A multi-source perception data fusion and intelligent early warning system for coal mine tunnel

Abstract

The present application relates to the technical field of mine safety monitoring, and is a multi-source perception data fusion and intelligent early warning system for coal mine tunnels, specifically comprising: a mine intrinsically safe intelligent management and control terminal deployed at the entrance of the tunnel, a ground dispatching platform and a mobile client, the terminal accesses multiple types of physical quantity monitoring data, performs real-time fusion analysis through an edge computing and data fusion unit, generates a dynamic safety situation level, and immediately drives a field sound and light alarm device to respond based on the level, automatically controls in combination with personnel identity and permission recognition, automatically links to broadcast an emergency plan and dynamically generates an optimal disaster avoidance route guide when an early warning is triggered, the system supports local storage and offline work, and data is synchronized to the platform after network recovery, which improves the field perception capability and active prevention and control efficiency, realizes risk prevention and control forward movement and emergency response closed-loop management.

Description

Technical Field

[0001] This invention relates to the field of mine safety monitoring technology, and is a multi-source sensing data fusion and intelligent early warning system for coal mine roadways. Background Technology

[0002] Coal mine underground roadways, especially single-heading tunnels, are relatively independent semi-enclosed spaces and are also high-risk areas for coal mine accidents. Currently, coal mines are generally equipped with a variety of safety monitoring devices, such as gas monitoring systems, mine pressure monitoring systems, personnel positioning systems, and ventilation monitoring systems.

[0003] Coal mines present complex underground environments and diverse disaster risks. Existing safety monitoring equipment operates with each subsystem functioning independently, resulting in severe data silos and fragmented safety information. This makes it difficult for workers to quickly obtain comprehensive safety information at tunnel entrances. Furthermore, early warnings require analysis by the ground dispatch center before being relayed down through multiple levels, leading to prolonged response times and limitations imposed by network stability. Personnel entering tunnels cannot obtain real-time information about the internal safety situation, resulting in a lack of safety confirmation. Entry into high-risk areas relies on self-reporting and manual checks, leading to insufficient control. Unauthorized personnel entering special work areas, especially when de-escalation zones are closed, lack relevant warnings, increasing the risk of accidents. Moreover, the emergency response and real-time monitoring and early warning systems are disconnected. When a disaster occurs, emergency broadcasts, emergency plan activation, and evacuation route guidance often rely on manual decision-making and notification, which is inefficient and prone to errors. Static evacuation routes may become ineffective due to changes in the disaster situation.

[0004] Based on this, the present invention discloses a multi-source sensing data fusion and intelligent early warning system for coal mine roadways. Summary of the Invention

[0005] The technical problem to be solved by this invention is that the existing technology cannot solve the problem of real-time, multi-dimensional safety status presentation at the entrance and on-site mandatory control, and the independent emergency broadcast system and emergency plan lack the ability to automatically link with the front-end real-time sensing data. Therefore, this invention proposes a multi-source sensing data fusion and intelligent early warning system for coal mine roadways.

[0006] The present invention provides a technical solution for a multi-source sensing data fusion and intelligent early warning system for coal mine roadways, comprising an intrinsically safe intelligent control terminal deployed at the roadway entrance, a ground dispatch platform, and a mobile client:

[0007] The intelligent control terminal is deployed at the entrance of the alleyway and configured to connect to multi-source sensors.

[0008] The terminal collects monitoring data of at least two different types of physical quantities in real time through the sensing data aggregation unit;

[0009] Locally on the terminal, the monitoring data is fused and analyzed in real time through edge computing and data fusion unit to generate a dynamic security status level.

[0010] Based on the security status level, immediately activate the local sound, light, and display alarm devices on the terminal to respond on-site;

[0011] When personnel approach, the system simultaneously identifies their identity and permissions, compares this information with the current security status level in real time, and executes on-site control actions based on preset strategies. When an alert is triggered or an emergency command is received, the system automatically links to execute voice broadcasts of emergency plans and graphical evacuation route guidance.

[0012] The locally generated operation logs and data are recorded to local storage and synchronized to the ground dispatch platform when network conditions permit.

[0013] Preferably, the multiple physical quantity monitoring data includes at least two or more of the following: gas concentration, roof pressure, wind speed, equipment operating status, personnel positioning information, oxygen concentration, water pressure and temperature for exploration and drainage, micro-vibration, ground sound, and drainage volume; the sensing data aggregation unit accesses the data through one or more of the following: industrial Ethernet, RS485, CAN bus, or wireless communication interface.

