristics cognitive model is called to query the air conditioner installation coordinates, furniture layout coordinates and load-bearing wall external corner geometric parameters in the functional space corresponding to the smart air conditioner. The building space physical characteristics cognitive model is invoked to generate the air outlet temperature and swing angle for the smart air conditioner based on the air conditioner installation coordinates, the furniture layout coordinates, and the geometric structural parameters of the load-bearing wall corner. The control command for the smart air conditioner is generated using the outlet air temperature and the swing angle; the control command is used to control the smart air conditioner to perform air supply service based on the outlet air temperature and the swing angle.

[0052] In this embodiment of the invention, an abnormal sensation is accurately identified by calling the cognitive model of the physical characteristics of the building space. The spatial airflow is optimized by combining the air conditioner installation coordinates, furniture layout coordinates and geometric parameters of the load-bearing wall corners within the space. Then, control commands for the air outlet temperature and swing angle of the smart air conditioner are generated and issued, thereby realizing the active reconstruction of the micro airflow and temperature field and eliminating differences in the sensation of comfort. Temperature sensing data refers to the real-time ambient temperature values ​​collected by temperature sensors deployed within the functional space. It should be noted that the deployment location of temperature sensors does not have to be unique. Temperature sensing data can be data collected by temperature sensors on edge nodes, data collected by temperature sensors configured in specific locations (such as furniture frequently used by users, such as sofas), or data collected by temperature sensors on user wearable devices. Posture and micro-motion characteristic data refers to the data collected by sensing components on the body posture (such as curling up, crossing arms) and the amplitude of small limb movements of people in the functional space, which are used to characterize the body's perceived temperature. Uneven physical comfort refers to an abnormal event category in which, within the same functional space, there are significant differences in perceived temperature or wind sensation among people in different areas due to airflow obstruction, rebound, or uneven distribution. Intelligent air conditioning refers to environmental control equipment that is deployed as an edge node within a functional space, capable of receiving external control commands and dynamically adjusting the outlet air temperature, wind speed, and swing angle. Air conditioner installation coordinates refer to the specific geometric position parameters of the air outlet of the indoor unit of a smart air conditioner in the three-dimensional coordinate system of the functional space. Furniture layout coordinates refer to the geometric positions and occupied areas of major furniture items such as sofas and dining tables in three-dimensional space within a functional space. The geometric parameters of the external corner of a load-bearing wall refer to the three-dimensional coordinates, corner angles, and surface geometric dimensions of the external corner of a load-bearing wall that protrudes into the functional space. The outlet air temperature refers to the target set temperature calculated by the central control terminal and used to instruct the intelligent air conditioner to output cooling or heating. The swing angle refers to the angle of the swing limit of the air guide plate and the air outlet airflow guiding unit of the intelligent air conditioner in the horizontal and vertical directions or the fixed direction of the sector. Control commands refer to control data packets generated and sent from the central control terminal to the smart air conditioner, which are used to drive the smart air conditioner to perform air delivery tasks according to the specified air outlet temperature and swing angle. Air supply service refers to the process by which a smart air conditioner delivers cool or hot air to a functional space according to the set temperature, wind speed, and airflow direction.

[0053] In practical implementation, the central control terminal can call the building space physical characteristic cognitive model to query the installation coordinates of the smart air conditioner in the living room, the furniture layout coordinates of the sofa and dining table, and the geometric structural parameters of the external corner of the load-bearing wall in the living room, and deduce the physical mechanism of cold air rebounding and converging in the sofa area after hitting the external corner of the load-bearing wall.

[0054] 1. Spatial physical parameter query and 3D entity mapping: 1.1 Extracting spatial geometry and equipment coordinate data: The central control terminal invokes the cognitive model, retrieves the 3D coordinate system definition of the current functional space (living room) from the model's static spatial topology database, and reads the geometric parameters of the following key entities: Smart air conditioner installation coordinates: Obtain the fixed three-dimensional coordinates (x, y, z) of the center of the indoor unit's air outlet. ac , y ac , z acAnd the initial normal vector v of the air outlet. out .

[0055] Geometric parameters of the load-bearing wall's external corner, obtaining the vertex coordinates (x, y) of the protruding corner (external corner) of the living room wall. corner, y corner , z corner ), the angle θ between the two walls on either side of the external corner wall And the normal vector n of the external corner wall wall .

[0056] Target furniture layout coordinates, obtain the bounding box vertex coordinate range of the sofa area and dining table in the living room coordinate system. Specifically target the area where abnormal physical sensations occur—the sofa area's 3D space V. sofa = [x min ,x max ] *[y min , y max ] *[z min , z max ].

[0057] 2. Simulation of the initial phase of an aerodynamic vector jet: Step 2.1: Construct the initial airflow stream for cold air: The cognitive model abstracts the cold air output from the air conditioner vent as a set of three-dimensional airflow lines or airflow particle bundles with velocity vectors and temperature attributes: Let the initial air velocity be v0 and the air temperature be Tcold. Based on the current angle of the air conditioner's air guide plate, a set of three-dimensional airflow direction vectors D = (dx, dy, dz) is emitted into space, starting from the coordinates of the air outlet.

[0058] Step 2.2, Spatial tracking of airflow path: The model calculates the propagation path equation of the airflow jet in three-dimensional space along the airflow direction vector D: P(t) = (x ac , y ac , z ac ) + t·D During the tracking process, the model incorporates aerodynamic drag formulas to update the velocity decay and momentum diffusion of the airflow stream.

[0059] 3. Physical calculation of collision and reflection at the external corner of a load-bearing wall: Step 3.1, Collision Intersection and Incident Angle Calculation: The model calculates the intersection of the airflow propagation path equation with the geometric plane of the load-bearing wall's external corner: Determine whether the airflow beam P(t) strikes the corner wall before reaching the target area; Calculate the coordinates Φ of the intersection point when the airflow beam impacts the external corner wall. t And calculate the incident angle θ based on the normal vector nwall of the positive angle. in .

[0060] Step 3.2, Calculation of airflow rebound path and momentum loss: Based on the principles of fluid mechanics wall collision and reflection, the cognitive model calculates the outgoing direction vector R of the airflow after reflection from a rigid surface (load-bearing wall) at a sharp angle: R = D-2(D·n wall )n wall Simultaneously, the model incorporates the surface roughness coefficient μ from the wall material identification data. wall Calculate the wall friction loss and wind speed momentum attenuation after the airflow collides with the corner, and obtain the effective wind speed v after reflection. reflect =η·vhit, where η is the energy reflection retention coefficient.

