Indoor large-space intelligent energy-saving system based on identity recognition and control method thereof

By combining distributed acquisition and control nodes, flexible UWB antennas, and a data training center, the problem of fine-grained control of smart home systems in complex large-space scenarios has been solved, achieving efficient control of temperature, humidity, air pressure, and air quality, and improving the system's recognition and scheduling capabilities.

CN118938693APending Publication Date: 2024-11-12ZHEJIANG PROVINCE INST OF ARCHITECTURAL DESIGN & RES
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410842837.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-06-27
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

Existing smart home control systems struggle to achieve precise time- and zone-based temperature, humidity, air pressure, and air quality control in large spaces and complex scenarios. They also lack sufficient recognition and scheduling capabilities in multi-person scenarios and are deficient in full-scenario hybrid capabilities.

Method used

By employing distributed acquisition nodes, distributed control nodes, flexible UWB antennas, and a data training center, combined with identity recognition technology, and through digital twin modeling and automated training, precise positioning and intelligent control are achieved.

Benefits of technology

It achieves highly reliable, precise, time- and zone-based temperature, humidity, air pressure, and air quality control in large indoor environments, improving system efficiency and user experience, and providing a flexible and scalable solution.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118938693A_ABST
    Figure CN118938693A_ABST
Patent Text Reader

Abstract

The invention discloses an indoor large-space intelligent energy-saving system based on identity recognition and a control method thereof. The intelligent energy-saving system comprises distributed acquisition nodes, distributed control nodes, a flexible UWB antenna and a data training center. The control method comprises the steps of scheme deployment, indoor space UWB positioning, automatic training and instance operation. According to the invention, high-reliability intelligent regulation and control of temperature, humidity, air pressure and air quality in refined time-sharing and zoning under an indoor large scene and a complex mode are realized. And the energy-saving control system with higher efficiency and better experience is realized through intelligent regulation and control of temperature, humidity, air pressure and air quality in a refined time-sharing and zoning manner. And a flexible and extensible solution is provided for an indoor scene with a high security requirement.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of smart home, and in particular relates to an indoor large space intelligent energy-saving system based on identity recognition and a control method thereof. Background Art

[0002] Smart homes detect the indoor environment by installing sensors such as temperature sensors and humidity sensors indoors, and then use control systems to control indoor equipment to adjust the indoor environment accordingly.

[0003] The existing control system has the following defects:

[0004] 1. Most of them are aimed at small indoor scenes. Large scenes with complex indoor structures have higher requirements for sensor resolution accuracy, system architecture support, and algorithm scheduling.

[0005] 2. Most of the time, identification and dispatching are performed on family units with a small number of people. Indoor scenarios with multiple people pose higher challenges to identification and dispatching capabilities.

[0006] 3. Most of them are targeted at a single scenario, have no time-sharing and partitioning capabilities, and lack full-scenario mixing capabilities.

[0007] Therefore, there is a need for an intelligent energy-saving system that can recognize identities and is targeted at large indoor spaces. Summary of the invention

[0008] The purpose of the present invention is to provide an indoor large space intelligent energy-saving system based on identity recognition and a control method thereof, so as to realize high-reliability intelligent regulation of temperature, humidity, pressure and air quality in a refined time-sharing and zone-based manner for large indoor scenes and complex modes.

[0009] In order to achieve the above object, the present invention adopts the following technical solution:

[0010] Indoor large space intelligent energy-saving system based on identity recognition, including: distributed collection nodes, distributed control nodes, flexible UWB antennas and data training centers;

[0011] Distributed collection nodes are used to calculate indoor positioning, collect indoor sound, temperature and humidity data, TVOC data, air pressure data and formaldehyde data, and send the collected data to the distributed control nodes;

[0012] Distributed control nodes are used for data communication, real-time control parameter settings, and indoor positioning calculations;

[0013] Flexible UWB antenna for identification and indoor positioning;

[0014] The data training center is used for data storage, data cleaning, data labeling, inference SEM model training, and push training and updated SEM models.