[0014] Preferably, the early warning model built into the edge computing and data fusion unit has an adaptive optimization mechanism, which can fine-tune the early warning threshold and fusion algorithm weight parameters in real time based on the historical monitoring data of the roadway, the early warning history record, and the parameter adjustment instructions issued by the external platform. It uses a deep neural network model to optimize the weight allocation, analyzes time series data through an LSTM model, and dynamically corrects the gas mutation threshold. The adaptive optimization mechanism utilizes the correlation between historical microseismic signals and roof displacement to predict weights through an LSTM model.

[0015] Preferably, the data fusion algorithm includes, but is not limited to, DS evidence theory or adaptive weighted fusion algorithm, whose output is a dynamic security situation level, graphically displayed on the screen of the human-computer interaction and display unit in the form of red, yellow, and green situation lights and key parameter curves.

[0016] Preferably, the personnel identification and authorization unit integrates RFID and UWB card reading modules, which synchronously obtain personnel identity information, safety training validity period and current electronic work ticket with the ground dispatch platform or through local caching, and automatically execute the action logic of release, warning or prohibition in combination with the comprehensive security situation level generated in real time by the edge computing and data fusion unit.

[0017] Preferably, the emergency broadcast linkage and disaster evacuation guidance unit is configured as follows:

[0018] Receive and parse structured emergency instruction packets from ground dispatch platforms or emergency broadcast systems, wherein the emergency instruction packets contain at least the warning level, the affected area, and the recommended action instructions;

[0019] When a local warning is triggered or the emergency instruction package is received, the system automatically matches and broadcasts the pre-stored voice emergency plan content that corresponds to the current risk type and location.

[0020] Based on real-time personnel location information and risk source locations, the mine's disaster avoidance route map leading to designated safety exits or refuge chambers is dynamically generated and highlighted on the screen of the human-computer interaction and display unit.

[0021] Preferably, the local storage and offline working unit ensures that the core functions of the sensing data aggregation unit, edge computing and data fusion unit, personnel identification and permission recognition unit, and emergency broadcast linkage and disaster avoidance guidance unit can continue to operate independently when network communication is interrupted, and data synchronization will be performed after the network is restored.

[0022] Preferably, the intelligent control terminal includes: a sensing data aggregation unit, an edge computing and data fusion unit, a human-computer interaction and display unit, a personnel identity and permission recognition unit, an emergency broadcast linkage and disaster avoidance guidance unit, and a local storage and offline working unit;

[0023] The data aggregation unit receives physical quantity monitoring data in real time; the processing and fusion unit runs the fusion algorithm to generate security assessment results; the human-computer interaction unit displays the assessment results; and the identity recognition unit verifies personnel information.

[0024] Preferably, the control actions include: issuing voice and text warnings based on the safety situation level when the personnel's work permit area does not match the risk level or when safety training has expired, and sending alarm information to a remote platform.

[0025] Preferably, it also includes algorithm iteration optimization: the processing and fusion unit periodically uses local historical data of the roadway to adjust the algorithm parameters through supervised learning training to improve the accuracy of early warning for specific working conditions, and is equipped with a mobile client, i.e. a specially equipped mobile phone, which can read data directly from the display terminal via Bluetooth.

[0026] Compared with the prior art, the technical effects of the present invention are as follows:

[0027] 1. This invention utilizes an intrinsically safe intelligent control terminal deployed at the entrance of a mine roadway to connect in real time to multiple sensors such as gas concentration and roof pressure. By collecting and processing various physical quantity monitoring data through a sensing data aggregation unit, it moves safety information and control checkpoints forward to the source of risk. Before entering the roadway, workers can intuitively obtain the dynamic safety status level through a local audible and visual alarm device and a human-machine interaction unit, ensuring that risks are known before entry. This solves the data barrier problem of existing systems, prevents blind entry into high-risk areas from the source, improves on-site perception capabilities and proactive prevention efficiency, and avoids decision-making delays caused by fragmented safety information.

[0028] 2. This invention utilizes an edge computing and data fusion unit to run an adaptive weighted fusion algorithm, combined with a personnel identity and permission recognition unit for multi-dimensional verification, and links with an emergency broadcasting and disaster evacuation guidance unit to achieve a closed-loop system encompassing risk perception, early warning, contingency plan execution, and on-site guidance. The system can automatically match the voice emergency plan and dynamically generate the optimal evacuation route map when an early warning is triggered, ensuring accurate and efficient emergency response. Simultaneously, the built-in LSTM model analyzes historical data to fine-tune thresholds, improving early warning accuracy and intelligence, and solving problems such as reliance on manual response and static route failure.