[0061] 4. Determining the convergence of spatial airflow reflection trajectory and target area: Step 4.1, Tracking the secondary propagation of reflected airflow: The cognitive model takes the collision intersection point Φt as a new starting point and calculates the spatial trajectory equation of the reflected airflow beam along the reflection direction vector R: P reflect (s)=Φ t +s·R Step 4.2, Calculation of bounding box intersection and wind energy convergence: The model compares the reflected airflow trajectory Preflect(s) with the previously retrieved 3D bounding box V of the sofa area. sofa Perform spatial intersection determination: Collision matching: Determines whether the propagation line segments of most reflected airflow beams penetrate or fall into V. sofa Within the space; Convergence density superposition: Calculates the amount of material falling into the sofa area V per unit time. sofa Internal reflected airflow and cooling density Φ cold .

[0062] 5. Determination of physical mechanism and output of anomaly root cause: Step 5.1, Encapsulation of Mechanism Derivation Conclusions: When the cold air flux Φcold falling into the sofa area Vsofa exceeds the human comfort threshold, the cognitive model completes the closed-loop deduction of the physical mechanism and outputs the deduction conclusion: the cold airflow emitted from the air outlet of the smart air conditioner does not block the furniture and directly collides with the corner of the load-bearing wall; due to the physical reflection effect of the hard wall surface, the original airflow direction is changed, causing the reflected low temperature and high flow velocity airflow to converge again in the sofa area, thus causing people in this area to experience localized uneven comfort due to feeling colder / direct blowing.

[0063] For example, when the temperature and humidity sensor data in the living room are within the normal range, but the Doppler motion sensing component detects that people in the sofa area are curling up or hugging their arms, indicating cold, the central control terminal calls the building space physical characteristics cognitive model to determine that an abnormality in the unevenness of perceived comfort caused by the physical characteristics of the building space has occurred. Subsequently, the central control terminal calls the cognitive model to query the installation coordinates of the smart air conditioner in the living room, the furniture layout coordinates of the sofa and dining table, and the geometric parameters of the external corner of the load-bearing wall in the living room, and deduce the physical mechanism of cold air rebounding and converging in the sofa area after hitting the external corner of the load-bearing wall. Based on this, the cognitive model calculates the sweeping angle that avoids the rebound area of ​​the external corner and the matching outlet air temperature, and uses the outlet air temperature and sweeping angle to generate control commands and send them to the smart air conditioner to drive the smart air conditioner to adjust the direction of the air guide plate and the set temperature to perform the air supply service. This invention, by introducing attitude and micro-motion characteristic data, breaks through the traditional coarse control mode that relies solely on readings from a single temperature and humidity sensor, and achieves precise perception of microscopic differences in spatial sensation. Simultaneously, by deeply integrating the geometric structural parameters of air conditioner installation coordinates, furniture layout, and the external corners of load-bearing walls, it reveals the physical causes of cold air collision and reflection from an aerodynamic perspective. This leads to the generation of sweeping angle and temperature control parameters that can accurately avoid impacting the external corners and optimize airflow paths. Without significantly reducing the set temperature or increasing energy consumption, it eliminates the phenomenon of localized direct cold air blowing and uneven sensation, greatly improving the comfort and energy efficiency of the living environment.

[0064] In an optional embodiment of the present invention, the edge node is a refrigeration device equipped with a compressor, the service data is the current operating frequency of the compressor, and the floor vibration spectrum data of the functional space corresponding to the refrigeration device, and the step of calling the building space physical characteristic cognitive model to generate control commands for the edge node includes: When the building space physical characteristics cognitive model determines the event type of the abnormal event as a structural resonance-type abnormality caused by the vibration characteristics of the compressor and the floor slab of the functional space based on the current operating frequency and the floor vibration spectrum data, the building space physical characteristics cognitive model is invoked to extract the main peak frequency in the floor vibration spectrum data, and a target safe operating frequency that avoids the main peak frequency is determined based on the main peak frequency and the current operating frequency. The target safe operating frequency is used to generate control commands for the refrigeration equipment; the control commands are used to control the refrigeration equipment to adjust the operating frequency of the compressor to the target safe operating frequency.

[0065] This invention aims to address the issue of structural resonance noise caused by the coupling between the operating frequency of a variable frequency compressor in a refrigeration device and the inherent frequency of a building's floor structure. By calling a cognitive model of the physical characteristics of the building space to extract the main peak frequency in the floor vibration spectrum, the invention accurately calculates the target safe operating frequency that avoids the resonance peak and does not affect the refrigeration effect. Then, it generates control commands to drive the compressor to switch to the safe frequency band, thereby achieving the active avoidance and elimination of physical resonance noise. The current operating frequency refers to the frequency of the variable frequency drive electrical signal (in Hertz Hz) that the refrigeration equipment equipped with a compressor outputs or maintains operation in real time at the current moment. Floor vibration spectrum data refers to the data collected by continuously monitoring the floor structure of the functional space where the refrigeration equipment is located through a micromechanical vibration sensing unit, which reflects the distribution of vibration amplitude of the floor at different frequencies. Structural resonance anomalies refer to an abnormal event type in which the operating frequency of the variable frequency compressor closely matches the inherent resonant frequency of building components such as floor slabs or walls, and the amplitude exceeds the preset safety threshold, resulting in amplified vibration and low-frequency noise. The main peak frequency refers to the specific frequency point in the floor vibration spectrum data where the vibration energy is the highest and the corresponding vibration amplitude is the largest. The target safe operating frequency refers to the compressor adjustment frequency calculated by the central control terminal, which maintains a safe frequency difference from the main peak frequency (such as avoiding the resonance peak range) and can maintain the normal cooling effect of the refrigeration equipment. Refrigeration equipment refers to equipment deployed as an edge node within a functional space, equipped with a variable frequency compressor, and capable of dynamically adjusting the compressor speed / operating frequency by receiving control commands (such as variable frequency air conditioner cabinets or wall-mounted units). Control commands refer to control commands generated by the central control terminal and sent to the refrigeration equipment drive controller, used to forcibly change the operating frequency of the compressor. For example, the central control terminal reads the current operating frequency (e.g., 48Hz) of the variable frequency air conditioner compressor in real time through a private communication protocol, and simultaneously uses the built-in micro-mechanical vibration sensor to collect real-time vibration spectrum data of the floor structure of the functional space. When the central control terminal calls the building space physical characteristic cognitive model to perform frequency domain comparison and determines that the compressor operating frequency matches the main peak frequency of the floor vibration and causes structural resonance-like anomalies, the cognitive model automatically extracts the main peak frequency in the vibration spectrum and calculates the target safe operating frequency (e.g., 49.5Hz) that deviates from the main peak frequency range and does not affect the cooling effect. Subsequently, the central control terminal uses the target safe operating frequency to generate a frequency update control command and sends it to the air conditioner drive controller, driving the variable frequency compressor to instantly switch to the target safe operating frequency, thereby eliminating resonance. In practice, the target safe operating frequency can be calculated in the following way.