[0015] Furthermore, the distributed collection nodes include UWB base stations, microphones, temperature and humidity sensors, TVOC sensors, air pressure and altitude sensors, formaldehyde sensors, and RISC-V edge computing MCUs;

[0016] UWB base station, used to transmit UWB signals and communicate with flexible UWB antennas to calculate indoor positioning;

[0017] Microphone, used to collect indoor sounds and distinguish human voice and breathing sounds;

[0018] Temperature and humidity sensor, used to collect temperature and humidity data;

[0019] TVOC sensor, used to collect TVOC data;

[0020] Barometric altitude sensor, used to collect barometric pressure data and calculate altitude;

[0021] Formaldehyde sensor, used to collect formaldehyde data;

[0022] RISC-V edge computing MCU is used to send collected data to distributed control nodes and communicate with nearby nodes for collaborative collection.

[0023] Furthermore, the distributed control nodes include CAN bus control, air conditioning VRF control, fresh air control, humidifier control, local timing system and inference computing nodes;

[0024] CAN bus control, providing CAN device access interface and data communication;

[0025] Air conditioning VRF control, access the air conditioning VRF system in the control area, and perform parameter settings and real-time regulation;

[0026] Fresh air control: access the fresh air system in the control area to set parameters and adjust them in real time;

[0027] Humidifier control: access the humidifier equipment in the control area to set parameters and adjust them in real time;

[0028] The local timing system synchronizes the timing of UWB base stations in the area for indoor positioning calculations;

[0029] Infer computing nodes, obtain regional positioning data, and calculate UWB precise positioning; obtain regional voiceprint data, infer and calculate crowd distribution; obtain regional collection node data and prepare inference parameters; use SEM model to calculate application parameters of each device and adjust the operation of each device in real time.

[0030] The control method of the indoor large space intelligent energy-saving system based on identity recognition comprises the following steps:

[0031] S1: deployment plan;

[0032] S2: UWB positioning in indoor space;

[0033] S3: Automated training;

[0034] S4: Run the example.

[0035] Preferably, the specific method of step S1 is as follows:

[0036] S11: measure basic parameters of large indoor spaces;

[0037] S12: Digital twin modeling is performed based on the basic parameters obtained by measurement, and simulation tests are performed;

[0038] S13: Install each distributed acquisition node and distributed control node based on the node deployment plan generated by the digital twin model;

[0039] S14: deploy a local timing system to correct the time of all nodes;

[0040] S15: Parameter test collection based on multiple scenarios generated by the model;

[0041] S16: Customized automated training pipeline deployment. Through automated training, a more efficient, accurate and better experience model can be generated at any time according to the changes in the scene, and pushed to each distributed control node on a regular basis.

[0042] Preferably, in step S12, the digital twin model generates a UWB base station deployment plan and a customized identification mark for the UWB antenna; and the digital twin model tunes the respiratory sound recognition model.

[0043] Preferably, the specific method of step S2 is as follows:

[0044] S21: UWB base station sends positioning frame;

[0045] S22: UWB antenna receives signal and returns data;

[0046] S23: The UWB base station receives data, calculates relevant information, and submits it to the control node;

[0047] S24: The control node removes invalid data based on the time-continuous digital twin model;

[0048] S25: The control node infers and calculates indoor precise positioning data;

[0049] S26: The control node updates the time-continuous digital twin model based on the latest data.

[0050] Preferably, the specific method of step S3 is as follows:

[0051] S31: The collection data of all distributed collection nodes and the operation data of distributed control nodes are stored in the time series library of the data training center;

[0052] S32: classifying the summary data into scenarios;

[0053] S33: Perform data verification and identify scenarios that need to be optimized;

[0054] S34: further model training is performed in the data training center to generate updated model data;

[0055] S35: Use OTA to update the node firmware and push new model data;

[0056] S36: Regularly use manual reinforcement labeling for the data in the automated training pipeline to ensure that model iterations do not overfit.

[0057] Preferably, the specific method of step S4 is as follows:

[0058] S41: The control node is deployed as an edge computing node to control the model;

[0059] S42: The collection node collects data through integrated sensors to obtain relevant parameters such as temperature and humidity, air pressure and wind direction, air quality, and positioning data and summarizes them in the data training center;

[0060] S43: The data training center identifies the correct scene settings through human scene data analysis;

[0061] S44: each scene parameter is sent to each control node;

[0062] S45: Each control node infers the optimal fitting parameters of the control interval based on the real-time data of the relevant domain acquisition node, and controls the relevant area equipment to execute instructions;

[0063] S46: The control node collects relevant evaluation data within a fixed period and iteratively infers the control parameters of the next stage.

