Intelligent control method and system of coal lighting intelligent lamp based on cloud collaboration

Abstract

The application discloses an intelligent control method and system of a coal lighting intelligent lamp based on cloud cooperation, and relates to the technical field of cloud cooperation. According to the coal lighting intelligent lamp and a plurality of passage lighting lamps configured for a coal mine passage, a corresponding working illumination area is determined, virtual simulation is performed in combination with underground environment information collected by the coal lighting intelligent lamp and a plurality of electronic fence areas, and a corresponding area dynamic diagram is output. If the coal lighting intelligent lamp outputs an emergency signal, a cloud cooperation mechanism between the cloud and the coal lighting intelligent lamp is triggered, a plurality of sub-areas of different risk levels are output, and early warning linkage is performed in combination with an emergency light end of the coal lighting intelligent lamp to output light bands of different colors to the coal mine passage. An intelligent control system of the coal lighting intelligent lamp is constructed based on a plurality of risk characteristics, illumination positions of the light bands and a working state of the coal lighting intelligent lamp, and the accuracy of the intelligent control system of the coal lighting intelligent lamp is improved.

Description

Technical Field

[0001] This invention relates to the field of cloud collaboration, and in particular to an intelligent control method and system for a cloud-based intelligent coal lighting lamp. Background Technology

[0002] Coal mining operations are conducted in harsh environments with complex underground tunnels and various hidden hazards such as gas, dust, and roof water damage. As a primary means for miners to perceive their environment, the level of intelligence in lighting systems directly impacts the safety and efficiency of coal mine production. Traditional miner lamps or fixed passageway lighting typically provide only basic illumination and cannot dynamically generate targeted electronic fence zones based on the miner's real-time location, movement trajectory, and geological structures within the tunnel (such as faults and old goaf areas). In complex tunnel environments, miners often struggle to intuitively perceive their relative spatial relationship to hazards (such as gas accumulation areas and geologically fractured zones), relying solely on experience or static markers for disaster avoidance. The dynamic area maps fail to reflect the current safety situation in real time, affecting their accuracy and resulting in low precision in the intelligent control system of coal mine lighting. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of the prior art. This invention provides an intelligent control method and system for coal-fired lighting smart lamps based on cloud collaboration.

[0004] This invention provides an intelligent control method for smart coal-fired lighting lamps based on cloud collaboration, comprising:

[0005] Miners wear smart coal lighting lamps and walk dynamically along the coal mine passages. The smart coal lighting lamps are connected to the corresponding cloud based on the Internet of Things. In the cloud, multiple electronic fence areas are determined based on the location data of the smart coal lighting lamps and the distribution map of the coal mine passages.

[0006] The corresponding working lighting area is determined based on the intelligent coal lighting lamp and the multiple channel lighting lamps configured in the coal mine channel. The underground environmental information collected by the intelligent coal lighting lamp and multiple electronic fence areas are combined to perform virtual simulation and output the corresponding dynamic map of the area.

[0007] If the coal mine lighting smart lamp outputs an emergency signal, it triggers a cloud-based collaborative mechanism between the cloud and the coal mine lighting smart lamp. During the collaborative process, the dynamic map of the area is dynamically identified to output multiple sub-areas with different risk levels. The emergency light terminal of the coal mine lighting smart lamp is used for early warning linkage to output warning light bands of different colors to the coal mine passage. Each warning light band is dynamically adjusted as the position of the coal mine lighting smart lamp is adjusted.

[0008] Based on the analysis of the emergency signal, multiple emergency contents are determined, and linked verification is performed in conjunction with the current images collected by the coal lighting smart lamp to output multiple risk features. The multiple risk features, the illumination positions of each warning light strip, and the working status of the coal lighting smart lamp are cross-matched, and an attention mechanism is incorporated into the matching process, thereby gradually building an intelligent control system for the coal lighting smart lamp.

[0009] This invention provides an intelligent control system for a cloud-based collaborative coal-fired power plant lighting smart lamp. This cloud-based collaborative intelligent control system is applied to the aforementioned cloud-based collaborative intelligent control method for coal-fired power plant lighting smart lamps. The cloud-based collaborative intelligent control system for coal-fired power plant lighting smart lamps includes:

[0010] The electronic fence module is used by miners to wear smart coal lighting lamps and move dynamically along the coal mine passage. The smart coal lighting lamps are connected to the corresponding cloud based on the Internet of Things. In the cloud, multiple electronic fence areas are determined based on the location data of the smart coal lighting lamps and the distribution map of the coal mine passage.

[0011] The virtual simulation module is used to determine the corresponding working lighting area based on the intelligent coal lighting lamp and the multiple channel lighting lamps configured in the coal mine channel, and to perform virtual simulation by combining the underground environmental information collected by the intelligent coal lighting lamp and multiple electronic fence areas, and output the corresponding dynamic map of the area.

[0012] The cloud collaboration module is used to trigger a cloud collaboration mechanism between the cloud and the smart coal lighting lamp when the smart coal lighting lamp outputs an emergency signal. During the collaboration process, the module dynamically identifies the regional dynamic map to output multiple sub-regions with different risk levels. It also combines the emergency light terminal of the smart coal lighting lamp to perform early warning linkage to output warning light bands of different colors to the coal mine passage. Each warning light band is dynamically adjusted as the position of the smart coal lighting lamp is adjusted.

[0013] The intelligent control module is used to determine multiple emergency contents based on the analysis of the emergency signal, and to perform linkage verification in conjunction with the current image collected by the coal lighting intelligent lamp to output multiple risk features. It cross-matches multiple risk features, the illumination position of each warning light strip and the working status of the coal lighting intelligent lamp, and incorporates an attention mechanism in the matching process, thereby gradually building an intelligent control system for the coal lighting intelligent lamp.

[0014] Compared with the prior art, the beneficial effects of the present invention are:

[0015] (1) Miners wear smart coal lighting lamps and walk dynamically along the coal mine passage. The smart coal lighting lamps are connected to the corresponding cloud based on the Internet of Things. In the cloud, multiple electronic fence areas are determined based on the location data of the smart coal lighting lamps and the distribution map of the coal mine passage. The corresponding working lighting area is determined based on the smart coal lighting lamps and the multiple channel lighting lamps configured in the coal mine passage. The underground environmental information collected by the smart coal lighting lamps and the multiple electronic fence areas are combined to perform virtual simulation and output the corresponding dynamic map of the area. The introduction of multiple electronic fence areas further controls the working lighting area and improves the accuracy of the dynamic map of the area.

[0016] (2) If the coal lighting smart lamp outputs an emergency signal, it triggers the cloud collaboration mechanism between the cloud and the coal lighting smart lamp. During the collaboration process, the regional dynamic map is dynamically identified to output multiple sub-regions with different risk levels. The emergency light terminal of the coal lighting smart lamp is used for early warning linkage to output warning light strips of different colors to the coal mine passage. Each warning light strip is dynamically adjusted as the position of the coal lighting smart lamp is adjusted. Multiple emergency contents are determined according to the analysis of the emergency signal. The current image collected by the coal lighting smart lamp is used for linkage verification to output multiple risk features. The multiple risk features, the illumination position of each warning light strip and the working status of the coal lighting smart lamp are cross-matched. During the matching process, the attention mechanism is combined to gradually build the intelligent control system of the coal lighting smart lamp. The cloud collaboration between the cloud and the coal lighting smart lamp is further controlled. The multiple risk features, the illumination position of each warning light strip and the working status of the coal lighting smart lamp are fully considered, which improves the accuracy of the intelligent control system of the coal lighting smart lamp. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the intelligent control method for a cloud-based collaborative coal-fired lighting smart lamp according to an embodiment of the present invention.

[0018] Figure 2 This is a flowchart illustrating step S11 in the intelligent control method for a cloud-based collaborative coal lighting smart lamp according to an embodiment of the present invention.

[0019] Figure 3 This is a flowchart illustrating step S12 in the intelligent control method for a cloud-based collaborative coal lighting smart lamp according to an embodiment of the present invention.

[0020] Figure 4 This is a flowchart illustrating step S13 in the intelligent control method for a cloud-based collaborative coal lighting smart lamp according to an embodiment of the present invention.

[0021] Figure 5This is a flowchart illustrating step S14 in the intelligent control method for a cloud-based collaborative coal lighting smart lamp according to an embodiment of the present invention.