[0066] The central control terminal first reads the current operating frequency data sent by the variable frequency air conditioner controller in real time through a private communication interface, and simultaneously triggers micro-mechanical vibration sensors deployed on the functional space floor to collect time-domain vibration acceleration signals. Subsequently, the central control terminal performs Fourier transform processing on the collected time-domain vibration signals, converting the time-domain signals into frequency-domain signals, generating floor vibration spectrum data containing the correspondence between frequency and amplitude. The cognitive model inputs the current operating frequency of the variable frequency compressor to the frequency domain comparison module and defines a comparison window centered on the current operating frequency in the floor vibration spectrum data. The cognitive model retrieves the amplitude values ​​within this window and calculates the proportion and magnitude of vibration energy at the current operating frequency. When the vibration amplitude at the current operating frequency exceeds the preset resonance safety threshold, and the amplitude at that frequency reaches a local extremum, the cognitive model determines that the compressor operating frequency and the natural frequency of the floor have physically coupled, confirming a structural resonance anomaly. After confirming the resonance anomaly, the cognitive model runs a peak retrieval algorithm to scan the entire frequency band of the floor vibration spectrum data, accurately locating the frequency value corresponding to the point of maximum vibration amplitude and extracting it as the main peak frequency. Subsequently, the cognitive model calculates the half-power bandwidth corresponding to the main peak frequency based on the damping ratio and quality factor parameters of the floor slab material, thereby defining the resonant danger frequency range, including the main peak frequency. The cognitive model uses the current cooling capacity demand of the refrigeration equipment as a constraint, setting the minimum allowable frequency deflection for the compressor. The model searches for candidate frequencies in the neighboring frequency bands outside the resonant danger frequency range, eliminating low-frequency points that would cause a significant decrease in cooling capacity and high-frequency points that exceed the compressor's safe operating speed. The cognitive model evaluates the predicted amplitude and cooling efficiency corresponding to each candidate frequency, selecting the frequency value closest to the main peak frequency and located outside the resonant danger frequency range, and determines this as the target safe operating frequency.

[0067] This invention, through cross-domain frequency domain comparison of the equipment's operating status with the building structure's vibration spectrum, accurately pinpoints the essence of resonance noise caused by the dynamic coupling between the equipment and the building's physical structure, overcoming the limitation of traditional systems that cannot identify the source of structural resonance. Simultaneously, by using a cognitive model of the physical characteristics of building space to calculate a small-amplitude safety target frequency jump variable, instantaneous frequency avoidance and resonance noise elimination with zero perception are achieved without the need for additional hardware vibration damping accessories or shutdown maintenance. This ensures the continuous and stable operation of the refrigeration equipment and significantly improves the acoustic comfort and living quality of the living environment. In an optional embodiment of the present invention, the step of determining the location information and material identification data of the target wall, and calculating the volume of the functional space includes: Transmit broadband pulsed microwave signals to the edge nodes and receive the reflected echoes of the broadband pulsed microwave signals; Determine the time difference between the transmitted and received signals, the echo intensity, and the spectral characteristics of the reflected echo; Based on the time difference between the transmitted and received signals, the echo intensity, and the spectral characteristics, the location information and material identification data of the target wall are calculated. Collect ambient acoustic signals within the functional space; Determine the reverberation decay time corresponding to the environmental acoustic signal; The volume of the functional space is calculated based on the reverberation decay time.

[0068] This invention aims to utilize the microwave and acoustic multimodal sensing capabilities built into the central control terminal to analyze the geometric position and physical medium material of the wall by actively transmitting and receiving broadband pulse microwave signals. At the same time, it combines the environmental acoustic reverberation attenuation characteristics to calculate the three-dimensional volume of different functional spaces, thereby providing high-precision physical topology and geometric structure foundation data for constructing a refined spatial framework model of a building. Wideband pulse microwave signal refers to an ultra-wideband (UWB) electromagnetic wave signal that is actively transmitted from a central control terminal into the building space, and has a wide frequency range and an extremely short pulse width. A reflected echo refers to an electromagnetic wave signal that is reflected when a broadband pulse microwave signal encounters a physical medium such as a target wall during propagation and is then re-received by the central control terminal. The signal transmission and reception time difference refers to the time interval between the moment when the central control terminal transmits a broadband pulse microwave signal and the moment when it receives the corresponding reflected echo (used to estimate the transmission distance). Echo intensity refers to the amplitude or power of the electromagnetic wave energy of the reflected echo signal received by the central control terminal. Spectral characteristics refer to the energy distribution pattern, attenuation response, and spectral attenuation curve of the reflected echo signal at different frequency components (used to distinguish different media materials such as reinforced concrete and lightweight partition walls). Ambient acoustic signals refer to sound wave signals emitted by sound sources within a functional space or existing naturally, which are captured by the acoustic acquisition array built into the central control terminal after being reflected multiple times by the interior walls and objects of the space. Reverberation decay time refers to the time interval required for the sound energy density in the functional space to decrease by a specified decibel (e.g., 60 dB) during reflection after the sound source stops emitting sound. For example, when the central control terminal first starts or performs the initial spatial scan, it first transmits broadband pulse microwave signals to the edge nodes within the building space through its built-in ultra-wideband microwave module and receives the reflected echoes from each wall. Subsequently, the central control terminal analyzes the transmission and reception time difference, echo intensity, and spectral characteristics of the reflected echoes to accurately calculate the position information of the target wall in three-dimensional space and identify the material identification data of the target wall based on the differences in spectral characteristics. At the same time, the central control terminal uses a high-precision acoustic acquisition array to collect environmental acoustic signals within the functional space, extracts the reverberation decay time of the acoustic signals in the current space, and calculates the volume size data of the corresponding functional space based on the reverberation decay time, thereby completing the accurate perception and database construction of the static physical characteristics of the building space. This invention, through the integration of ultra-wideband microwave active scanning and environmental acoustic reverberation analysis technologies, overcomes the shortcomings of traditional smart home systems that cannot perceive the physical structure of buildings. It enables the automated and adaptive establishment of the physical geometric features of buildings without requiring users to manually draw diagrams or input floor plan parameters. Specifically, by cross-analyzing the time difference, intensity, and spectral three-dimensional features of microwave signals, it can accurately identify the geometric location and deep medium material of hidden walls. Combined with the reverberation decay time, it can accurately calculate the three-dimensional volume of the space, providing an extremely reliable physical foundation for subsequent modeling and refined collaborative control, greatly reducing the deployment threshold and operation and maintenance costs of the system.