[0064] Preferably, the scenes in step S43 include standard indoor scenes, VIP scenes, and open / glass-transmitting scenes.

[0065] Due to the adoption of the above technical solution, the present invention has the following beneficial effects:

[0066] 1. Realize highly reliable intelligent control of temperature, humidity, pressure and air quality in a refined time-sharing and zone-based manner for large indoor scenes and complex modes.

[0067] 2. Through refined time-sharing and zone-based intelligent regulation of temperature, humidity, air pressure and air quality, an energy-saving control system with higher efficiency and better experience is achieved.

[0068] 3. Provide flexible and scalable solutions for indoor scenarios with high security requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] The present invention will be further described below with reference to the accompanying drawings.

[0070] Figure 1 Schematic diagram of an indoor large space intelligent energy-saving system based on identity recognition.

[0071] Figure 2 A flowchart of the deployment scenario.

[0072] Figure 3 This is a flowchart for automated training.

[0073] Figure 4 The following is a flowchart of the running example. DETAILED DESCRIPTION

[0074] like Figure 1 The indoor large space intelligent energy-saving system based on identity recognition shown includes: distributed collection nodes, distributed control nodes, flexible UWB antennas and data training centers;

[0075] Distributed collection nodes are used to calculate indoor positioning, collect indoor sound, temperature and humidity data, TVOC data, air pressure data and formaldehyde data, and send the collected data to the distributed control nodes;

[0076] Distributed control nodes are used for data communication, real-time control parameter settings, and indoor positioning calculations;

[0077] Flexible UWB antenna for identification and indoor positioning;

[0078] The data training center is used for data storage, data cleaning, data labeling, inference SEM model training, and push training and updated SEM models.

[0079] Distributed collection nodes include UWB base stations, microphones, temperature and humidity sensors, TVOC sensors, air pressure and altitude sensors, formaldehyde sensors, and RISC-V edge computing MCUs;

[0080] UWB base station, used to transmit UWB signals and communicate with flexible UWB antennas to calculate indoor positioning;

[0081] Microphone, used to collect indoor sounds and distinguish human voice and breathing sounds;

[0082] Temperature and humidity sensor, used to collect temperature and humidity data;

[0083] TVOC sensor, used to collect TVOC data;

[0084] Barometric altitude sensor, used to collect barometric pressure data and calculate altitude;

[0085] Formaldehyde sensor, used to collect formaldehyde data;

[0086] RISC-V edge computing MCU is used to send collected data to distributed control nodes and communicate with nearby nodes for collaborative collection.

[0087] Distributed control nodes include CAN bus control, air conditioning VRF control, fresh air control, humidifier control, local timing system and inference computing nodes;

[0088] CAN bus control, providing CAN device access interface and data communication;

[0089] Air conditioning VRF control, access the air conditioning VRF system in the control area, and perform parameter settings and real-time regulation;

[0090] Fresh air control: access the fresh air system in the control area to set parameters and adjust them in real time;

[0091] Humidifier control: access the humidifier equipment in the control area to set parameters and adjust them in real time;

[0092] The local timing system synchronizes the timing of UWB base stations in the area for indoor positioning calculations;

[0093] Infer computing nodes, obtain regional positioning data, and calculate UWB precise positioning; obtain regional voiceprint data, infer and calculate crowd distribution; obtain regional collection node data and prepare inference parameters; use SEM model to calculate application parameters of each device and adjust the operation of each device in real time.

[0094] like Figure 2 The control method of the indoor large space intelligent energy-saving system based on identity recognition shown includes the following steps:

[0095] S1 deployment plan:

[0096] S11: measure basic parameters of large indoor spaces;

[0097] S12: Digital twin modeling is performed based on the basic parameters obtained from the measurements, and simulation tests are performed. The digital twin model generates a UWB base station deployment plan and a customized identification tag for the UWB antenna.

[0098] S13: Install each distributed acquisition node and distributed control node based on the node deployment plan generated by the digital twin model;

[0099] S14: deploy a local timing system to correct the time of all nodes;

[0100] S15: Parameter test collection based on multiple scenarios generated by the model;

[0101] S16: Customized automated training pipeline deployment. Through automated training, a more efficient, accurate and better experience model can be generated at any time according to the changes in the scene, and pushed to each distributed control node on a regular basis.

