Intelligent robot fire early warning detection device
Through the intelligent robot fire warning detection device, integrated multimodal sensors and large model analysis, the monitoring blind spots and data security issues of the traditional fire warning system are solved, and efficient fire warning and decision support are achieved.
Patent Information
- Application Number
- CN202510956878.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-11
- Publication Date
- 2025-09-30
AI Technical Summary
Traditional fire warning systems have difficulty responding in a timely manner in the incipient stage of a fire, are unable to accurately locate the source of the fire, and have monitoring blind spots in large and complex venues. They are unable to upload data to the cloud in real time for in-depth analysis, resulting in the inability to quickly obtain comprehensive and accurate on-site information when a fire occurs.
An intelligent robot fire warning detection device is used, integrating mobile mechanisms, environmental perception modules, data processing modules and Internet platforms. It utilizes composite high-precision smoke sensors, infrared flame sensors, high-precision MEMS temperature sensors and image acquisition units, combined with the YOLOv7 model and the Transformer large model to achieve multimodal data fusion analysis and emergency decision-making, and ensure data security through quantum key distribution and alliance chain architecture.
It has achieved accurate prediction of fire spread direction and impact range in fire scenarios, generated highly accurate emergency decision-making plans, ensured data security and privacy, and reduced fire losses.
Smart Images
Figure CN120726751A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of fire early warning, and in particular to an intelligent robot fire early warning detection device. Background Art
[0002] Fire poses a serious threat to human life and property, but traditional early warning systems suffer from numerous drawbacks. Common smoke alarms and heat detectors rely on smoke spread or a significant temperature rise to trigger an alarm, making them incapable of responding promptly to a fire's incipient stages. Furthermore, these devices cannot accurately locate the fire source. In large and complex locations such as shopping malls, warehouses, and multi-story factories, limited monitoring range and signal transmission delays can easily create blind spots, making it difficult to detect a fire in its early stages.
[0003] With the development of technology, although some fire detection equipment with mobile functions has made certain progress, their intelligence level is low and they lack effective integration with the Internet platform. They are unable to upload the collected data to the cloud in real time for in-depth analysis. As a result, when a fire occurs, relevant personnel cannot quickly obtain comprehensive and accurate on-site information, miss the best time to deal with it, and cause greater losses. Therefore, an intelligent robot fire early warning detection device is proposed. Summary of the Invention
[0004] To this end, the present invention provides an intelligent robot fire warning detection device to solve the above-mentioned problems in the prior art.
[0005] In order to achieve the above object, the present invention provides the following technical solutions: According to a first aspect of the present invention, an intelligent robot fire warning detection device includes a moving mechanism, the moving mechanism includes a base plate, the bottom of the base plate is fixedly connected to a mounting piece, one side of the mounting piece is fixedly connected to a first motor, and one end of the first motor shaft is fixedly connected to a driving wheel; a control mechanism, the control mechanism includes a control terminal and an antenna, the bottom of the control terminal is fixedly connected to the bottom of the base plate, the surface of the control terminal is provided with a mainboard, the antenna is provided on the surface of the mainboard, the top of the base plate is fixedly connected to a mounting bracket, and the control terminal is provided with a data processing module, a communication module and an Internet platform; an environmental perception module, the environmental perception module includes a composite high-precision smoke sensor, an infrared flame sensor, a high-precision MEMS temperature sensor and an image acquisition unit.
[0006] Furthermore, the image acquisition unit includes a second motor, the surface of the second motor is fixedly connected to the surface of the mounting bracket, one end of the second motor's rotating shaft is fixedly connected to the mounting bracket, the surface of the mounting bracket is fixedly connected to a third motor, one end of the third motor's rotating shaft is fixedly connected to an image acquisition component, and the image acquisition component includes a 4K high-definition camera and a non-cooled thermal imaging camera.
[0007] Furthermore, the data processing module processes the information collected by the environmental perception module, and internally includes a CPU / GPU / NPU multi-core heterogeneous processor and a YOLOv7 model. The CPU / GPU / NPU multi-core heterogeneous processor adopts an edge computing architecture; the YOLOv7 model is combined with transfer learning technology to realize multimodal data fusion analysis.