[0069] In an optional embodiment of the present invention, the step of determining the blocking attenuation parameter of the target wall to the radio signal, as well as the thermal performance parameters and thermal inertia level parameters of the target wall, includes: Test data packets are sent to the edge nodes of each of the functional spaces at a preset fixed transmission power, and the received signal strength indication data fed back by the edge nodes are obtained. The difference between the received signal strength indication data and the preset theoretical value of free space propagation loss is calculated to obtain the instantaneous blocking attenuation data. Obtain historical barrier attenuation data, and perform a weighted average calculation on the instant barrier attenuation data and the historical barrier attenuation data to generate the barrier attenuation parameter. Acquire outdoor meteorological data and indoor temperature change rate data within the functional space; Based on the outdoor meteorological data and the indoor temperature change rate data, the thermal insulation and heat storage capacity of the functional space is calculated. The thermal performance parameters and thermal inertia level parameters of the target wall are determined based on the heat insulation and heat storage capacity.

[0070] This invention aims to smoothly generate stable barrier attenuation parameters that characterize the electromagnetic barrier performance of walls by dynamically testing the measured received strength of wireless signals and combining it with theoretical loss calculations; at the same time, it calculates the thermal insulation and heat storage capacity of the space by combining outdoor weather and indoor temperature dynamic changes, and accurately calibrates the thermal performance parameters and thermal inertia level parameters of the target wall, thereby endowing the cognitive model of the physical characteristics of building space with accurate physical attributes of radio propagation impedance and thermal response dimensions. The preset fixed transmission power refers to the electromagnetic wave signal output power value that the central control terminal presets and maintains constant when sending test data packets to edge nodes. Test data packets refer to specific format data frames sent by the central control terminal specifically for measuring the quality of wireless communication links and signal transmission loss. The Received Signal Strength Indicator (RSSI) refers to the radio signal power measured by the edge node after receiving the test data packet and fed back to the central control terminal. The theoretical calculation value of free space propagation loss refers to the theoretical energy attenuation value that a wireless signal should have after being transmitted a specified distance in an ideal free space without any obstacles, calculated based on the electromagnetic wave propagation theory formula. Real-time blocking attenuation data refers to the electromagnetic wave energy loss value of a single measurement caused solely by the physical obstruction of the target wall, calculated by subtracting the measured received signal strength indication data from the theoretically calculated free-space propagation loss. Historical barrier attenuation data refers to the set of real-time barrier attenuation data for the same target wall that has been tested multiple times before the current moment and stored locally by the central control terminal. Weighted average calculation refers to a mathematical calculation method that assigns different time weights to historical data and real-time data (such as giving more weight to recent data) and sums them up to achieve an average, in order to eliminate occasional measurement noise. Outdoor meteorological data refers to data on current ambient temperature, solar radiation intensity, and wind speed that affect building heat exchange, obtained through external meteorological interfaces or outdoor sensors. Indoor temperature change rate data refers to the data on the magnitude or rate of change of indoor temperature collected by temperature sensors deployed in functional spaces within a unit time interval. Thermal insulation and heat storage capacity refers to the comprehensive physical ability of a functional space and its enclosure to resist heat loss and store and absorb heat when external meteorological conditions change or indoor heat sources fluctuate. For example, the central control terminal periodically sends test data packets to the edge nodes in the study at a preset fixed transmission power, and receives the received signal strength indication data fed back by the edge nodes. The central control terminal calculates the difference between this data and the preset theoretical value of free space propagation loss to obtain the instantaneous blocking attenuation data, and retrieves historical blocking attenuation data to perform a weighted average calculation to generate blocking attenuation parameters for calibrating the target wall. At the same time, the central control terminal obtains the current outdoor meteorological data and the indoor temperature change rate data collected by the temperature sensor in the study, calculates the heat preservation and heat storage capacity of the study, and accurately determines the thermal performance parameters and thermal inertia level parameters of the target wall (such as calibrating it as a high thermal inertia insulation zone or a low thermal inertia heat leakage zone), and finally completes the assignment and dynamic update of the physical attribute parameters in the cognitive model. This invention, by subtracting natural losses in free space and employing a weighted average smoothing algorithm, effectively eliminates occasional multipath interference and temporary obstruction noise from radio waves in the daily environment, thus calibrating stable and reliable electromagnetic attenuation parameters for the wall. Simultaneously, by combining outdoor weather data with indoor dynamic temperature rise / fall rates, it achieves non-destructive, dynamic self-learning and accurate identification of the wall's thermal performance and thermal inertia level, avoiding the tedious process of manually entering wall thermal parameters. This enables the cognitive model to not only accurately characterize the wall's radio penetration impedance but also precisely predict its heat storage and transfer characteristics, providing solid data support for subsequent differentiated energy storage scheduling and refined environmental management.

[0071] In an optional embodiment of the present invention, the central control terminal is configured with a corresponding cloud server, and after the step of sending the control command to the edge node, the method further includes: The service data is re-collected and updated according to the preset sampling period, and the updated service data is used to generate the performance evaluation result of the edge node executing the current service based on the control command. When the performance evaluation result is determined to meet the preset conditions, control experience sample data is generated based on the business data, the control instructions, and the performance evaluation result. The model data of the cognitive model of physical characteristics of building space is adjusted using the aforementioned regulatory experience sample data to obtain an updated cognitive model of physical characteristics of building space, and the updated cognitive model of physical characteristics of building space is confirmed as the first updated model. Extract the model gradient information generated during the training process of the first updated model; The model gradient information is encrypted using a preset encryption algorithm to generate ciphertext gradient data. The encrypted gradient data is uploaded to the cloud server; the cloud server is used to aggregate the encrypted gradient data from different central control terminals using a federated averaging algorithm, generate global model parameters, and send the global model parameters to the central control terminal. The global model parameters are received, and the first updated model is updated again using the global model parameters to generate the second updated model.