[0022] Figure 6 This is a schematic diagram of the structure of the intelligent control system for a cloud-based collaborative coal lighting smart lamp in an embodiment of the present invention. Detailed Implementation

[0023] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0024] Please see Figures 1 to 6 A cloud-based collaborative intelligent control method for coal-fired lighting smart lamps is proposed and applied to cloud-based collaborative scenarios. The cloud-based collaborative intelligent control method for coal-fired lighting smart lamps includes:

[0025] Step S11: Miners wear smart coal lighting lamps and walk dynamically along the coal mine passage. The smart coal lighting lamps are connected to the corresponding cloud based on the Internet of Things. In the cloud, multiple electronic fence areas are determined based on the location data of the smart coal lighting lamps and the distribution map of the coal mine passage.

[0026] Step S12: Determine the corresponding working lighting area based on the intelligent coal lighting lamp and the multiple channel lighting lamps configured in the coal mine channel, and perform virtual simulation by combining the underground environmental information collected by the intelligent coal lighting lamp and multiple electronic fence areas, and output the corresponding dynamic map of the area.

[0027] Step S13: If the coal mine lighting smart lamp outputs an emergency signal, the cloud collaboration mechanism between the cloud and the coal mine lighting smart lamp is triggered. During the collaboration process, the regional dynamic map is dynamically identified to output multiple sub-regions with different risk levels. The emergency light terminal of the coal mine lighting smart lamp is used for early warning linkage to output warning light strips of different colors to the coal mine passage. Each warning light strip is dynamically adjusted as the position of the coal mine lighting smart lamp is adjusted.

[0028] Step S14: Based on the analysis of the emergency signal, multiple emergency contents are determined, and the current images collected by the coal lighting smart lamp are linked for verification to output multiple risk features. The multiple risk features, the illumination positions of each warning light strip, and the working status of the coal lighting smart lamp are cross-matched, and an attention mechanism is incorporated during the matching process to gradually build an intelligent control system for the coal lighting smart lamp.

[0029] refer to Figure 2 In step S11, the specific steps are as follows:

[0030] S111: The coal mine lighting smart lamp is worn on the miner's head and provides dynamic lighting to the coal mine tunnel as the miner moves. When the miner walks in the coal mine tunnel, the current position of the miner is displayed according to the position data of the coal mine lighting smart lamp. At the same time, the position data of the coal mine lighting smart lamp is collected and synchronized to the cloud at high frequency in combination with the Internet of Things of the coal mine tunnel.

[0031] S112: Obtain the distribution map of the coal mine tunnels, which includes the tunnel topology, geological faults, and hidden disaster-causing factors. At the same time, after receiving the data, the cloud performs spatial mapping between the location data of the coal mine lighting smart lamps and the distribution map of the coal mine tunnels. Through a dynamic geometric calculation mechanism, based on the miner's current location and combined with preset safety location thresholds, it adaptively divides multiple electronic fence areas, including absolute no-entry zones, over-limit warning zones, and regular passage zones.

[0032] In the embodiments of this application, the smart coal lighting lamp is worn on the miner's head and provides dynamic lighting to the coal mine tunnel as the miner moves. The miner walks in the coal mine tunnel, and the current position of the miner is displayed according to the position data of the smart coal lighting lamp. At the same time, the position data of the smart coal lighting lamp is collected and, combined with the Internet of Things of the coal mine tunnel, the position data is synchronized to the cloud at high frequency, thus introducing high-frequency synchronization of position data to the cloud.

[0033] At this time, the coal mine lighting smart lamp, as an integrated intelligent terminal on the miner's head, is fixed to the mining safety helmet through a rigid connection mechanism. As the miner walks along the coal mine tunnel, the coal mine lighting smart lamp communicates with the external positioning base station based on its built-in inertial navigation system, and calculates its own three-dimensional spatial coordinates in the tunnel coordinate system in real time. This position data is not only used for positioning, but also serves as an input parameter for dynamic lighting control. The coal mine lighting smart lamp dynamically adjusts the light distribution curve of the main light source according to the walking speed and attitude angle to achieve beam-following lighting and ensure the uniformity of illuminance within the miner's field of vision. At the same time, the system renders the miner's current position marker in the local cache in real time, as a reference source for subsequent data synchronization.

[0034] The data acquisition unit inside the smart coal lighting lamp captures raw location information at millisecond intervals. This information typically includes UWB ultra-wideband ranging data, IMU inertial measurement unit acceleration and angular velocity data, and RFID tag information. The system performs timestamp alignment and filtering on the above-mentioned multi-source heterogeneous data to remove positioning noise caused by underground multipath effects and generate high-confidence location data packets. This process is completed by the edge computing microprocessor built into the smart coal lighting lamp, ensuring the accuracy and integrity of the data before transmission.

[0035] Through an IoT wireless sensor network (such as a 5G private network or an industrial Wi-Fi Mesh network) deployed within the coal mine tunnels, the smart coal lighting lamps concurrently upload encapsulated location data packets to the cloud server using a preset high-frequency transmission protocol (such as MQTT or CoAP). This transmission link adopts a QoS service quality assurance mechanism to ensure that the location data stream can still maintain real-time synchronization with low latency and low packet loss rate, even in the complex underground electromagnetic environment and when the signal is attenuated due to the curvature of the tunnel. After receiving the data stream, the cloud immediately performs data parsing and verification to complete the mapping update from the physical world location to the digital twin spatial coordinates.

[0036] Specifically, in a coal mine tunnel, which is defined as a deep transport roadway with complex geological structure, branches, and non-uniform signal coverage; the intelligent coal lighting lamps worn by miners have high-precision positioning modules and adaptive dimming functions.

[0037] When miners wearing smart coal mine lighting enter the coal mine tunnels for inspection work, as the miners move at a speed of approximately 1.5 meters per second, the smart coal mine lighting uses its built-in UWB positioning module to interact with positioning base stations deployed on the tunnel walls in real time to calculate the miners' current absolute coordinates as (X: 150.5m, Y: 30.2m, Z: -450.0m). At the same time, based on the changes in the pitch angle of the miners' heads, the smart coal mine lighting automatically adjusts the beam projection angle by 15 degrees to illuminate any low-lying obstacles in front of the coal mine tunnel, achieving dynamic lighting tracking.

[0038] The microprocessor of the smart coal mine lighting lamp encapsulates data packets containing coordinates, attitude, and device status at a frequency of 20Hz. These data packets are then synchronized to the cloud control center via a 5G industrial IoT base station deployed within the coal mine tunnel. Even if signal fluctuations occur at bends in the coal mine tunnel, the breakpoint resume mechanism enabled by the smart coal mine lighting lamp ensures the continuity of the location data chain. This allows the cloud-based digital twin platform to display the miner's precise movement trajectory and current location status within the coal mine tunnel in real time without delay, providing real-time data support for subsequent electronic fence determination.

[0039] Furthermore, a distribution map of the coal mine tunnels is obtained, which includes the tunnel topology, geological faults, and hidden disaster-causing factors. Simultaneously, after receiving the data, the cloud platform spatially maps the location data of the smart coal mine lighting lamps with the distribution map of the coal mine tunnels. Through a dynamic geometric calculation mechanism, based on the miner's current location and combined with preset safety location thresholds, multiple electronic fence areas are adaptively divided. These multiple electronic fence areas include absolutely prohibited areas, over-limit warning areas, and regular passage areas.

[0040] At this point, the cloud control center retrieves the coal mine tunnel distribution map of the target area from the geological database and the mine geographic information system. This distribution map is not a traditional two-dimensional planar drawing, but a multi-source fusion vector map that integrates the three-dimensional topological structure of the tunnels, the location of fault fracture zones revealed by geological exploration, and hidden disaster-causing factors (such as old goaf areas and waterlogged areas) that have been identified through geophysical exploration in the past. The system vectorizes and encodes the above-mentioned geological risk elements and assigns them specific spatial attribute labels in the digital twin space, constructing a digital base map with risk weights to provide a decision-making basis for subsequent spatial mapping.

[0041] When the cloud receives the high-frequency location data stream from step S111, it starts the spatial mapping engine. This engine aligns and registers the real-time 3D coordinates uploaded by the coal mine lighting smart lamps with the loaded coal mine tunnel distribution map. Through the point-area inclusion relationship, the system determines the specific location of the coordinates in the vector map and retrieves the geological attribute data within the preset range around the location in real time. This process realizes the transformation from simple "location coordinates" to contextual information with "environmental attributes", establishing the relative spatial relationship between miners and geological risk sources.