[0008] Furthermore, the communication module is equipped with dual-mode 5G / 4G and dual-band WiFi, supports TCP / IP and UDP protocols, adopts JPEG-XR / H.265 compression algorithm and AES-256 encryption, and optimizes the protocol stack to achieve high-speed and low-latency data transmission, with a transmission rate of 1Gbps / delay <100ms.
[0009] Furthermore, the Internet platform is based on the Hadoop architecture and integrates the Transformer big model, the authority management unit, the fire equipment management unit, and the VR fire training module. The Transformer big model builds a knowledge graph based on fire regulations and cases, and combines meteorological and geographical data to predict the spread of fire and generate emergency decision-making plans that include evacuation routes and equipment linkage.
[0010] Furthermore, the authority management unit uses quantum key distribution and one-time pad encryption to transmit data, stores data based on a consortium chain architecture, and implements dynamic identity authentication and authority management through a zero-trust architecture. At the same time, it adopts a federated learning framework, and the intelligent robots in each region only upload encrypted model parameters. The global model is aggregated and updated through a central server, and differential privacy technology is combined to protect data privacy.
[0011] Furthermore, the fire-fighting equipment management unit constructs a virtual model of the fire-fighting equipment through digital twin technology, collects equipment operation data in real time and synchronizes the status, and automatically controls the fire-fighting equipment according to the fire situation.
[0012] Furthermore, the composite high-precision smoke sensor includes a multi-sensitivity detection unit and a spectral analysis function. The composite high-precision smoke sensor automatically switches the detection mode through the multi-sensitivity detection unit. The infrared flame sensor adopts three-band infrared detection technology, and the temperature sensor combines MEMS technology with a high-precision A / D conversion circuit to achieve a temperature measurement accuracy of ±0.1°C.
[0013] Furthermore, the Internet platform simulates multiple fire scenarios through VR technology, provides fire training courses with gesture and voice interaction, sets graded emergency drill tasks and evaluates user operations in real time.
[0014] Furthermore, the data processing module adopts feature-level fusion technology to extract flame morphology video, temperature distribution image, smoke concentration, and text keyword features to construct a multi-dimensional fire feature vector.
[0015] The present invention has the following advantages: through the setting of an internet platform, based on the Transformer large model, a fire protection business-specific knowledge base is formed after self-supervised learning training, which can perform in-depth fusion analysis of multimodal data. In complex fire scenarios, the model can simultaneously analyze data such as flame morphology captured by high-definition cameras, temperature distribution captured by thermal imaging cameras, and gas concentration changes collected by sensors. Combined with knowledge of fire regulations, historical cases, and other knowledge, it can accurately predict the direction of fire spread and the scope of impact, and automatically generate emergency decision-making plans including evacuation route planning and fire equipment linkage strategies. Compared with traditional systems that can only issue simple warnings, this device has a higher decision accuracy rate, which buys golden time for fire rescue. Data transmission adopts the quantum key distribution protocol, which utilizes the non-cloning property of quantum states to achieve unconditional secure encryption, eliminating the risk of data theft or tampering. The storage link is based on a blockchain system with a consortium chain architecture, which ensures data integrity and traceability through timestamps and smart contracts. The zero-trust security architecture dynamically authenticates and manages permissions for all access requests, ensuring the security and privacy of fire protection data even in industrial scenarios with complex network environments. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] Figure 1 This is a schematic diagram of the main structure of an intelligent robot fire early warning detection device provided by the present invention.
[0017] Figure 2 This is a side structural schematic diagram of an intelligent robot fire early warning detection device provided by the present invention.
[0018] Figure 3 This is a schematic diagram of the control mechanism structure of an intelligent robot fire early warning detection device provided by the present invention.
[0019] Figure 4 This is a schematic diagram of the bottom-up structure of an intelligent robot fire early warning detection device provided by the present invention.
[0020] Figure 5 This is a structural schematic diagram of the image acquisition unit of an intelligent robot fire warning detection device provided by the present invention.
[0021] Figure 6 This is a structural block diagram of the control system of an intelligent robot fire early warning detection device provided by the present invention.