[0029] 3. This invention ensures the continuous operation of core functions during network interruptions through local storage and offline working units, and utilizes intrinsically safe design to adapt to harsh environments, achieving high reliability and longevity of the system. In addition, it deeply integrates multiple technology modules such as data fusion, edge computing, and identity recognition, and builds a complete solution around the alleyway entrance scenario, solving collaborative problems such as information gaps, response delays, and management deficiencies. Furthermore, it improves the accuracy of early warnings for specific working conditions by periodically iterating algorithm parameters through an adaptive optimization mechanism, forming a technological integration innovation that enhances the robustness and overall efficiency of the system. Attached Figure Description

[0030] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0031] Figure 1 This is a flowchart illustrating a multi-source sensing data fusion and intelligent early warning system for coal mine roadways according to the present invention.

[0032] Figure 2 This is a schematic diagram of the intelligent control terminal of the present invention. Detailed Implementation

[0033] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0034] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0035] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0036] Example 1:

[0037] like Figure 1 and Figure 2 As shown, the present invention provides a multi-source sensing data fusion and intelligent early warning system for coal mine roadways, which specifically includes an intrinsically safe intelligent control terminal for mining deployed at the entrance of the target roadway, a remote monitoring platform deployed at the ground dispatch center, and a mobile client accessible to authorized operators.

[0038] The following describes the closed-loop process of the system from normal operation to risk emergency response, using a typical operational scenario:

[0039] 1. System Initialization and Normal Operation

[0040] An intelligent control terminal is installed at the entrance of the tunnel, and the gas concentration sensor, roof pressure sensor, temperature sensor, wind speed sensor and micro-vibration monitoring probe arranged in the tunnel are connected to the terminal's sensing data aggregation unit through the industrial Ethernet and CAN bus interface.

[0041] After the terminal is initialized, it synchronizes the historical monitoring data, personnel information database and standard emergency plan database of the tunnel from the ground dispatch platform and caches them in local storage and offline work unit.

[0042] During normal system operation, the data collected by various sensors are all within the preset safety threshold range. At this time, the edge computing and data fusion unit within the terminal performs real-time fusion analysis of the multi-source data.

[0043] The unit first performs data preprocessing to filter out environmental noise; then, it uses an adaptive weighted fusion algorithm to assign dynamic weights to different sensor data (for example, gas concentration and roof pressure data have relatively high weights); finally, it compares the data with historical data from the same period and evaluates the dynamic safety status level as "green" (low safety level).

[0044] The human-computer interaction and display unit then displays the current safety status using red, yellow, and green status lights and a stable parameter curve.

[0045] The personnel identification and authorization unit (integrated RFID reader) is in standby mode. When authorized worker A, wearing a positioning card, approaches the entrance, the identification unit reads his / her identity information within 0.5 seconds and compares it with the locally cached safety training validity period and electronic work ticket authorization.

[0046] After successful comparison, the terminal controls the gate to automatically allow passage, while a brief welcome message and a summary of the current environmental parameters are displayed on the screen.

[0047] 2. Moderate Risk Warning and Control

[0048] Suppose that due to localized heating at the tunnel face, the temperature sensor in the roadway detects that the temperature has risen by 6°C within 5 minutes, approaching the preset warning threshold. At the same time, the gas sensor data shows that although the concentration has not exceeded the standard, it is showing a slow upward trend.

[0049] The sensing data aggregation unit transmits the aforementioned abnormal data to the edge computing and data fusion unit in real time.

[0050] The fusion unit initiates a hierarchical processing mechanism:

[0051] The first layer involves fusing multi-source data such as temperature, gas concentration, and wind speed using DS evidence theory to calculate the overall risk probability.

[0052] The second layer compares the current data pattern with similar warning event records in the local historical database, such as "localized warming accompanied by gas fluctuations".

[0053] The third layer assigns higher weights to temperature and gas concentration for preliminary processing according to preset rules. Through per-second calculations, the unit determines that the current situation has been upgraded to "yellow" (medium safety level).

[0054] The system immediately initiates a coordinated response:

[0055] (1) On-site control: The sound and light alarm device activates the yellow rotating warning light and intermittent prompt sound, and the human-machine interaction and display unit screen switches to display, highlighting the temperature and gas change trends on the key parameter curves, and marking abnormal areas in the form of a heat map.