[0042] Based on the miner's real-time location, a dynamic geometric calculation mechanism is triggered in the cloud to construct a buffer geometry centered on the miner that dynamically changes as the miner moves. The system logically divides the buffer area according to a preset safe location threshold (this threshold is not a fixed value, but the radius is dynamically adjusted according to the surrounding geological risk level). Based on the risk attributes in the geological distribution map, different sectors or sub-regions are judged: if a certain area is less than the safe threshold from a geological fault or hidden disaster source, it is marked as a high-risk area; otherwise, areas with stable geological structures and no obstacles are marked as safe areas. Three types of electronic fence areas are output through Boolean operations: the "absolutely prohibited area" which is strictly prohibited from entering, the "over-limit warning area" which is close to the risk edge, and the "normal passage area" which allows normal operation.

[0043] Specifically, when the cloud executes step S112, a detailed distribution map of the coal mine tunnel is loaded. The map clearly marks the existence of a known F3 geological fault 300 meters ahead of the tunnel excavation, and there is a risk of water accumulation in old goaf areas near the fault (a hidden disaster-causing factor).

[0044] Miners wearing smart coal mine lighting lamps are inspecting the main roadway of the coal mine, with their location coordinates received in real time by the cloud. During spatial mapping, the system identifies that the miner is currently 150 meters away from the F3 fault. Immediately, a dynamic geometric calculation mechanism is activated, establishing a dynamic safety buffer zone centered on the miner's current coordinates and based on the "safe position threshold." Due to the presence of the high-risk source, the F3 fault, the system shrinks the safety buffer radius, designating a fan-shaped area within 20 meters of the fault as an absolute no-entry zone (marked in red), corresponding to the geological fault fracture zone. The transition area extending 30 meters outward from the fault edge is designated as a warning zone (marked in yellow), indicating that the miner is approaching a danger source. The area behind the miner and with intact side support is designated as a regular passage zone (marked in green). This classification is dynamic and continuous; with each slight movement of the miner within the coal mine roadway, the cloud recalculates the relative distance between the miner and the F3 fault, updating the boundaries of the electronic fence in real time.

[0045] refer to Figure 3 In step S12, the specific steps are as follows:

[0046] S121: Real-time monitoring of the cloud. The cloud uses the real-time pose of the coal mine lighting smart lamp as a spatial anchor point, and coordinates the use of the illumination tilt angle and luminous flux parameters of multiple channel lights that have been networked in the coal mine channel. The real-time pose of the coal mine lighting smart lamp and the illumination tilt angle and luminous flux parameters of multiple channel lights are input into the same virtual space. In this virtual space, a working lighting area that is calculated in real time as the miner moves is constructed in combination with the light field reconstruction mechanism.

[0047] S122: The intelligent coal lighting lamp has multiple environmental sensors built in. During the lighting process, the intelligent coal lighting lamp collects corresponding underground environmental information through multiple environmental sensors, including gas concentration, temperature and humidity and dust particles.

[0048] S123: The cloud integrates multi-source heterogeneous data from multiple underground environmental information sources and multiple electronic fence areas, and uses a physics engine to build a full-element virtual simulation environment in the cloud. Ray tracing and real-time rendering technology are applied to the virtual simulation environment to further output the corresponding regional dynamic map. This regional dynamic map is stored in the coal-fired intelligent lighting lamp and can intuitively reflect the current environmental safety situation, light blind spots, and fence intrusion risks.

[0049] In the embodiments of this application, real-time monitoring is performed on the cloud. The cloud uses the real-time pose of the coal mine lighting smart lamp as a spatial anchor point, and coordinates the illumination tilt angle and luminous flux parameters of multiple channel lights that have been networked in the coal mine channel. The real-time pose of the coal mine lighting smart lamp and the illumination tilt angle and luminous flux parameters of multiple channel lights are input into the same virtual space. In this virtual space, a working lighting area that is calculated in real time as the miner moves is constructed in conjunction with a light field reconstruction mechanism. This approach is compatible with the overall consideration of virtual space combined with light field reconstruction mechanism, ensuring the accuracy of the working lighting area calculated in real time as the miner moves.

[0050] At this time, the cloud monitoring module maintains a high-frequency polling state of the underground Internet of Things, and sets the received real-time pose data of coal mine lighting smart lights (including three-dimensional spatial coordinates and Euler angle attitude information) as the spatial anchor point of the current scene. With this anchor point as the center, the cloud uses a topology network retrieval method to search for networked channel lights in the coal mine channel that are within the communication coverage or logical association range of the anchor point. The system sends query commands to the control nodes of these target lights and transmits back the current mechanical illumination tilt angle (pitch angle and yaw angle) and the luminous flux parameters of the electric drive of each channel light in real time, thereby constructing a real-time correlation data cluster of "people-lights-environment".

[0051] A high-fidelity digital twin virtual space is constructed in the cloud, mapping the real-time pose of the coal-fired intelligent lighting lamps as dynamic entity objects to the vicinity of the origin of the virtual coordinate system. At the same time, the tilt angle and luminous flux parameters of multiple channel lighting lamps called in collaboration are transformed into the spatial pose attributes and luminous intensity attributes of the virtual light source. This process realizes the accurate mapping of physical entity parameters to virtual model attributes, ensuring that the lighting scene in the virtual space maintains geometric and optical consistency with the real physical environment underground, and providing accurate boundary conditions for subsequent light field calculations.

[0052] In the virtual space, the system initiates a light field reconstruction mechanism. Based on photometric principles and a roadway wall reflection model, it simulates the physical process of light emanating from a light source and propagating, reflecting, and attenuating within the roadway. It calculates in real time the superposition state of the main optical axis of the coal lighting smart lamp and the light spots of multiple channel lighting lamps in space, eliminates invalid lighting areas blocked by roadway equipment or corners, and outputs a dynamically calculated "working lighting area" that moves with the miner. This area accurately represents the three-dimensional spatial polygon covered by effective lighting within the current miner's field of vision and working range.

[0053] Specifically, when a miner wearing a smart coal mine lighting lamp walks to a sharp bend in the coal mine tunnel, the cloud monitoring module captures the real-time posture data uploaded by the smart coal mine lighting lamp, showing that the miner's head posture is turned 45 degrees to the left (preparing to observe the blind spot of the bend); the cloud immediately sets this posture as a spatial anchor point and coordinates the use of the L1 and L2 tunnel lighting lamps located on both sides of the bend; after checking, the current illumination tilt angle of L1 lamp is 30 degrees downward and the luminous flux is 3000 lumens, while the L2 lamp is horizontal and the luminous flux is 2500 lumens.

[0054] The cloud synchronously inputs the pose coordinates of the coal mine lighting smart lamps and the aforementioned parameters of lamps L1 and L2 into the virtual space; the light field reconstruction mechanism begins to operate: due to the branch bends in the coal mine tunnel, the system simulates and calculates that part of the beam of lamp L1 is blocked by the protruding rock wall of the tunnel, forming a shadow area; while the coal mine lighting smart lamp, as a moving light source, has its light spot pointing towards the blind spot of the bend; by vector superposition and illuminance integration of these three beams of light, an irregular polygonal working lighting area is calculated in real time. This area dynamically extends as the miner moves forward, not only covering the working face directly in front of the miner, but also eliminating the visual blind spot at the bend through the complementary light fields of the coal mine lighting smart lamp and lamp L1, realizing dynamic adaptive lighting for complex tunnel environments.

[0055] Furthermore, the intelligent coal lighting lamp has multiple built-in environmental sensors. During the lighting process, the intelligent coal lighting lamp collects corresponding underground environmental information through multiple environmental sensors. The multiple underground environmental information includes gas concentration, temperature and humidity, and dust particles.

[0056] At this time, the intelligent coal lighting lamp integrates a multi-sensor fusion sensing module. During the lighting operation, the main control unit coordinates the excitation of various environmental sensors according to the preset sampling frequency. For the gas environment, electrochemical sensors or infrared optical sensors convert the gas concentration in the roadway air into electrical signals in real time. For the meteorological environment, temperature and humidity sensors use thermistor and humidity resistor elements to sense the current temperature and relative humidity. For the aerosol environment, laser scattering dust sensors emit beams of specific wavelengths and receive particulate scattering signals to measure the concentration of dust particles in the air. All sensors work in parallel, converting physical environmental parameters into analog electrical signals and quantizing them into digital signals via analog-to-digital converters (ADCs).