[0072] In this embodiment of the invention, successful control experience is transformed into local positive sample data to adaptively fine-tune the local cognitive model. Under the premise of strictly protecting user privacy, encrypted model gradient information is extracted and uploaded to the cloud to participate in the distributed federated averaging algorithm aggregation. Finally, the local model is updated again using the global model parameters issued by the cloud, thereby realizing the adaptive fine-tuning of the intelligent control capability of a single home host and the continuous iterative evolution of collective intelligence. The preset sampling period refers to the fixed time interval set in advance by the central control terminal after issuing control commands, which is used to collect edge node and environmental status data again. Updating business data refers to the latest business or environmental status data of the edge node that is re-collected by the central control terminal after a preset sampling period following the execution of control commands at the edge node. The performance evaluation result refers to the evaluation conclusion or value obtained by the central control terminal after comparing the updated business data with the abnormal business data or the target preset standard before the regulation, which is used to quantitatively evaluate whether the regulation was successful or not. Preset conditions refer to the evaluation thresholds or standards set in advance to determine whether the control is successful (such as network speed recovering to above the predetermined threshold, the disappearance of the feeling of coldness, or the reduction of the floor resonance amplitude to a safe range). The regulatory experience sample data refers to the training positive sample data that is jointly packaged and combined with the original business data before regulation, the specific control instructions issued, and the evaluation results of successful execution that meet the preset conditions. Model data refers to the weights, biases, or decision parameters of the neural network within the cognitive model of the physical characteristics of architectural space, used to determine the mapping relationship between input and output. The first updated model refers to the phased model with local personalized adaptive capabilities obtained after incrementally fine-tuning the original cognitive model of the physical characteristics of building space using locally generated regulatory experience sample data. Model gradient information refers to the partial derivative matrix data generated during the backpropagation training of the first updated model using the control experience sample data, which is used to indicate the direction and magnitude of model parameter updates. Encryption algorithms refer to mathematical encryption protocols pre-configured by the central control terminal for homomorphic encryption or high-strength, high-density processing of model gradient information to prevent data from being reverse-engineered and cracked. Ciphertext gradient data refers to gradient data packets generated after being processed by encryption algorithms, which completely erase the original user privacy information and are in an unreadable ciphertext state. A cloud server refers to a central computing platform deployed on the network side to receive and process encrypted data from thousands of central control terminals and perform global algorithm aggregation. Federated averaging is a distributed machine learning algorithm that protects data privacy by generating better global model parameters through weighted summation and averaging of gradients of massive amounts of encrypted or abstract models in the cloud. Global model parameters refer to the globally optimized neural network weight parameters calculated by the cloud server after aggregating massive terminal gradient information through a federated averaging algorithm. These parameters have higher generalization ability and stronger inference ability. The second update model refers to the composite cognitive model that combines local personalization and global collective intelligence, which is ultimately generated by the central control terminal after receiving the global model parameters sent from the cloud and overwriting the first update model. For example, after the central control terminal sends control commands to the edge nodes, it re-collects and updates service data (such as network transmission rate after channel switching or temperature, humidity and human posture data after regulation) according to a preset sampling period, and uses the updated service data to generate an execution effect evaluation result. When the execution effect evaluation result is determined to meet the preset conditions (i.e., regulation is successful), the central control terminal encapsulates and generates regulation experience sample data based on the service data before regulation, the issued control commands and the execution effect evaluation result, and uses the sample data to fine-tune the local model data to obtain the first updated model. Then, the central control terminal extracts the model gradient information generated by the first updated model during training, encrypts it using a preset encryption algorithm to generate ciphertext gradient data and uploads it to the cloud server. The cloud server aggregates the ciphertext gradient data from different central control terminals using a federated averaging algorithm to generate global model parameters and distributes them. Finally, the central control terminal receives the global model parameters and uses them to update the first updated model again to generate the final second updated model. In this embodiment of the invention, by establishing a result evaluation and adaptive fine-tuning mechanism after issuing instructions, closed-loop verification of control actions is achieved. This ensures that the system can transform each successful regulation into the self-learning and evolutionary power of the local model, improving the personalized fit of local regulation to specific family environments. At the same time, by introducing federated learning and gradient encryption technology, under the premise of strictly isolating the privacy of family layouts, user routines, and sensitive behaviors, a group collaborative evolution mechanism of "data not leaving home, wisdom shared" is realized. This enables a single central control terminal to absorb the regulation experience of massive heterogeneous residences, significantly improving the generalization ability, decision accuracy, and collaborative regulation effect of the cognitive model in the face of complex and unknown scenarios.

[0073] To enable those skilled in the art to better understand the embodiments of the present invention, an example is used below to illustrate the embodiments of the present invention.

[0074] refer to Figure 2 , Figure 2 This is a flowchart illustrating a device control method provided in an embodiment of the present invention; Phase 1: Hardware Configuration and Initial Deployment of Home PCs 1. Hardware Components: In this embodiment, the home host is physically a simply designed desktop or wall-mounted smart gateway device. In addition to a high-performance multi-core central processing unit and a neural network processing unit dedicated to accelerating artificial intelligence algorithms, it also highly integrates the following key functional modules: Ultra-wideband microwave scanning module: It consists of one or more ultra-wideband radio frequency transceiver chips and a micro array antenna, and is responsible for transmitting and receiving microwave signals for space scanning.

[0075] High-precision ambient acoustic acquisition array: Consists of multiple miniature silicon microphones arranged at specific geometric intervals to capture detailed features of ambient sound.

[0076] Micromechanical vibration sensing unit: A high-precision three-axis or six-axis microelectromechanical system accelerometer and gyroscope used to monitor minute structural vibrations.

[0077] Multi-protocol wireless IoT gateway module: A radio frequency chipset that supports multiple wireless communication protocols such as Zigbee, Thread, Bluetooth Low Energy, and conventional Wi-Fi, ensuring broad compatibility and stable connection with various new and old smart devices in the home.

[0078] 2. Initialization process: After connecting the home host to power and network, users are guided through basic network configuration and user account binding via the accompanying mobile application. The system then prompts the user that "a spatial environment adaptability scan is in progress to optimize the subsequent experience," and automatically enters the autonomous perception and modeling phase of the building space structure. During this phase, users do not need to perform any manual measurements or upload floor plans; the home host will automatically complete the preliminary detection and modeling of the main building envelope and spatial layout using its microwave scanning and acoustic acquisition modules.

[0079] Phase Two: The Daily Operation and Dynamic Evolution of the Cognitive Model of Architectural Spatial Physical Characteristics 1. Data structure and storage of the model: The home computer uses a highly efficient data structure designed specifically for describing the physical world to store a cognitive model of the physical characteristics of architectural spaces. This structure can be understood as a dynamic database composed of "objects" and "attributes." The "objects" in the database mainly include three categories: functional space partitioning units (such as "living room" and "master bedroom"), building envelope entities (such as "the partition wall between the living room and study" and "the south-facing window of the master bedroom"), and equipment terminal entities (such as "one Gree brand air conditioner unit"). Each "object" is accompanied by a set of "attribute" fields that are dynamically updated over time.

[0080] 2. Daily update mechanism for physical property parameters: The core of the system lies in its continuous autonomous learning and parameter correction capabilities. Taking the building envelope entity of "the partition wall between the living room and the study" as an example, the update process of its wireless signal attenuation parameter is as follows: The home console will create a background scheduled task that will proactively initiate communication quality tests with a specific wireless smart speaker in the study multiple times a day when all devices are idle.

[0081] In each test, the home host sends a set of test data packets at a fixed power and records the signal strength indication value received by the smart speaker.