[0102] S2 Indoor UWB Positioning:

[0103] S21: UWB base station sends positioning frame;

[0104] S22: UWB antenna receives signal and returns data;

[0105] S23: The UWB base station receives data, calculates relevant information, and submits it to the control node;

[0106] S24: The control node removes invalid data based on the time-continuous digital twin model;

[0107] S25: The control node infers and calculates indoor precise positioning data;

[0108] S26: The control node updates the time-continuous digital twin model based on the latest data.

[0109] S3: Figure 3 Automated training shown:

[0110] S31: The collection data of all distributed collection nodes and the operation data of distributed control nodes are stored in the time series library of the data training center;

[0111] S32: classifying the summary data into scenarios;

[0112] S33: Perform data verification and identify scenarios that need to be optimized;

[0113] S34: further model training is performed in the data training center to generate updated model data;

[0114] S35: Use OTA to update the node firmware and push new model data;

[0115] S36: Regularly use manual reinforcement labeling for the data in the automated training pipeline to ensure that model iterations do not overfit.

[0116] S4: Figure 4 The running example shown.

[0117] S41: The control node is deployed as an edge computing node to control the model;

[0118] S42: The collection node collects data through integrated sensors to obtain relevant parameters such as temperature and humidity, air pressure and wind direction, air quality, and positioning data and summarizes them in the data training center;

[0119] S43: The data training center identifies the correct scene settings through human scene data analysis;

[0120] For different scenarios

[0121] i. Standard indoor scene, using the evaluation function:

[0122] AT = 1.07T + 0.2e-0.65 - 2.7

[0123]

[0124] Where AT is the perceived temperature (°C), T is the air temperature (°C), e is the water vapor pressure (hPa), V is the wind speed (m / s), and RH is the relative humidity (%).

[0125] ii. The VIP scenario implements tracking mode through indoor positioning and uses the evaluation function:

[0126] PMV = [0.303e -0 . 036M +0.028]L

[0127] Among them, MV is the predicted mean vote (Predicted Mean Vot e ), is an indicator of human thermal comfort, M is the human metabolic rate (W / m 2 ), L is the solution of the human body heat balance equation, which represents the heat load of the human body (W / m 2 ).

[0128] iii. For open scenes or glass-transmitting scenes, use the evaluation function:

[0129] T s =0.6T a +0.4T r

[0130] Among them, T s is the human body temperature in indoor buildings (℃), T a is the indoor air temperature (°C), T r is the indoor radiation temperature (℃).

[0131] S44: each scene parameter is sent to each control node;

[0132] S45: Each control node infers the optimal fitting parameters of the control interval based on the real-time data of the relevant domain acquisition node, and controls the relevant area equipment to execute instructions;

[0133] S46: The control node collects relevant evaluation data within a fixed period and iteratively infers the control parameters of the next stage.

[0134] The scenes in step S43 include standard indoor scenes, VIP scenes, and open / glass-transmitting scenes.

[0135] The above are only specific embodiments of the present invention, but the technical features of the present invention are not limited thereto. Any simple changes, equivalent replacements or modifications made based on the present invention to solve basically the same technical problems and achieve basically the same technical effects are all included in the protection scope of the present invention.

Claims

1. An indoor large space intelligent energy-saving system based on identity recognition, characterized by include: Distributed acquisition nodes, distributed control nodes, flexible UWB antennas and data training centers; The distributed collection node is used to calculate indoor positioning, collect indoor sound, temperature and humidity data, TVOC data, air pressure data and formaldehyde data, and send the collected data to the distributed control node; The distributed control node is used for data communication, real-time control parameter setting, and indoor positioning calculation; The flexible UWB antenna is used for identity identification and indoor positioning; The data training center is used for data storage, data cleaning, data labeling, inference SEM model training, and push training and updated SEM models.

2. According to claim 1, the indoor large space intelligent energy-saving system based on identity recognition is characterized in that: The distributed collection nodes include a UWB base station, a microphone, a temperature and humidity sensor, a TVOC sensor, an air pressure and altitude sensor, a formaldehyde sensor and a RISC-V edge computing MCU; The UWB base station is used to transmit UWB signals to communicate with the flexible UWB antenna to calculate indoor positioning; The microphone is used to collect indoor sounds and distinguish human voice and breathing sounds; The temperature and humidity sensor is used to collect temperature and humidity data; The TVOC sensor is used to collect TVOC data; The air pressure altitude sensor is used to collect air pressure data and calculate the altitude; The formaldehyde sensor is used to collect formaldehyde data; The RISC-V edge computing MCU is used to send the collected data to the distributed control node and communicate with adjacent nodes for collaborative collection.