[0057] The edge computing unit built into the intelligent coal lighting lamp performs signal conditioning and feature extraction on the acquired raw digital signals, and performs filtering to remove noise data caused by underground electromagnetic interference or equipment vibration. The system performs nonlinear correction on the gas concentration data to ensure that the measurement accuracy meets the coal mine safety regulations. At the same time, it combines temperature and humidity data to perform cross-compensation correction on the output of the dust sensor to eliminate dust measurement deviations caused by high humidity environment. The gas concentration value, temperature and humidity value, and dust concentration value are encapsulated into a structured underground environmental information data frame to complete the solution process from the bottom sensor signal to the high-level environmental semantics.

[0058] Specifically, when miners wearing smart coal mine lighting lamps reach the deep excavation face of a coal mine tunnel, the environmental parameters exhibit significant unique characteristics due to the underground location and relatively obstructed ventilation. The smart coal mine lighting lamp's built-in multimodal environmental sensing array automatically activates a high-frequency sampling mode. While the lighting beam illuminates the rock wall ahead, its built-in electrochemical gas sensor captures minute fluctuations in the methane concentration in the air in real time, outputting a concentration value of 0.65%. Simultaneously, the temperature and humidity sensor detects abnormal environmental parameters in the area due to the combined effects of geothermal heat and dust suppression spray, with a temperature of 32℃ and a relative humidity of 95%. At the same time, the laser scattering dust sensor detects a surge in the concentration of respirable dust particles suspended due to recent blasting operations, reaching 120 mg / m³. Meanwhile, the smart coal mine lighting lamp's microprocessor performs real-time fusion and calculation of this underground environmental information, identifying a special environmental state of "coexistence of high gas risk and high dust concentration." This data is not only temporarily stored in a local log but also provides accurate real-time physical field data support for mapping this environmental information to a virtual simulation space in subsequent steps.

[0059] Therefore, the cloud integrates multi-source heterogeneous data from multiple underground environmental information sources and multiple electronic fence areas, and uses a physics engine to construct a full-element virtual simulation environment in the cloud. Ray tracing and real-time rendering technology are applied to the virtual simulation environment to further output corresponding regional dynamic maps. These regional dynamic maps are stored in coal-fired intelligent lighting lamps and can intuitively reflect the current environmental safety situation, light blind spots, and fence intrusion risks. The introduction of multiple electronic fence areas further controls the working lighting area and improves the accuracy of the regional dynamic maps.

[0060] At this point, the cloud data center receives the underground environmental information data packet (containing time-series data of gas, temperature, humidity, and dust) uploaded from the coal-fired intelligent lighting lamp, as well as the electronic fence vector data generated in step S112. The system adopts a spatiotemporal alignment method to map the heterogeneous environmental scalar data to the spatial polygonal topology network of the electronic fence, realizing the deep integration of environmental attributes and spatial areas. On this basis, the cloud uses a physics engine to construct a full-element virtual simulation environment. This environment not only includes the geometric mesh model of the tunnel, but also integrates aerodynamic properties and light physical properties, transforming the gas concentration distribution into a visualized density field and the electronic fence into a volume field with collision detection properties, forming the basis of a digital twin.

[0061] Based on a virtual simulation environment, the cloud-based physical rendering engine performs global ray tracing calculations. The system simulates the physical processes of reflection, refraction, and scattering of photons emitted by the light source in the tunnel walls, equipment surfaces, and air dust media, accurately calculating the illuminance distribution and color temperature performance within the working lighting area. The rendering engine combines real-time gas concentration field and dust particle size to simulate the attenuation effect and Tyndall effect of light in turbid media, thereby identifying lighting blind spots caused by tunnel bends or equipment obstruction. Through texture mapping and volume rendering technology, it outputs a dynamic map of the region containing a safety situation spectrum, illuminance cloud map, and fence boundaries. This dynamic map exists in the form of a highly compressed video stream or keyframe image stream.

[0062] The cloud-generated dynamic map of the area is transmitted via IoT link and stored in the local storage unit of the coal-fired smart lighting lamp, serving as the baseline map for local decision-making. This dynamic map uses layer overlay technology to intuitively reflect the current safety status of the environment: the risk level of exceeding the standard for gas or dust concentration is represented by red, yellow, and green gradients; gray-scale shadow areas are used to mark blind spots; and flashing boundary lines are used to indicate the risk of intrusion into the electronic fence. Miners can perceive their spatial position relative to the risk source in real time through the miniature display screen configured on the coal-fired smart lighting lamp or through the coded signals projected by the light, realizing a closed-loop feedback from data collection to visual perception.

[0063] Specifically, when the cloud executes step S123, it performs multi-source heterogeneous data fusion between the environmental data uploaded by the coal-fired lighting smart lamps, such as "gas concentration 0.65% and dust 120mg / m³", and the vector data of the "absolute no-entry zone" near the F3 fault in the coal mine passage. Based on this, the cloud physics engine constructs a full-element virtual simulation environment for the coal mine passage, simulating the light scattering phenomenon caused by excessive dust concentration, as well as the gas accumulation cloud in the F3 fault area.

[0064] The system performs ray tracing and real-time rendering, calculating a blind spot (illuminance below 10 Lux) formed by the combined effects of rock wall obstruction and dust at the current tunnel corner, and marking it in the virtual view with a dark gray shadow. Simultaneously, the system renders areas of abnormal gas concentration as a semi-transparent red warning layer. The final output dynamic map of the area is then sent and stored in the local memory of the coal mine lighting smart lamp. At this point, the miner can visually see on the miniature display screen of the coal mine lighting smart lamp: a red risk area (predicted gas exceedance) and a gray blind spot at the bend in the mine tunnel ahead, with a green cursor indicating that they are at the edge of the "normal passage zone," thus gaining advance knowledge of the safety situation and blind spots ahead, providing a visualized basis for emergency response decisions.

[0065] refer to Figure 4 In step S13, the specific steps are as follows:

[0066] S131: Real-time monitoring of smart coal lighting lamps. When a smart coal lighting lamp detects an emergency and outputs an emergency signal, the emergency signal is directly transmitted to the cloud and triggers a cloud-to-cloud collaboration mechanism between the cloud and the smart coal lighting lamp. In this cloud-to-cloud collaboration mechanism, the collaborative data stream between the cloud and the smart coal lighting lamp is marked. The cloud uses a spatiotemporal graph convolutional network in the collaborative data stream to perform frame-by-frame dynamic recognition of the regional dynamic map, thereby extracting the emergency spread trend and the spatial relationship of personnel. Then, the regional dynamic map is decoupled from multiple factors, and multiple sub-regions with different risk levels are output, covering extremely high risk level, high risk level, or medium risk level.

[0067] S132: Mark the communication link between the cloud and the coal lighting smart lamp. The cloud outputs the corresponding high-frequency control command to the coal lighting smart lamp along the communication link, and combines it with the emergency control matrix of the emergency light terminal of the coal lighting smart lamp to perform multi-factor fusion, thereby triggering the early warning linkage between the emergency light terminals of the coal lighting smart lamp.

[0068] S133: By driving the multi-color LED chip of the coal mine lighting smart lamp, warning light strips of different colors corresponding to the risk level of each sub-area are projected onto the ground and side walls of the coal mine passage. The projection angle, length and color temperature gradient of the warning light strip can be dynamically adjusted adaptively as the position of the coal mine lighting smart lamp is adjusted, so as to construct wind shelter guidance content for miners.

[0069] In the embodiments of this application, real-time monitoring of smart coal-fired lighting lamps is performed. When a smart coal-fired lighting lamp outputs an emergency signal due to an emergency situation, the emergency signal is directly transmitted to the cloud and triggers a cloud-to-cloud collaboration mechanism between the cloud and the smart coal-fired lighting lamp. In this cloud-to-cloud collaboration mechanism, the collaborative data stream between the cloud and the smart coal-fired lighting lamp is marked. The cloud uses a spatiotemporal graph convolutional network in the collaborative data stream to perform frame-by-frame dynamic recognition of the regional dynamic map, thereby extracting the emergency spread trend and the spatial relationship of personnel. Then, the regional dynamic map is decoupled from multiple factors, and multiple sub-regions with different risk levels are output. The risk levels cover extremely high risk level, high risk level, or medium risk level.

[0070] At this time, the system performs real-time monitoring of the entire chain of the intelligent coal lighting lamps. When the threshold judgment logic of the internal sensors of the intelligent coal lighting lamps or the manual triggering device of the miners detects an abnormal situation, an emergency signal with a timestamp and a unique device identifier is immediately generated. This signal is transmitted to the cloud via the IoT communication module, which prioritizes the use of a high-reliability channel, breaking the conventional polling mode. After receiving the signal, the cloud immediately activates the cloud collaboration mechanism to establish a dedicated two-way high-bandwidth, low-latency data transmission tunnel. In this tunnel, the cloud control commands and the sensing data streams uploaded by the intelligent coal lighting lamps are collaboratively marked to ensure the integrity of the data and the highest priority processing rights during emergency rescue.