[0082] The system compares each measured signal strength reading with a theoretically calculated value (i.e., the signal strength it would have if it were propagating in a completely open, unobstructed space). The difference between the two is the instantaneous attenuation obtained in this test.

[0083] The system does not directly use the values ​​from a single test. Instead, it calculates a weighted average with historical data to obtain a smooth, stable average barrier attenuation that reflects the long-term characteristics of the wall. If a user hangs a large decorative painting with a metal backing on this partition wall, the system will detect a sudden change in the value in subsequent days of testing and automatically update the model to reflect this new physical reality.

[0084] Phase Three: Implementation Process of Reasoning and Multi-Device Collaborative Control of AI Decision-Making Bodies Taking a complete event-triggered control process as an example, this paper details the specific working steps of the system during operation.

[0085] 1. Event triggering and state snapshot capture: Suppose that one evening, the temperature and humidity sensors in the living room indicate that the ambient temperature and humidity are within the human comfort zone. However, a microwave Doppler motion sensing component installed on the ceiling above the sofa area detects a family member in the sofa area exhibiting a static, micro-motion pattern of curling up with arms crossed—a pattern highly correlated with the human posture of "feeling cold." Simultaneously, a motion sensing component located in the dining area detects another family member performing light housework. This combination of conditions triggers a "discrepancy in perceived comfort event."

[0086] 2. Cognitive Model Query and Problem Simulation: The AI ​​decision-making system within the home control unit was immediately activated. It first queried the building's spatial physics model to obtain complete information about the living room functional area: including the installation location and airflow direction of the air conditioner's indoor unit, the relative coordinates of the sofa and dining table, and the specific location and dimensions of the load-bearing wall corner that might cause airflow rebound. The decision-making system then rapidly performed a simulation within a lightweight, built-in fluid dynamics model. This model was based on fundamental physical rules such as "airflow propagates in straight lines, reflects off obstacles, and diffuses through openings." The simulation results were presented as a visualized airflow path diagram and temperature distribution cloud map, clearly confirming the previous assessment: after the air conditioner's cold air struck the load-bearing wall corner, it formed a converging cold airflow that directly impacted the sofa area.

[0087] 3. Calculation and decision-making of multi-objective optimization schemes: Having identified the physical causes, the decision-making body begins calculating solutions. Its objective function is set as "to alleviate the coldness of the sofa area while minimizing the increase in energy consumption, and to minimize the deterioration of the comfort of the dining area." It generates and evaluates the simulation effects of multiple candidate solutions: Option A: Simply increase the air conditioner's set temperature. Simulation results show that this will cause the entire living room to heat up, making people in the dining area feel even hotter, and cooling down again will require more energy. This option received a low score.

[0088] Option B: Adjust only the angle of the air conditioner vent vanes. Simulation results show that by limiting the vane oscillation range, the amount of air blowing directly onto the corner of the load-bearing wall can be significantly reduced, resulting in a decrease in airflow speed in the sofa area and a rise in perceived temperature. However, it has little effect on the dining area. This option receives a moderate rating.

[0089] Option C (Preferred Option): Combining Option B with a ceiling fan. Simulation results show that Option B solves the main cause of the excessive cold in the sofa area; simultaneously, activating the ceiling fan's ultra-low speed reverse mode gently delivers coolness from near the ceiling to the dining area, resulting in a more uniform overall temperature field. Furthermore, the ceiling fan consumes very little power. This option received the highest score.

[0090] 4. Instruction encapsulation, issuance, and execution confirmation: After selecting option C, the decision-making body converts it into a sequence of executable instructions for the devices. It then sends encapsulated control data packets to the indoor air conditioner unit and the ceiling fan controller via the local network. After the instructions are issued, the home control unit will query the operating status parameters of the air conditioner and ceiling fan again after a few seconds to confirm that the instructions have been successfully received and executed.

[0091] 5. Post-evaluation of the regulatory effect and model fine-tuning: After running stably for about fifteen minutes in the new controlled state, the home control unit will again read the various sensor data from the sofa area and dining area, and analyze the human posture characteristics captured by the motion sensing components. If the posture of the person in the sofa area changes from "curled up with arms crossed" to "relaxed and stretched out," the system determines that the control has achieved the expected positive effect. This successful experience—including the sensor data pattern that triggered the event, the selected control scheme, and the changes in environmental parameters after the control—will be recorded as a positive sample to fine-tune the decision preference model of the local artificial intelligence decision-making agent, enabling it to make judgments and responses faster and more accurately when faced with similar situations in the future.

[0092] Phase Four: Continuous Evolution of Swarm Intelligence Under Privacy Protection Mechanisms The continuous evolution capability of this invention relies on a privacy-preserving federated learning framework, the specific implementation steps of which are as follows: 1. Localized model training and data isolation: All raw data involving users' personal behavioral patterns (such as daily routines and temperature preferences in different regions) and home physical environment (such as specific apartment layouts) are strictly restricted to processing and computation within the local storage space of the home host and are never uploaded to external networks in any raw form. The daily fine-tuning and personalized training of the AI ​​decision-making model are entirely based on this local data.

[0093] 2. Encrypted upload of model gradient information: When the system is idle (e.g., late at night), the home host extracts the update direction and magnitude information (i.e., "gradients") of the model weight parameters generated during recent local training. This gradient information contains absolutely no possibility of being reverse-engineered to the original user data. For example, it records a generalized rule such as "when a building component with high barrier properties is detected, the priority weight of network traffic scheduling is increased by a small amount," rather than "the wall of user Zhang San's study blocks Wi-Fi by 15dB." After extracting this abstract, untraceable gradient information, the home host encrypts it using a high-strength encryption algorithm and then uploads it through a secure channel to the cloud aggregation server of the device manufacturer (such as Gree).

[0094] 3. Cloud-based model aggregation and iteration: The cloud server receives encrypted model gradient information from tens of thousands of households. Instead of decrypting this information, the server uses a federated averaging algorithm to mathematically aggregate the massive amounts of gradient data directly within the encrypted domain. The result is a globally superior AI model with stronger generalization capabilities. For example, by aggregating local learning experiences from tens of thousands of home computers in frigid northern regions and humid southern regions, the cloud model gains a deeper and more comprehensive understanding of the relationship between the thermal performance of building envelopes and environmental control strategies.

[0095] 4. Secure distribution of the updated model: The parameters of the newly generated global AI model, aggregated and generated in the cloud, are also encrypted and distributed back to each home host in batches over the network. After receiving the parameters, the home host decrypts them in a secure local environment and uses them to update its local AI decision-making model.

[0096] Through this mechanism of "data stays within reach, intelligence is shared," the intelligence level of each home host will continuously iterate and evolve as the entire user group uses it, while the privacy of any single family is firmly protected.