[0022] In the figure: 11, base plate; 12, mounting part; 13, first motor; 14, driving wheel; 21, control terminal; 22, main board; 23, antenna; 24, mounting bracket; 31, infrared flame sensor; 32, high-precision MEMS temperature sensor; 33, composite high-precision smoke sensor; 34, image acquisition unit; 341, second motor; 342, mounting bracket; 343, third motor; 344, image acquisition component. DETAILED DESCRIPTION
[0023] The following describes the implementation of the present invention using specific embodiments. Those skilled in the art will readily understand the other advantages and benefits of the present invention from the disclosure herein. Obviously, the embodiments described are only a portion of the present invention, not all of it. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without inventive effort are intended to fall within the scope of protection of the present invention. Example
[0024] like Figures 1 to 6 As shown, an intelligent robot fire warning detection device in an embodiment of the first aspect of the present invention includes a moving mechanism, which includes a base plate 11, a mounting member 12 is fixedly connected to the bottom of the base plate 11, a first motor 13 is fixedly connected to one side of the mounting member 12, and a driving wheel 14 is fixedly connected to one end of the rotating shaft of the first motor 13; a control mechanism, which includes a control terminal 21 and an antenna 23, the bottom of the control terminal 21 is fixedly connected to the bottom of the base plate 11, a mainboard 22 is provided on the surface of the mainboard 22, the antenna 23 is provided on the surface of the mainboard 22, and a mounting bracket 24 is fixedly connected to the top of the base plate 11, and a data processing module, a communication module and an Internet platform are provided in the control terminal 21; an environmental perception module, which includes a composite high-precision smoke sensor 33, an infrared flame sensor 31, a high-precision MEMS temperature sensor 32 and an image acquisition unit 34; In the above embodiment, it should be noted that after the intelligent robot is started, the mobile mechanism operates first, the base plate 11 serves as the basic support, and the mounting member 12 thereon fixes the first motor 13, and the first motor 13 drives the driving wheel 14 to rotate, so that the robot moves stably in the preset area, creating dynamic monitoring conditions for the environmental perception module, and the environmental perception module immediately enters the working state. The composite high-precision smoke sensor 33 uses a multi-sensitivity detection unit and a spectral analysis function to monitor the environmental smoke concentration in real time. When the smoke concentration is detected to reach the threshold of 0.01mg / m³, the multi-sensitivity detection unit automatically switches the mode, accurately detecting the smoke concentration. Distinguish between fire smoke and interference sources; the infrared flame sensor 31, leveraging three-band infrared detection technology, can capture characteristic flame spectra from 100 meters away. The high-precision MEMS temperature sensor 32, based on MEMS technology and high-precision A / D conversion circuitry, continuously monitors temperature changes with a high accuracy of ±0.1°C and a 10Hz sampling frequency. Simultaneously, the second motor 341 and third motor 343 of the image acquisition unit 34 coordinate to adjust the angle of the image acquisition component 344. A 4K high-definition camera and an uncooled thermal imaging camera simultaneously capture visible light and thermal infrared images, synthesizing a multi-dimensional image of the fire scene using a data fusion algorithm. The technical effect achieved by the above embodiment is as follows: the second motor 341 and the third motor 343 of the image acquisition unit 34 cooperate to adjust the angle of the image acquisition component 344, the 4K high-definition camera and the uncooled thermal imaging camera synchronously capture visible light and thermal infrared images, and synthesize a multi-dimensional picture of the fire scene through a data fusion algorithm. Example
[0025] like Figures 1 to 6 As shown, an intelligent robot fire warning detection device includes all the contents of Example 1. In addition, the image acquisition unit includes a second motor 341, the surface of the second motor 341 is fixedly connected to the surface of the mounting bracket 24, one end of the second motor 341 shaft is fixedly connected to the mounting bracket 342, the surface of the mounting bracket 342 is fixedly connected to the third motor 343, one end of the third motor 343 shaft is fixedly connected to the image acquisition component 344, the image acquisition component includes a 4K high-definition camera and a non-cooled thermal imaging camera, the data processing module processes the information collected by the environmental perception module The system processes data in real time and includes a CPU / GPU / NPU multi-core heterogeneous processor and a YOLOv7 model. The CPU / GPU / NPU multi-core heterogeneous processor adopts an edge computing architecture. The YOLOv7 model combines transfer learning technology to achieve multimodal data fusion analysis. The communication module is equipped with dual-mode 5G / 4G and dual-band WiFi, supports TCP / IP and UDP protocols, uses JPEG-XR / H.265 