[0071] In the cloud-based collaborative data stream, the system calls a pre-trained spatiotemporal graph convolutional network model to perform frame-by-frame dynamic scanning and recognition of the regional dynamic map generated in step S123. This network model combines the feature extraction capability of graph convolutional networks (GCN) for non-Euclidean spatial structures (such as tunnel topology networks) with the ability of temporal convolutional networks (TCN) to capture dynamic changes in time series. By performing convolution operations on the dynamic map of consecutive frames, the system extracts the spatiotemporal evolution features of disasters or abnormal situations from the complex image pixels, and analyzes the spatial diffusion trajectory of the risk source (emergency spread trend) and the motion vector relationship of miners relative to the risk source (personnel spatial relationship).

[0072] Based on the extracted spatiotemporal features, the cloud-based system performs multi-factor decoupling analysis on the regional dynamic map, decomposing the single environmental data layer into a disaster physical field (such as fire spread and gas diffusion direction), a personnel location field, and a disaster evacuation path topology field. The system calculates risk weight values ​​for different spatial locations based on the distance to the risk source, diffusion speed, and roadway ventilation conditions, and performs regional segmentation accordingly. The overall regional dynamic map is divided into multiple sub-regions with clear boundaries and assigned corresponding risk level labels, including "extremely high risk level" with direct life threat, "high risk level" with potential harm risk, and "medium risk level" which is relatively safe, forming a visualized risk zoning decision map.

[0073] Specifically, when a miner wearing a smart coal mine lighting lamp walks near an old goaf in a coal mine passage, the gas sensor of the smart coal mine lighting lamp detects that the concentration instantly exceeds the alarm threshold, immediately generates a "gas over-limit emergency signal" and transmits it directly to the cloud, instantly triggering the cloud collaboration mechanism; the cloud then establishes a dedicated collaborative data stream and locks the virtual simulation view of the coal mine passage.

[0074] In this collaborative data stream, the cloud uses a spatiotemporal graph convolutional network to analyze the regional dynamic map of the coal mine tunnel frame by frame. Considering that the coal mine tunnel has branches and uneven wind speed, the network model accurately captures that the gas is spreading rapidly downwind along the main tunnel, and due to the backflow effect of the branch tunnel, some gas is flowing back into the branch tunnel. Based on this, the system extracts that the emergency spread trend is spreading to the depth of the main tunnel, and at the same time calculates that the miner's current position is upwind of the spread path and is moving towards a dead-end branch tunnel, making the spatial relationship between personnel extremely dangerous.

[0075] The system decouples multiple factors and overlays the gas concentration field with the miner's trajectory. The final risk zoning map shows that the dead-end side roadway that the miner is about to enter is marked as "extremely high risk level" (dead-end roads are prone to accumulating high concentrations of gas), the main roadway area where the miner is currently located is marked as "high risk level" because the concentration is rising, and the main roadway exit direction behind the miner is marked as "medium risk level" because of good ventilation. This classification result provides the core decision-making basis for the subsequent lighting guidance of intelligent coal lighting.

[0076] Furthermore, the communication link between the cloud and the smart coal lighting lamp is marked. The cloud outputs corresponding high-frequency control commands to the smart coal lighting lamp along the communication link, and combines the emergency control matrix of the emergency light terminal of the smart coal lighting lamp to perform multi-factor fusion, thereby triggering the early warning linkage between the emergency light terminals of the smart coal lighting lamp, thus introducing the early warning linkage between the emergency light terminals of the smart coal lighting lamp.

[0077] At this point, after the cloud-based collaborative mechanism is activated, the system logically marks the IoT communication link between the cloud server and the smart coal lighting lamp, designating it as an "emergency dedicated transmission channel" and implementing a queue scheduling strategy with the highest priority to avoid network congestion. Based on the risk level sub-region division results generated in step S131, the cloud generates a high-frequency control command sequence for a specific device ID. This command sequence includes control bytes such as light color temperature parameters, strobe frequency parameters, luminous flux output amplitude, and beam projection angle. The cloud sends these commands to the embedded control unit of the smart coal lighting lamp along the marked communication link at millisecond intervals to ensure real-time synchronization between control commands and risk status.

[0078] After receiving control commands from the cloud, the smart coal-fired lighting lamps call upon their built-in "emergency control matrix" to perform localized strategy fusion. This matrix is ​​a multi-dimensional mapping function, with input variables including risk level commands from the cloud, real-time environmental feedback values ​​collected by local sensors, and the miner's current posture data. The system uses a weighted fusion method to calculate the global risk decisions from the cloud and the local perception status of the lamp, outputting the final driving parameters. For example, when the cloud command requires a "red warning" but the local detection shows that the dust concentration is too high and affects penetration, the matrix will automatically adjust the light intensity compensation coefficient to ensure optimal visibility in complex environments.

[0079] Based on the fused driving parameters, the intelligent coal lighting lamp driving circuit triggers the early warning linkage mechanism of the emergency light terminal; the main control chip precisely controls the LED driving current through pulse width modulation (PWM) technology, driving the multi-color temperature lamp array to perform specific optical actions. This linkage mechanism not only controls the flashing and color changing of individual lamps, but also achieves spatiotemporal synchronization of lighting effects through preset protocols with surrounding channel lighting lamps (such as DMX512 or wireless Mesh networking protocols); at this time, the intelligent coal lighting lamp is not only a lighting tool, but also transforms into an audible and visual alarm, providing miners with intuitive sensory warnings through differentiated light effect codes (such as "red light flashing mode" representing extremely high risk), completing the closed loop from cloud decision-making to physical execution.

[0080] Specifically, when the cloud determines that the dead-end side road in front of the miner is an "extremely high-risk area" and the main roadway where the miner is currently located is a "high-risk area", the system immediately marks the communication link between the cloud and the coal lighting smart lamp in red and raises its priority to the highest level. The cloud then generates a set of high-frequency control commands containing "red light band (620nm), strobe frequency 10Hz, and luminous flux 100%", and sends them to the coal lighting smart lamp without delay along the marked link.

[0081] After receiving the instruction, the intelligent coal mine lighting lamp immediately initiates multi-factor fusion calculation using its built-in emergency control matrix. Considering the high dust concentration (120mg / m³) in the coal mine passage, the matrix method determines that the penetration of ordinary red light is insufficient. Therefore, it automatically corrects the light intensity drive current, increases the light flux output to 120% of the rated power (overload drive mode), and locks the beam projection angle based on the miner's head posture facing the danger zone.

[0082] The emergency light terminal of the coal mine lighting smart lamp instantly triggers an early warning linkage: the main light source switches from conventional white light to a high-penetration red light flashing mode, flashing intensely at a frequency of 10 times per second, directly illuminating the entrance to the dead end ahead; at the same time, the coal mine lighting smart lamp links with the nearby passage light groups through a wireless mesh network, causing them to switch to constant red light synchronously, forming a clear "danger boundary light band" in the dim and complex coal mine passage, intuitively warning miners that the area ahead is absolutely prohibited from entering, thus effectively avoiding the risk of accidental entry due to poor visibility or delayed information.

[0083] Therefore, by driving the multi-color LED chip of the coal mine lighting smart lamp, warning light strips of different colors corresponding to the risk level of each sub-area are projected onto the ground and side walls of the coal mine passage. The projection angle, length and color temperature gradient of the warning light strip can be dynamically adjusted adaptively as the position of the coal mine lighting smart lamp is adjusted, so as to construct wind shelter guidance content for miners.

[0084] At this point, based on the risk level sub-regions defined in step S131, the system drives the multi-primary-color LED chip array integrated inside the coal mine lighting smart lamp to perform a light mixing method. By adjusting the PWM duty cycle of the red, green, blue, and warm white light chips, the system precisely synthesizes warning light colors that strictly correspond to the risk level color spectrum of each sub-region. The coal mine lighting smart lamp uses its precise optical lens group to project these beams of light with specific color temperature and chromaticity onto the ground and side walls of the coal mine passage, forming a highly visible "warning light strip" in physical space. This light strip, as a visual augmented reality sign, directly covers the surface of the risk area, filling the gap where traditional signs are invisible in the complex underground environment.