[0097] This invention also provides a device control method applied to an edge node. In a distributed smart home network where the edge node is located, a corresponding central control terminal is configured. The central control terminal is placed inside a building, which is divided into multiple functional spaces by multiple walls. One or more edge nodes are deployed within each functional space. The central control terminal is used to: determine the location information and material identification data of a target wall, and calculate the volume of the functional space; transmit and receive radio signals with the edge nodes through the target wall; and construct an initial space using the location information, the material identification data, and the volume data. The method includes: establishing a spatial frame model; determining the attenuation parameters of the target wall for the radio signal, as well as the thermal performance parameters and thermal inertia level parameters of the target wall; training the initial spatial frame model based on the attenuation parameters, the thermal performance parameters, and the thermal inertia level parameters to generate a building space physical characteristic cognitive model; monitoring the service data of the edge node; when the service data determines that the current service of the edge node has an abnormal event related to the building space physical characteristics of the building, invoking the building space physical characteristic cognitive model to generate control commands for the edge node, and sending the control commands to the edge node. Receive the control command; The current service is executed based on the control parameters specified in the control command.

[0098] For the edge node embodiment, since it is basically similar to the central control terminal embodiment, the description is relatively simple. For relevant details, please refer to the description of the central control terminal embodiment.

[0099] It should be noted that, for the sake of simplicity, the method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments of the present invention are not limited to the described order of actions, because according to the embodiments of the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions involved are not necessarily essential to the embodiments of the present invention.

[0100] Reference Figure 3 The diagram illustrates a structural block diagram of a device control apparatus provided in an embodiment of the present invention, which may specifically include the following modules: The data calculation module 301 is used to determine the location information and material identification data of the target wall, and to calculate the volume of the functional space; the central control terminal is used to transmit and receive radio signals through the target wall and the edge nodes. The initial spatial framework model construction module 302 is used to construct an initial spatial framework model using the location information, the material identification data, and the volume size data. The building space physical characteristics determination module 303 is used to determine the blocking attenuation parameter of the target wall to the radio signal, as well as the thermal performance parameters and thermal inertia level parameters of the target wall. The building space physical characteristics cognitive model generation module 304 is used to train the initial space frame model based on the barrier attenuation parameter, the thermal performance parameter and the thermal inertia level parameter, and generate the building space physical characteristics cognitive model. The control instruction generation module 305 is used to monitor the service data of the edge node. When it is determined from the service data that an abnormal event related to the physical characteristics of the building space occurs in the current service of the edge node, the control instruction is generated for the edge node by calling the building space physical characteristics cognitive model and sent to the edge node. The control instruction is used to control the edge node to execute the current service based on the adjustment parameters of the control instruction.

[0101] This invention provides a device control apparatus. The method is applied to an edge node. In a distributed smart home network where the edge node is located, a corresponding central control terminal is set up. The central control terminal is placed inside a building, which is divided into multiple functional spaces by multiple walls. One or more edge nodes are deployed within each functional space. The central control terminal is used to: determine the location information and material identification data of a target wall, and calculate the volume of the functional space; the central control terminal is used to transmit and receive radio signals through the target wall and the edge nodes; and to construct an initial space using the location information, the material identification data, and the volume data. The device includes: a spatial frame model; determining the blocking attenuation parameters of the target wall for the radio signal, as well as the thermal performance parameters and thermal inertia level parameters of the target wall; training the initial spatial frame model based on the blocking attenuation parameters, the thermal performance parameters, and the thermal inertia level parameters to generate a building space physical characteristic cognitive model; monitoring the service data of the edge node, and when it is determined through the service data that the current service of the edge node has an abnormal event related to the building space physical characteristics of the building, calling the building space physical characteristic cognitive model to generate control commands for the edge node, and sending the control commands to the edge node. The device includes: A control command receiving module is used to receive the control commands; The current service execution module is used to execute the current service based on the control parameters of the control command.

[0102] As the device embodiment is basically similar to the method embodiment, the description is relatively simple, and relevant parts can be found in the description of the method embodiment.

[0103] In addition, embodiments of the present invention also provide an electronic device, such as... Figure 4 As shown, it includes a processor 401, a communication interface 402, a memory 403, and a communication bus 404, wherein the processor 401, the communication interface 402, and the memory 403 communicate with each other through the communication bus 404. Memory 403 is used to store computer programs; When the processor 401 executes the program stored in the memory 403, it implements any of the device control methods described in the above embodiments: The communication bus mentioned above can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used to represent it in the diagram, but this does not mean that there is only one bus or one type of bus.

[0104] The communication interface is used for communication between the aforementioned terminal and other devices.

[0105] The memory may include random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0106] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0107] like Figure 5As shown, in another embodiment of the present invention, a computer-readable storage medium 501 is also provided, which stores instructions that, when executed on a computer, cause the computer to perform the device control method described in the above embodiments.

[0108] This invention also provides a chip, which includes a processor and a communication interface. The communication interface is coupled to the processor. The processor is used to run programs or instructions to implement the various processes of the above-described device control method embodiments and achieve the same technical effect. To avoid repetition, it will not be described again here.

[0109] It should be understood that the chip mentioned in the embodiments of the present invention may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0110] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of the present invention is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0111] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0112] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of the present invention.

Claims

1. A device control method, characterized in that, The method is applied to a central control terminal located within a building, which is divided into multiple functional spaces by multiple walls. In the distributed smart home network where the central control terminal is located, one or more edge nodes are configured within each of the functional spaces. The method includes: The location information and material identification data of the target wall are determined, and the volume of the functional space is calculated; the central control terminal is used to transmit and receive radio signals through the target wall and the edge nodes. An initial spatial framework model is constructed using the location information, the material identification data, and the volume size data. Determine the parameters of the target wall's blocking and attenuation of the radio signal, as well as the thermal performance parameters and thermal inertia level parameters of the target wall; The initial spatial frame model is trained based on the barrier attenuation parameter, the thermal performance parameter, and the thermal inertia level parameter to generate a cognitive model of the physical characteristics of the building space. The system monitors the service data of the edge node. When it is determined from the service data that an abnormal event related to the physical characteristics of the building space occurs in the current service of the edge node, the system calls the building space physical characteristics cognitive model to generate control instructions for the edge node and sends the control instructions to the edge node. The control instructions are used to control the edge node to execute the current service based on the adjustment parameters of the control instructions.