compression algorithms and AES-256 encryption, and optimizes the protocol stack to achieve high-speed and low-latency data transmission, with a transmission rate of 1Gbps and a latency of <100ms. In the above embodiment, it should be noted that the collected data is processed and transmitted by the communication module of the control mechanism, dual-mode 5G / 4G and dual-band WiFi ensure network connection, JPEG-XR / H.265 compression algorithm compresses images and videos by 10-50 times, AES-256 encryption ensures data security, and the optimized protocol stack achieves a 1Gbps transmission rate and a delay of less than 100ms, quickly uploading the data to the Internet platform, and the data processing module performs preliminary analysis of the data. The CPU / GPU / NPU multi-core heterogeneous processor is based on the edge computing architecture and completes data noise reduction, format conversion and other pre-processing locally to reduce the pressure on the cloud; the YOLOv7 model combines transfer learning technology and adopts feature-level fusion to extract multi-dimensional features such as flame morphology, temperature distribution, and smoke concentration, construct a fire feature vector, identify fire hazards and assess the risk level, and transmit the analysis results to the Internet platform; The technical effect achieved by the above embodiment is: using feature-level fusion method to extract multi-dimensional features such as flame shape, temperature distribution, smoke concentration, etc., construct fire feature vectors, identify fire hazards and assess risk levels, and transmit the analysis results to the Internet platform. Example
[0026] like Figures 1 to 6As shown, an intelligent robot fire warning detection device includes all the contents of Example 2. In addition, the Internet platform is based on the Hadoop architecture and integrates the Transformer large model, the authority management unit, the fire equipment management unit, and the VR fire training module. The Transformer large model builds a knowledge graph based on fire regulations and cases, combines meteorological and geographical data to predict the spread of fire, and generates an emergency decision-making plan including evacuation routes and equipment linkage. The authority management unit uses quantum key distribution and one-time pad encryption to transmit data, stores data based on the alliance chain architecture, and realizes dynamic identity authentication and authority management through the zero-trust architecture. At the same time, a federated learning framework is adopted. The intelligent robots in each region only upload encrypted model parameters, and the global model is updated through aggregation by the central server. The differential privacy technology is combined to protect data privacy. The fire equipment management unit can realize dynamic identity authentication and authority management through the zero-trust architecture. The management unit uses digital twin technology to build a virtual model of firefighting equipment, collects equipment operation data in real time and synchronizes status, and automatically controls firefighting equipment according to the fire situation. The composite high-precision smoke sensor 33 includes a multi-sensitivity detection unit and spectral analysis function. The composite high-precision smoke sensor 33 automatically switches detection modes through the multi-sensitivity detection unit. The infrared flame sensor 31 uses three-band infrared detection technology. The temperature sensor combines MEMS technology with high-precision A / D conversion circuit to achieve a temperature measurement accuracy of ±0.1°C. The Internet platform uses VR technology to simulate multi-scenario fires, provide fire training courses with gesture and voice interaction, set up hierarchical emergency drill tasks and evaluate user operations in real time. The data processing module uses feature-level fusion technology to extract flame morphology videos, temperature distribution images, smoke concentration, and text keyword features to construct a multi-dimensional fire feature vector. In the above embodiment, it should be noted that after the Internet platform receives the data, the Transformer large model predicts the fire spread trend and generates an emergency decision-making plan based on the fire knowledge graph and combined with meteorological and geographical data. The authority management unit uses quantum key distribution and one-time pad encryption to ensure the security of data transmission, stores data based on the alliance chain architecture, and realizes dynamic authority control through the zero-trust architecture; the federated learning framework coordinates the update of models in each region under the premise of protecting privacy, and the fire equipment management unit uses digital twin technology to synchronize the status of fire equipment in real time, and automatically links fire hydrants, sprinkler systems and other equipment according to the decision plan. In addition, the VR fire training module can simulate fire scenes and provide interactive training. Finally, the Internet platform transmits the instructions back to the control terminal through the communication module, controls the robot to adjust the detection strategy or coordinate with external fire resources to complete the fire warning and disposal closed loop.