[0085] The main control unit of the intelligent coal mine lighting lamp calculates the miner's movement speed, position coordinates, and head posture angle in real time, and adaptively adjusts the projection geometry parameters of the warning light strip accordingly. In terms of projection angle, the system dynamically corrects the optical axis pitch and yaw angles by combining IMU inertial measurement unit data to ensure that the light strip always stably covers the target area. In terms of length, the system dynamically adjusts the extension range of the light strip by adjusting the focal length of the light spot or by combining beam widening technology, based on the cross-sectional dimensions of the roadway and the distance to the boundary of the risk area. At the same time, the system controls the gradual change of color temperature gradient according to the distance from the risk source, that is, presenting a continuous transition of color temperature from cool (high-risk red / blue) to warm (safe green / yellow) from the risk center to the periphery, using color psychology to enhance the risk avoidance guidance effect.

[0086] As miners adjust their positions, the relative spatial relationship between the warning light strip and the risk area changes in real time. The system constructs dynamic risk avoidance guidance content through continuous light strip redrawing. When miners turn or move, the cloud collaboratively updates the risk situation, and the coal mine lighting smart lights immediately recalculate the projection parameters to ensure that the "safe direction" is always guided by a low-saturation or green light strip, while the "dangerous direction" is blocked by a high-saturation red light strip. This dynamic light and shadow navigation mechanism can provide miners with intuitive escape path guidance without cognitive load conversion in dusty and dimly lit tunnels, constructing a clear visual safety passage.

[0087] Specifically, when a miner encounters a gas over-limit emergency in a coal mine tunnel, the intelligent coal lighting immediately executes step S133; the cloud identifies the dead-end side tunnel ahead as an "extremely high-risk area," the main tunnel exit behind as a "medium-risk area," and an "escape route"; the intelligent coal lighting drives the multi-color LED chip to project a high-saturation red warning light strip onto the ground ahead, which precisely lands at the entrance of the side tunnel, clearly marking "absolutely prohibited"; at the same time, a green guiding light strip is projected onto the ground behind the miner, pointing towards the safety exit.

[0088] As miners begin to retreat, the intelligent coal mine lighting dynamically adjusts based on their position: when miners run faster, the green light strip automatically extends its projection distance by adjusting the lens focal length, ensuring that the guidance range always covers the miners' field of vision; at the same time, the light strip displays a red-yellow-green color temperature gradient at the end closest to the danger source, using color gradation to indicate the risk attenuation trend; in the environment of extremely high dust concentration and extremely low visibility in coal mine tunnels, this colorful light strip, which dynamically extends and changes direction with the miners' movement, becomes the only visual lifeline in the miners' eyes, realizing the construction of intuitive and efficient risk avoidance guidance content.

[0089] refer to Figure 5 In step S14, the specific steps are as follows:

[0090] S141: The cloud performs multi-semantic parsing on the emergency signals uploaded by the smart coal lighting lamps and identifies multiple emergency contents during the parsing process. These multiple emergency contents present emergency information in different dimensions. Simultaneously, the current images captured by the cameras built into the smart coal lighting lamps are retrieved, and the multimodal framework deployed in the cloud performs image-text linkage verification. Thus, multiple risk features are output through feature alignment and abnormal target detection. These multiple risk features cover the risk type, risk intensity, and corresponding risk content.

[0091] S142: Multiple risk characteristics and the illumination positions of various warning light strips are fused together, and the corresponding matching framework is determined during the fusion process. The working status of the coal lighting smart lamp is then input into the cross-matching network. In this cross-matching process, a spatial-channel joint attention mechanism is introduced to dynamically focus on the matching relationship between risk characteristics and light strip projection areas, thereby suppressing redundant environmental interference and gradually building an intelligent control system for coal lighting smart lamps with high robustness and self-learning ability.

[0092] In the embodiments of this application, the cloud performs multi-semantic parsing on the emergency signals uploaded by the smart coal lighting lamps and determines multiple emergency contents during the parsing process. The multiple emergency contents present emergency information in different dimensions. The current image captured by the camera built into the smart coal lighting lamp is retrieved simultaneously, and the image and text linkage is reviewed in a multimodal framework deployed in the cloud. Thus, multiple risk features are output through feature alignment and abnormal target detection. The multiple risk features cover the risk type, risk intensity and corresponding risk content, introducing the risk type, risk intensity and corresponding risk content.

[0093] At this time, after receiving the emergency signal uploaded by the smart coal lighting lamp, the cloud starts a multi-semantic parsing engine to deeply deconstruct the data packet. This parsing process goes beyond simple threshold triggering logic. Instead, it is based on knowledge graph technology to semantically associate the device status code, sensor value sequence, and trigger timestamp in the signal, thereby determining the emergency content in multiple dimensions. Specifically, the system transforms the original signal into a structured cluster of emergency information, including "object dimension" (such as the identity of the alarm subject), "time dimension" (such as duration and frequency), "spatial dimension" (such as the specific location of occurrence), and "physical dimension" (such as the specific gas concentration value or temperature value), forming a multi-dimensional emergency situation description.

[0094] While analyzing emergency signals, the cloud simultaneously retrieves current video frames or key images captured by the built-in camera of the coal-fired smart lighting lamp through a high-bandwidth link. The system inputs the analyzed structured emergency content and the real-time acquired visual images into a multimodal fusion framework deployed in the cloud to perform image-text linkage verification. This framework uses a cross-modal attention mechanism to align the text-based alarm information (such as "gas concentration exceeds the standard") with the image pixel features in the feature space, verifying whether the abnormal sensor values ​​have corresponding physical representations at the visual level (such as whether there is smoke, flames, or roof collapse), thereby eliminating false alarms caused by sensor drift or accidental touch.

[0095] Based on the image-text linkage verification results, the system performs feature alignment and abnormal target detection on the visual images; it extracts high-frequency features from the images through convolutional neural networks and, combined with semantic guidance from emergency content, identifies abnormal areas in the images; the system outputs multiple standardized risk features, which cover specific risk types (such as gas accumulation, water seepage, fire, etc.), risk intensity (such as smoke density level, flame spread range), and corresponding risk content (such as specific risk source descriptions); this step realizes a cognitive leap from abstract signals to concrete scenarios, providing accurate decision-making basis for subsequent intelligent control.

[0096] Specifically, when the intelligent coal lighting lamp detects an anomaly deep in the coal mine passage and sends an emergency signal, the cloud executes step S141; the cloud performs multi-semantic analysis on the emergency signal, deconstructing multiple emergency contents such as "gas concentration suddenly increased to 1.2%" and "location is near the F3 fault fracture zone", and judges that this is a complex early warning involving geology and gas environment.

[0097] The system synchronously retrieves current images captured by smart coal mine lighting cameras in the dimly lit environment of coal mine tunnels in the cloud. Within a multimodal framework deployed in the cloud, the system performs image-text linkage verification between the textual semantics of "high gas concentration" and the visual features in the images. Through low-light image enhancement technology, it identifies faint bluish-gray smoke aerosol features deep within the images and aligns these visual features with the semantics of "gas alarm." The system outputs multiple risk features: the risk type is determined to be "gas accumulation accompanied by dust and smoke from suspected roof collapse"; the risk intensity is "moderate diffusion"; and the risk content is "abnormal gas outburst caused by roof fracture in the F3 fault area." This process effectively eliminates the possibility of simple sensor failure and provides solid multimodal evidence to support the subsequent development of targeted risk avoidance plans.

[0098] Furthermore, multiple risk characteristics and the illumination positions of various warning light strips are fused using a multi-factor approach. During this fusion process, a corresponding matching framework is determined. This framework is then integrated with the operating status of the intelligent coal lighting lamp into a cross-matching network. This cross-matching process introduces a spatial-channel joint attention mechanism, dynamically focusing on the matching relationship between risk characteristics and the projection area of ​​the light strips. This suppresses redundant environmental interference and gradually constructs an intelligent control system for the intelligent coal lighting lamp with high robustness and self-learning capabilities. This system comprehensively considers multiple risk characteristics and the illumination positions of various warning light strips, ensuring the accuracy of the corresponding matching framework. Simultaneously, the cloud-to-cloud collaboration between the cloud and the intelligent coal lighting lamp is further controlled, fully considering multiple risk characteristics, the illumination positions of various warning light strips, and the operating status of the intelligent coal lighting lamp, thus improving the accuracy of the intelligent control system.

[0099] At this point, the cloud control center performs multi-factor fusion of the discrete risk features (risk type, intensity, and content) extracted in step S141 with the actual projection position parameters (three-dimensional coordinates and coverage area) of the warning light strip generated in step S133. During this process, the system establishes a matching framework based on semantic scene graphs, mapping abstract risk attributes to specific optical control parameters. For example, the feature of "high concentration of methane" is logically bound to "red light strip projection area" to establish the topological correspondence between risk and response, forming a standardized control strategy template, and providing structured input for subsequent neural network processing.