3. The indoor large space intelligent energy-saving system based on identity recognition according to claim 2 is characterized in that: The distributed control nodes include CAN bus control, air conditioning VRF control, fresh air control, humidifier control, local timing system and reasoning computing nodes; The CAN bus control provides a CAN device access interface and performs data communication; The air conditioning VRF control is connected to the air conditioning VRF system in the control area to perform parameter setting and real-time regulation; The fresh air control is connected to the fresh air system in the control area to set parameters and adjust them in real time; The humidifier control accesses the humidifier equipment in the control area to perform parameter setting and real-time regulation; the local timing system synchronizes the timing of the UWB base stations in the area for indoor positioning calculation; The inference computing node obtains regional positioning data and calculates UWB precise positioning; obtains regional voiceprint data and infers and calculates crowd distribution; obtains regional collection node data and prepares inference parameters; uses the SEM model to calculate the application parameters of each device and adjusts the operation of each device in real time.

4. A control method for an indoor large space intelligent energy-saving system based on identity recognition as claimed in claim 1, characterized in that: S1: deployment plan; S2: UWB positioning in indoor space; S3: Automated training; S4: Run the example.

5. The control method of the indoor large space intelligent energy-saving system based on identity recognition according to claim 4 is characterized in that: The specific method of step S1 is as follows: S11: measure basic parameters of large indoor spaces; S12: Digital twin modeling is performed based on the basic parameters obtained by measurement, and simulation tests are performed; S13: Install each distributed acquisition node and distributed control node based on the node deployment plan generated by the digital twin model; S14: deploy a local timing system to correct the time of all nodes; S15: Parameter test collection based on multiple scenarios generated by the model; S16: Customized automated training pipeline deployment. Through automated training, a more efficient, accurate and better experience model can be generated at any time according to the changes in the scene, and pushed to each distributed control node on a regular basis.

6. According to the control method of the indoor large space intelligent energy-saving system based on identity recognition according to claim 5, Features: In step S12, the digital twin model generates a UWB base station deployment plan and a customized identification tag for the UWB antenna.

7. The control method of the indoor large space intelligent energy-saving system based on identity recognition according to claim 4 is characterized in that: The specific method of step S2 is as follows: S21: UWB base station sends a positioning frame; S22: UWB antenna receives signal and returns data; S23: The UWB base station receives data, calculates relevant information, and submits it to the control node; S24: The control node removes invalid data based on the time-continuous digital twin model; S25: The control node infers and calculates indoor precise positioning data; S26: The control node updates the time-continuous digital twin model based on the latest data.

8. The control method of the indoor large space intelligent energy-saving system based on identity recognition according to claim 4 is characterized in that: The specific method of step S3 is as follows: S31: The collection data of all distributed collection nodes and the operation data of distributed control nodes are stored in the time series library of the data training center; S32: classifying the summary data into scenarios; S33: Perform data verification and identify scenarios that need to be optimized; S34: further model training is performed in the data training center to generate updated model data; S35: Use OTA to update the node firmware and push new model data; S36: Regularly use manual reinforcement labeling for the data in the automated training pipeline to ensure that model iterations do not overfit.

9. The control method of the indoor large space intelligent energy-saving system based on identity recognition according to claim 4 is characterized in that: The specific method of step S4 is as follows: S41: The control node is deployed as an edge computing node to control the model; S42: The collection node collects data through integrated sensors to obtain relevant parameters such as temperature and humidity, air pressure and wind direction, air quality, and positioning data and summarizes them in the data training center; S43: The data training center identifies the correct scene settings through human scene data analysis; S44: each scene parameter is sent to each control node; S45: Each control node infers the optimal fitting parameters of the control interval based on the real-time data of the relevant domain acquisition node, and controls the relevant area equipment to execute instructions; S46: The control node collects relevant evaluation data within a fixed period and iteratively infers the control parameters of the next stage.

10. The control method of the indoor large space intelligent energy-saving system based on identity recognition according to claim 4, characterized in that: The scenes in step S43 include standard indoor scenes, VIP scenes, and open / glass-transmitting scenes.