[0027] The technical effects achieved by the above embodiments are: data is stored based on the alliance chain architecture, and dynamic permission control is achieved through the zero-trust architecture; the federated learning framework coordinates the update of regional models while protecting privacy, and the fire equipment management unit uses digital twin technology to synchronize the status of fire equipment in real time, and automatically links fire hydrants, sprinkler systems and other equipment according to the decision-making plan.
[0028] Working principle: After the intelligent robot is started, the mobile mechanism operates first. The base plate 11 serves as the basic support. The mounting part 12 thereon fixes the first motor 13. The first motor 13 drives the driving wheel 14 to rotate, so that the robot can move stably within the preset area, creating dynamic monitoring conditions for the environmental perception module. The environmental perception module then enters the working state. The composite high-precision smoke sensor 33 uses a multi-sensitivity detection unit and a spectral analysis function to monitor the ambient smoke concentration in real time. When the smoke concentration reaches the threshold of 0.01mg / m³, the multi-sensitivity detection unit automatically switches the mode to accurately distinguish between fire smoke and interference sources. The infrared flame sensor 31, with its three-band infrared detection technology, can capture the characteristic light of flames from 100 meters away. spectrum; the high-precision MEMS temperature sensor 32, based on MEMS technology and high-precision A / D conversion circuit, continuously monitors temperature changes with a high accuracy of ±0.1°C and a sampling frequency of 10Hz. At the same time, the second motor 341 and the third motor 343 of the image acquisition unit 34 work together to adjust the angle of the image acquisition component 344. The 4K high-definition camera and the uncooled thermal imaging camera synchronously collect visible light and thermal infrared images, and synthesize a multi-dimensional picture of the fire scene through the data fusion algorithm. The collected data is processed and transmitted by the communication module of the control mechanism. Dual-mode 5G / 4G and dual-band WiFi ensure network connection. The JPEG-XR / H.265 compression algorithm compresses the image and video by 10-50 times, and AES-256 encryption is used for encryption. To ensure data security, the optimized protocol stack achieves a 1Gbps transmission rate and a delay of less than 100ms, quickly uploading data to the Internet platform. The data processing module performs preliminary analysis on the data. The CPU / GPU / NPU multi-core heterogeneous processor is based on the edge computing architecture to complete data noise reduction, format conversion and other pre-processing locally to reduce the pressure on the cloud. The YOLOv7 model combines transfer learning technology and adopts feature-level fusion to extract multi-dimensional features such as flame shape, temperature distribution, and smoke concentration, construct fire feature vectors, identify fire hazards and assess risk levels, and transmit the analysis results to the Internet platform. After the Internet platform receives the data, the Transformer large model is based on the fire knowledge graph and combined with Meteorological and geographical data are used to predict fire spread trends and generate emergency decision-making plans. The authority management unit uses quantum key distribution and one-time pad encryption to ensure data transmission security, stores data based on the alliance chain architecture, and realizes dynamic authority control through the zero-trust architecture; the federated learning framework coordinates the update of regional models while protecting privacy. The fire equipment management unit uses digital twin technology to synchronize the status of fire equipment in real time, and automatically links fire hydrants, sprinkler systems and other equipment according to the decision plan. In addition, the VR fire training module can simulate fire scenes and provide interactive training. Finally, the Internet platform transmits the instructions back to the control terminal through the communication module, controls the robot to adjust the detection strategy or coordinate with external fire resources to complete the fire warning and disposal closed loop.