[0100] The system inputs the fused dataset along with the real-time operating status of the smart coal-fired lighting lamps (such as remaining power, luminous flux attenuation, and heat dissipation status) into a cross-matching network. The core of this network introduces a spatial-channel joint attention mechanism: In the spatial dimension, attention weights focus on the spatial overlap between risk features and light projection areas in the tunnel space, strengthening attention to dangerous areas and suppressing responses to safe backgrounds; In the channel dimension, weights are assigned to different types of risk features (such as gas and dust), prioritizing the processing of feature channels with higher hazard levels. Through this dual suppression mechanism, complex electromagnetic interference and environmental noise underground are effectively filtered out, resolving the problem of multi-source data conflicts.

[0101] Through deep computation of the cross-matching network, the system outputs an optimized control strategy vector and dynamically adjusts the projection logic of the light strip. This process is not a one-time static output, but rather the construction of an intelligent control system with closed-loop feedback capability. The system continuously records changes in risk characteristics and light strip response effects, and updates network weights using online learning to gradually improve the robustness of the model. This means that as operational data accumulates, the system's ability to identify risks and coordinate lighting control in complex alleyway environments will continue to evolve, achieving a leap from "rule-driven" to "cognitive-driven".

[0102] Specifically, when the coal mine lighting smart lamp detects both "gas accumulation" and "dust suspension" risk characteristics simultaneously in the coal mine passage, step S142 is immediately initiated; the cloud integrates "gas accumulation (extremely high risk)" with the position of the red light band projected by the miner lamp on the ground in front, and "dust suspension (medium risk)" with the position of the yellow light band on the side wall, to establish a matching framework with gas risk avoidance as the core.

[0103] Considering that the current remaining power of the coal-fired intelligent lighting lamp is only 40% and it is in a high-load working state, the system inputs these parameters into the cross-matching network; the spatial-channel joint attention mechanism in the network starts to operate: in the spatial dimension, the network dynamically focuses on the F3 fault fracture zone area (high-risk overlapping area) in front, suppressing the light band adjustment interference in the safe area behind the roadway; in the channel dimension, "gas characteristics" are given a higher weight coefficient, reducing the computational power occupation of "dust characteristics".

[0104] The system has developed a highly robust control strategy: even under harsh conditions such as unstable electromagnetic signals in coal mine tunnels and dust interference with optical path recognition, the intelligent coal lighting lamp can still make intelligent decisions, concentrate limited power to maintain the high-frequency flashing of the red warning light strip in front, and appropriately reduce the brightness of the yellow light strip on the side wall. In this way, while ensuring the core risk avoidance function, it achieves optimal resource allocation and precise risk response with self-learning capabilities.

[0105] Please see Figure 6 , Figure 6 This is a schematic diagram of the structural composition of the intelligent control system for a cloud-based collaborative coal-fired lighting smart lamp according to an embodiment of the present invention; the intelligent control system for the cloud-based collaborative coal-fired lighting smart lamp is applied to the aforementioned intelligent control method for the cloud-based collaborative coal-fired lighting smart lamp; the intelligent control system for the cloud-based collaborative coal-fired lighting smart lamp includes:

[0106] The electronic fence module 21 is used for miners to wear smart coal lighting lamps and walk dynamically along the coal mine passage. The smart coal lighting lamps are connected to the corresponding cloud based on the Internet of Things. In the cloud, multiple electronic fence areas are determined based on the location data of the smart coal lighting lamps and the distribution map of the coal mine passage.

[0107] The virtual simulation module 22 is used to determine the corresponding working lighting area based on the intelligent coal lighting lamp and the multiple channel lighting lamps configured in the coal mine channel, and to perform virtual simulation based on the underground environmental information collected by the intelligent coal lighting lamp and multiple electronic fence areas, and output the corresponding dynamic map of the area.

[0108] The cloud collaboration module 23 is used to trigger the cloud collaboration mechanism between the cloud and the coal lighting smart lamp if the coal lighting smart lamp outputs an emergency signal. During the collaboration process, the dynamic map of the area is dynamically identified to output multiple sub-areas with different risk levels. It is also used to link the emergency light terminal of the coal lighting smart lamp to output warning light strips of different colors to the coal mine passage. Each warning light strip is dynamically adjusted as the position of the coal lighting smart lamp is adjusted.

[0109] The intelligent control module 24 is used to determine multiple emergency contents based on the analysis of the emergency signal, and to perform linkage verification in conjunction with the current image collected by the coal lighting intelligent lamp to output multiple risk features. It performs cross-matching of multiple risk features, the illumination position of each warning light strip and the working status of the coal lighting intelligent lamp, and incorporates an attention mechanism in the matching process, thereby gradually building an intelligent control system for the coal lighting intelligent lamp.

[0110] It should be noted that although multiple modules are mentioned in the detailed description above, this division is not mandatory. In fact, according to embodiments of this disclosure, the features and functions of two or more modules or modules described above can be embodied in one module. Conversely, the features and functions of one module described above can be further divided and embodied by multiple modules.

[0111] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the disclosure herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only.

[0112] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A smart control method for coal-fired lighting smart lamps based on cloud collaboration, characterized in that, include: Miners wear smart coal lighting lamps and walk dynamically along the coal mine passages. The smart coal lighting lamps are connected to the corresponding cloud based on the Internet of Things. In the cloud, multiple electronic fence areas are determined based on the location data of the smart coal lighting lamps and the distribution map of the coal mine passages. The corresponding working lighting area is determined based on the intelligent coal lighting lamp and the multiple channel lighting lamps configured in the coal mine channel. The underground environmental information collected by the intelligent coal lighting lamp and multiple electronic fence areas are combined to perform virtual simulation and output the corresponding dynamic map of the area. If the coal mine lighting smart lamp outputs an emergency signal, it triggers the cloud-to-cloud collaboration mechanism between the cloud and the coal mine lighting smart lamp. During the collaboration process, the regional dynamic map is dynamically identified to output multiple sub-regions with different risk levels. It also combines the emergency light terminal of the coal mine lighting smart lamp to carry out early warning linkage to output light strips of different colors to the coal mine passage. Each light strip is dynamically adjusted as the position of the coal mine lighting smart lamp is adjusted. Based on the analysis of the emergency signal, multiple emergency contents are determined, and linked verification is performed in conjunction with the current images collected by the coal lighting smart lamp to output multiple risk features. The multiple risk features, the illumination positions of each light strip, and the working status of the coal lighting smart lamp are cross-matched, and an attention mechanism is incorporated into the matching process to gradually build an intelligent control system for the coal lighting smart lamp.

2. The intelligent control method for coal-fired lighting smart lamps based on cloud collaboration according to claim 1, characterized in that, The miners wear smart coal mine lighting lamps and move dynamically along the coal mine tunnels. These lamps are connected to a corresponding cloud platform via the Internet of Things (IoT). Within this cloud platform, multiple electronic fence zones are determined based on the location data of the smart coal mine lighting lamps and the distribution map of the coal mine tunnels. These zones include: The smart coal mine lighting lamp is worn on the miner's head and provides dynamic lighting in the coal mine tunnel as the miner moves. As the miner walks in the tunnel, the location data of the smart coal mine lighting lamp is displayed to show the miner's current location. At the same time, the location data of the smart coal mine lighting lamp is collected and synchronized to the cloud at high frequency in conjunction with the Internet of Things in the coal mine tunnel.

3. The intelligent control method for coal-fired lighting smart lamps based on cloud collaboration according to claim 2, characterized in that, The miners wear smart coal mine lighting lamps and move dynamically along the coal mine tunnels. These smart lamps are connected to a corresponding cloud platform via the Internet of Things (IoT). Within this cloud platform, multiple electronic fence zones are determined based on the location data of the smart lamps and the distribution map of the coal mine tunnels. The system also includes: The distribution map of the coal mine tunnels is obtained, which includes the tunnel topology, geological faults, and hidden disaster-causing factors. At the same time, after receiving the data, the cloud performs spatial mapping between the location data of the coal mine lighting smart lamps and the distribution map of the coal mine tunnels. Through a dynamic geometric calculation mechanism, based on the miner's current location and combined with preset safety location thresholds, multiple electronic fence areas are adaptively divided. These multiple electronic fence areas include absolutely prohibited areas, over-limit warning areas, and regular passage areas.