2. The method according to claim 1, characterized in that, The business data is the network transmission data of the edge node, and the step of calling the building space physical characteristics cognitive model to generate control instructions for the edge node includes: Determine the transmission rate, retransmission rate, and packet loss rate of the network data; When the building space physical characteristics cognitive model determines the event type of the abnormal event as network congestion caused by signal attenuation due to the radio signal passing through the target wall, based on the transmission rate, retransmission rate and packet loss rate, the building space physical characteristics cognitive model is invoked to retrieve other services of other edge nodes; When the building space physical characteristics cognitive model determines that the other services are high-bandwidth services, the building space physical characteristics cognitive model is invoked to determine the network channel of the other edge nodes. The system invokes the building space physical characteristics cognitive model to find other network channels besides the network channel and generates channel switching instructions for the other network channels; the channel switching instructions are used to control the edge node to switch to the other network channel to execute the current service.

3. The method according to claim 1, characterized in that, The edge node is a smart air conditioner, the business data is temperature sensing data, and the user's posture and micro-motion characteristic data. The step of calling the building space physical characteristic cognitive model to generate control commands for the edge node includes: When the building space physical characteristics cognitive model determines the event type of the abnormal event as uneven body comfort caused by the building space physical characteristics based on the temperature sensing data and the posture and micro-motion characteristic data, the building space physical characteristics cognitive model is called to query the air conditioner installation coordinates, furniture layout coordinates and load-bearing wall external corner geometric parameters in the functional space corresponding to the smart air conditioner. The building space physical characteristics cognitive model is invoked to generate the air outlet temperature and swing angle for the smart air conditioner based on the air conditioner installation coordinates, the furniture layout coordinates, and the geometric structural parameters of the load-bearing wall corner. The control command for the smart air conditioner is generated using the outlet air temperature and the swing angle; the control command is used to control the smart air conditioner to perform air supply service based on the outlet air temperature and the swing angle.

4. The method according to claim 1, characterized in that, The edge node is a refrigeration device equipped with a compressor, the business data is the current operating frequency of the compressor, and the floor vibration spectrum data of the functional space corresponding to the refrigeration device, and the step of calling the building space physical characteristic cognitive model to generate control commands for the edge node includes: When the building space physical characteristics cognitive model determines the event type of the abnormal event as a structural resonance-type abnormality caused by the vibration characteristics of the compressor and the floor slab of the functional space based on the current operating frequency and the floor vibration spectrum data, the building space physical characteristics cognitive model is invoked to extract the main peak frequency in the floor vibration spectrum data, and a target safe operating frequency that avoids the main peak frequency is determined based on the main peak frequency and the current operating frequency. The target safe operating frequency is used to generate control commands for the refrigeration equipment; the control commands are used to control the refrigeration equipment to adjust the operating frequency of the compressor to the target safe operating frequency.

5. The method according to claim 1, characterized in that, The steps of determining the location information and material identification data of the target wall, and calculating the volume of the functional space include: Transmit broadband pulsed microwave signals to the edge nodes and receive the reflected echoes of the broadband pulsed microwave signals; Determine the time difference between the transmitted and received signals, the echo intensity, and the spectral characteristics of the reflected echo; The location information and material identification data of the target wall are calculated based on the time difference between the transmitted and received signals, the echo intensity, and the spectral characteristics. Collect ambient acoustic signals within the functional space; Determine the reverberation decay time corresponding to the environmental acoustic signal; The volume of the functional space is calculated based on the reverberation decay time.

6. The method according to claim 1, characterized in that, The steps of determining the blocking attenuation parameters of the target wall to the radio signal, as well as the thermal performance parameters and thermal inertia level parameters of the target wall, include: Test data packets are sent to the edge nodes of each of the functional spaces at a preset fixed transmission power, and the received signal strength indication data fed back by the edge nodes are obtained. The difference between the received signal strength indication data and the preset theoretical value of free space propagation loss is calculated to obtain the instantaneous blocking attenuation data. Obtain historical barrier attenuation data, and perform a weighted average calculation on the instant barrier attenuation data and the historical barrier attenuation data to generate the barrier attenuation parameter. Acquire outdoor meteorological data and indoor temperature change rate data within the functional space; Based on the outdoor meteorological data and the indoor temperature change rate data, the thermal insulation and heat storage capacity of the functional space is calculated. The thermal performance parameters and thermal inertia level parameters of the target wall are determined based on the heat insulation and heat storage capacity.

7. The method according to claim 1, characterized in that, The central control terminal is configured with a corresponding cloud server, and after the step of sending the control command to the edge node, it further includes: The service data is re-collected and updated according to the preset sampling period, and the updated service data is used to generate the performance evaluation result of the edge node executing the current service based on the control command. When the performance evaluation result is determined to meet the preset conditions, control experience sample data is generated based on the business data, the control instructions, and the performance evaluation result. The model data of the cognitive model of physical characteristics of building space is adjusted using the aforementioned regulatory experience sample data to obtain an updated cognitive model of physical characteristics of building space, and the updated cognitive model of physical characteristics of building space is confirmed as the first updated model. Extract the model gradient information generated during the training process of the first updated model; The model gradient information is encrypted using a preset encryption algorithm to generate ciphertext gradient data. The encrypted gradient data is uploaded to the cloud server; the cloud server is used to aggregate the encrypted gradient data from different central control terminals using a federated averaging algorithm, generate global model parameters, and send the global model parameters to the central control terminal. The global model parameters are received, and the first updated model is updated again using the global model parameters to generate the second updated model.

8. A device control apparatus, characterized in that, The method is applied to a central control terminal, which is located within a building divided into multiple functional spaces by multiple walls. In the distributed smart home network where the central control terminal is located, one or more edge nodes are configured within each of the functional spaces. The device includes: The data calculation module is used to determine the location information and material identification data of the target wall, and to calculate the volume of the functional space; the central control terminal is used to transmit and receive radio signals through the target wall and the edge nodes. An initial spatial framework model construction module is used to construct an initial spatial framework model using the location information, the material identification data, and the volume size data. The building space physical characteristics determination module is used to determine the blocking attenuation parameters of the target wall to the radio signal, as well as the thermal performance parameters and thermal inertia level parameters of the target wall. The building space physical characteristics cognitive model generation module is used to train the initial space frame model based on the barrier attenuation parameter, the thermal performance parameter and the thermal inertia level parameter, and generate the building space physical characteristics cognitive model. The control instruction generation module is used to monitor the service data of the edge node. When it is determined from the service data that an abnormal event related to the physical characteristics of the building space occurs in the current service of the edge node, the module calls the building space physical characteristics cognitive model to generate control instructions for the edge node and sends the control instructions to the edge node. The control instructions are used to control the edge node to execute the current service based on the adjustment parameters of the control instructions.

9. An electronic device, characterized in that, It includes a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein the program or instructions, when executed by the processor, implement the method as described in claims 1-7.

10. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions that, when executed by a processor, implement the method as described in claims 1-7.