Claims
1. An intelligent robot fire warning detection device, characterized in that: include A moving mechanism, the moving mechanism comprising a base plate (11), a mounting member (12) fixedly connected to the bottom of the base plate (11), a first motor (13) fixedly connected to one side of the mounting member (12), and a driving wheel (14) fixedly connected to one end of a rotating shaft of the first motor (13); A control mechanism, the control mechanism comprising a control terminal (21) and an antenna (23), the bottom of the control terminal (21) being fixedly connected to the bottom of the base plate (11), a mainboard (22) being provided on the surface of the control terminal (21), the antenna (23) being provided on the surface of the mainboard (22), a mounting frame (24) being fixedly connected to the top of the base plate (11), and a data processing module, a communication module and an Internet platform being provided in the control terminal (21); An environmental perception module comprises a composite high-precision smoke sensor (33), an infrared flame sensor (31), a high-precision MEMS temperature sensor (32), and an image acquisition unit (34).
2. The intelligent robot fire warning detection device according to claim 1, characterized in that: The image acquisition unit comprises a second motor (341), a surface of the second motor (341) is fixedly connected to a surface of a mounting frame (24), one end of a rotating shaft of the second motor (341) is fixedly connected to a mounting frame (342), a surface of the mounting frame (342) is fixedly connected to a third motor (343), and one end of a rotating shaft of the third motor (343) is fixedly connected to an image acquisition component (344), the image acquisition component comprising a 4K high-definition camera and a non-cooling thermal imaging camera.
3. The intelligent robot fire warning detection device according to claim 1, characterized in that: The data processing module processes the information collected by the environmental perception module and internally includes a CPU / GPU / NPU multi-core heterogeneous processor and a YOLOv7 model. The CPU / GPU / NPU multi-core heterogeneous processor adopts an edge computing architecture; the YOLOv7 model is combined with transfer learning technology to realize multimodal data fusion analysis.
4. The intelligent robot fire warning detection device according to claim 1, characterized in that: The communication module is equipped with dual-mode 5G / 4G and dual-band WiFi, supports TCP / IP and UDP protocols, adopts JPEG-XR / H.265 compression algorithm and AES-256 encryption, and optimizes the protocol stack to achieve high-speed and low-latency data transmission, with a transmission rate of 1Gbps / latency <100ms.
5. The intelligent robot fire warning detection device according to claim 1, characterized in that: The internet platform is based on the Hadoop architecture and integrates the Transformer large model, a rights management unit, a fire equipment management unit, and a VR fire training module. The Transformer large model constructs a knowledge graph based on fire regulations and cases, combines meteorological and geographical data to predict fire spread trends, and generates emergency decision-making plans that include evacuation routes and equipment linkage.
6. The intelligent robot fire warning detection device according to claim 5, characterized in that: The authority management unit uses quantum key distribution and one-time pad encryption to transmit data, stores data based on a consortium chain architecture, and implements dynamic identity authentication and authority management through a zero-trust architecture. At the same time, it adopts a federated learning framework, and the intelligent robots in each region only upload encrypted model parameters. The global model is updated through aggregation through a central server, and differential privacy technology is combined to protect data privacy.
7. The intelligent robot fire warning detection device according to claim 5, characterized in that: The fire-fighting equipment management unit uses digital twin technology to build a virtual model of fire-fighting equipment, collects equipment operation data in real time and synchronizes status, and automatically controls fire-fighting equipment according to the fire situation.
8. The intelligent robot fire warning detection device according to claim 1, characterized in that: The composite high-precision smoke sensor (33) includes a multi-sensitivity detection unit and a spectrum analysis function. The composite high-precision smoke sensor (33) automatically switches the detection mode through the multi-sensitivity detection unit. The infrared flame sensor (31) adopts a three-band infrared detection technology. The temperature sensor combines MEMS technology with a high-precision A / D conversion circuit to achieve a temperature measurement accuracy of ±0.1°C.
9. The intelligent robot fire warning detection device according to claim 1, characterized in that: The internet platform uses VR technology to simulate multi-scenario fires, provides firefighting training courses with gesture and voice interaction, sets graded emergency drill tasks, and evaluates user operations in real time.
10. The intelligent robot fire warning detection device according to claim 1, characterized in that: The data processing module adopts feature-level fusion technology to extract flame morphology video, temperature distribution image, smoke concentration, and text keyword features to construct a multi-dimensional fire feature vector.