4. The intelligent control method for coal-fired lighting smart lamps based on cloud collaboration according to claim 1, characterized in that, The process involves determining the corresponding working lighting area based on the intelligent coal mine lighting lamp and multiple channel lighting lamps configured in the coal mine passage, and performing virtual simulation by combining the underground environmental information collected by the intelligent coal mine lighting lamp and multiple electronic fence areas, and outputting the corresponding dynamic map of the area, including: Real-time monitoring is performed on the cloud. The cloud uses the real-time pose of the coal mine lighting smart lamp as a spatial anchor point and coordinates the illumination tilt angle and luminous flux parameters of multiple channel lights that have been networked in the coal mine channel. The real-time pose of the coal mine lighting smart lamp and the illumination tilt angle and luminous flux parameters of multiple channel lights are input into the same virtual space. In this virtual space, a working lighting area that is calculated in real time as the miner moves is constructed in combination with the light field reconstruction mechanism.

5. The intelligent control method for coal-fired lighting intelligent lamps based on cloud collaboration according to claim 4, characterized in that, The process of determining the corresponding working lighting area based on the intelligent coal mine lighting lamp and multiple channel lighting lamps configured in the coal mine channel, and performing virtual simulation by combining the underground environmental information collected by the intelligent coal mine lighting lamp and multiple electronic fence areas, and outputting the corresponding dynamic map of the area, also includes: The intelligent coal lighting lamp has multiple environmental sensors built in. During the lighting process, the intelligent coal lighting lamp collects corresponding underground environmental information through multiple environmental sensors, including gas concentration, temperature and humidity and dust particles. The cloud platform integrates multi-source heterogeneous data from multiple underground environmental information sources and multiple electronic fence areas, and uses a physics engine to construct a full-element virtual simulation environment in the cloud. Ray tracing and real-time rendering technologies are applied to the virtual simulation environment to further output corresponding dynamic maps of the area. These dynamic maps are stored in the coal-fired intelligent lighting lamps and can intuitively reflect the current environmental safety situation, light blind spots, and fence intrusion risks.

6. The intelligent control method for coal-fired lighting smart lamps based on cloud collaboration according to claim 1, characterized in that, If the coal mine lighting smart lamp outputs an emergency signal, it will trigger a cloud-based collaborative mechanism between the cloud and the coal mine lighting smart lamp. During the collaborative process, the regional dynamic map will be dynamically identified to output multiple sub-regions with different risk levels. The emergency light terminal of the coal mine lighting smart lamp will be used for early warning linkage to output warning light strips of different colors to the coal mine passage. The various warning light strips are dynamically adjusted according to the position of the intelligent coal lighting lamps, including: The system monitors smart coal-fired lighting lamps in real time. When a smart coal-fired lighting lamp detects an emergency and outputs an emergency signal, the emergency signal is directly transmitted to the cloud, triggering a cloud-to-cloud collaboration mechanism between the cloud and the smart coal-fired lighting lamp. In this cloud-to-cloud collaboration mechanism, the collaborative data stream between the cloud and the smart coal-fired lighting lamp is marked. The cloud uses a spatiotemporal graph convolutional network in the collaborative data stream to perform frame-by-frame dynamic recognition of the regional dynamic map, thereby extracting the emergency spread trend and the spatial relationship between people. This allows for multi-factor decoupling of the regional dynamic map, resulting in the output of multiple sub-regions with different risk levels, including extremely high risk, high risk, and medium risk.

7. The intelligent control method for coal-fired lighting smart lamps based on cloud collaboration according to claim 6, characterized in that, If the coal mine lighting smart lamp outputs an emergency signal, it will trigger a cloud-based collaborative mechanism between the cloud and the coal mine lighting smart lamp. During the collaborative process, the regional dynamic map will be dynamically identified to output multiple sub-regions with different risk levels. The emergency light terminal of the coal mine lighting smart lamp will be used for early warning linkage to output warning light strips of different colors to the coal mine passage. The various warning light strips are dynamically adjusted according to the position of the intelligent coal lighting lamps, and also include: The communication link between the cloud and the smart coal-fired lighting lamp is marked. The cloud outputs corresponding high-frequency control commands to the smart coal-fired lighting lamp along this communication link, and combines multiple factors with the emergency control matrix of the emergency light terminal of the smart coal-fired lighting lamp to trigger early warning linkage between the emergency light terminals of the smart coal-fired lighting lamp. By driving the multi-color LED chip of the coal mine lighting smart lamp, warning light strips of different colors corresponding to the risk level of each sub-area are projected onto the ground and side walls of the coal mine passage. The projection angle, length and color temperature gradient of the warning light strip can be dynamically adjusted adaptively as the position of the coal mine lighting smart lamp is adjusted, so as to construct wind shelter guidance content for miners.

8. The intelligent control method for coal-fired lighting smart lamps based on cloud collaboration according to claim 1, characterized in that, The process involves determining multiple emergency responses based on the analysis of the emergency signal, and then performing a linked verification based on the current images collected by the intelligent coal lighting lamp to output multiple risk features. This is followed by cross-matching of these risk features, the illumination positions of various warning light strips, and the operating status of the intelligent coal lighting lamp. An attention mechanism is incorporated into the matching process to gradually construct an intelligent control system for the intelligent coal lighting lamp, including: The cloud performs multi-semantic parsing on the emergency signals uploaded by the smart coal lighting lamps and identifies multiple emergency contents during the parsing process. These multiple emergency contents present emergency information from different dimensions. Simultaneously, the current images captured by the cameras built into the smart coal lighting lamps are retrieved and reviewed in a multi-modal framework deployed in the cloud. This allows for the output of multiple risk features through feature alignment and abnormal target detection. These risk features cover the risk type, risk intensity, and corresponding risk content.

9. The intelligent control method for a cloud-based collaborative coal-fired lighting smart lamp according to claim 8, characterized in that, The process involves determining multiple emergency responses based on the analysis of the emergency signal, and performing a linked verification with the current images collected by the intelligent coal lighting lamp to output multiple risk features. This includes cross-matching these risk features, the illumination positions of various warning light strips, and the operating status of the intelligent coal lighting lamp, incorporating an attention mechanism during the matching process. This gradually constructs an intelligent control system for the intelligent coal lighting lamp. The process also includes: By fusing multiple risk characteristics and the illumination positions of various warning light strips, and determining the corresponding matching framework during the fusion process, the working status of the coal lighting smart lamp is further incorporated into the cross-matching network. This introduces a spatial-channel joint attention mechanism in the cross-matching process, dynamically focusing on the matching relationship between risk characteristics and the light strip projection area, thereby suppressing redundant environmental interference and gradually building an intelligent control system for coal lighting smart lamps with high robustness and self-learning capabilities.

10. An intelligent control system for a cloud-based collaborative coal-fired lighting smart lamp, characterized in that, The intelligent control system of the cloud-based collaborative coal lighting smart lamp is applied to the intelligent control method of the cloud-based collaborative coal lighting smart lamp as described in any one of claims 1-9. The intelligent control system for the cloud-based collaborative coal-fired lighting smart lamp includes: The electronic fence module is used by miners to wear smart coal lighting lamps and move dynamically along the coal mine passage. The smart coal lighting lamps are connected to the corresponding cloud based on the Internet of Things. In the cloud, multiple electronic fence areas are determined based on the location data of the smart coal lighting lamps and the distribution map of the coal mine passage. The virtual simulation module is used to determine the corresponding working lighting area based on the intelligent coal lighting lamp and the multiple channel lighting lamps configured in the coal mine channel, and to perform virtual simulation by combining the underground environmental information collected by the intelligent coal lighting lamp and multiple electronic fence areas, and output the corresponding dynamic map of the area. The cloud collaboration module is used to trigger a cloud collaboration mechanism between the cloud and the smart coal lighting lamp when the smart coal lighting lamp outputs an emergency signal. During the collaboration process, the module dynamically identifies the regional dynamic map to output multiple sub-regions with different risk levels. It also combines the emergency light terminal of the smart coal lighting lamp to perform early warning linkage to output warning light bands of different colors to the coal mine passage. Each warning light band is dynamically adjusted as the position of the smart coal lighting lamp is adjusted. The intelligent control module is used to determine multiple emergency contents based on the analysis of the emergency signal, and to perform linkage verification in conjunction with the current image collected by the coal lighting intelligent lamp to output multiple risk features. It cross-matches multiple risk features, the illumination position of each warning light strip and the working status of the coal lighting intelligent lamp, and incorporates an attention mechanism in the matching process, thereby gradually building an intelligent control system for the coal lighting intelligent lamp.

Images