[0056] (2) Personnel control: If maintenance personnel B (whose work permit is limited to the electromechanical maintenance area) attempts to enter, the identity and permission identification unit will immediately compare their permissions with the current "yellow" risk level (associated mining area).

[0057] Upon detecting a permission mismatch, the system automatically executes a preset policy: the gate remains closed;

[0058] The screen simultaneously displays a prominent graphic warning: "The current area is at increased risk. Your permissions are not valid. Entry is prohibited." At the same time, the information about this unauthorized approach and the personnel involved is synchronized to the ground dispatch platform via the local area network.

[0059] (3) Pre-set emergency plan: The emergency broadcast linkage and disaster avoidance guidance unit preloads the voice emergency plan text and basic disaster avoidance route map that match the risk type ("high temperature gas anomaly") into memory, ready to be broadcast at any time.

[0060] 3. High-risk emergency response and intelligent guidance

[0061] If temperature and gas concentration continue to deteriorate, existing coal mine safety monitoring systems mostly rely on static thresholds or simple rules for early warning, which cannot effectively predict sudden risk changes, resulting in delayed warnings or high false alarm rates. In contrast, this invention uses edge computing and data fusion units to analyze time-series data based on the LSTM (Long Short-Term Memory) model to predict the possibility of sudden changes in gas concentration in a short period of time, thereby dynamically correcting the safety status level to "red" (high safety level) and triggering gas or roof collapse early warnings.

[0062] The system has entered a mandatory emergency closed-loop state:

[0063] (1) Immediate response: The audible and visual alarm device activates a red flashing light and a high-decibel continuous alarm sound. The display unit switches to a dynamic disaster avoidance guidance interface in full screen.

[0064] (2) Mandatory personnel control and guidance: All personnel access requests are suspended. For personnel still near the entrance, the system obtains their precise location in real time through the UWB positioning module. The emergency broadcast linkage and disaster avoidance guidance unit combines the real-time location of the risk source (predicted gas accumulation point) with the personnel location, and uses AI algorithms to dynamically plan the optimal disaster avoidance route to the nearest refuge chamber, which is displayed on the screen as a highlighted arrow and dynamically updated.

[0065] At the same time, a clear and concise voice command is automatically broadcast: "Red Alert! All personnel, please immediately proceed to the designated safety exit or refuge chamber according to the mine's designated evacuation route!"

[0066] (3) Offline capability assurance: At this time, it is assumed that the local network downhole is interrupted. The local storage and offline working unit ensure that all the above-mentioned sensing, fusion, alarm, personnel identification and emergency guidance functions continue to operate independently, and the process data, operation logs and final handling results of the entire event are completely recorded locally.

[0067] (4) Post-event synchronization and optimization: After the network is restored, all data recorded locally is automatically synchronized to the ground dispatch platform. The data analysis module on the platform can use the full chain of data of this event to conduct a new round of supervised learning training on the early warning model (especially the gas mutation threshold and fusion algorithm weights), thereby completing the iterative optimization of the algorithm parameters and improving the accuracy of early warning for similar working conditions.

[0068] The above embodiments fully demonstrate the closed-loop working mode of this system, from data perception and aggregation → edge fusion analysis and dynamic situation generation → on-site hierarchical early warning → intelligent personnel identification and access control → automatic matching of emergency commands and dynamic disaster avoidance guidance → offline operation and data feedback optimization. This solution solves the problems of data isolation, delayed response, disconnected control, and static contingency plans in traditional systems, moving the risk prevention and control checkpoint forward to the roadway site, and realizing intelligent, adaptive, and strongly closed-loop safety early warning and emergency response.

[0069] Example 2:

[0070] Example of a data fusion early warning method based on multi-dimensional quantitative evaluation and adaptive feedback

[0071] This embodiment provides a method for multi-source sensing data fusion and intelligent early warning in coal mine roadways, the core steps of which correspond to those described in claim 1, including:

[0072] Deploy intelligent control terminals at the entrance of the alleyway and configure them to connect to multi-source sensors;

[0073] The sensing data aggregation unit collects various types of physical quantity monitoring data in real time.

[0074] Locally on the terminal, the monitoring data is fused and analyzed in real time through edge computing and data fusion unit to generate a dynamic security status level.

[0075] Based on the security status level, immediately activate the local sound, light, and display alarm devices on the terminal to respond on-site;

[0076] When personnel approach, their identity and permissions are simultaneously identified and compared in real time with the current security status level. On-site control actions are executed according to preset policies, and photos or videos are recorded.

[0077] When an early warning is triggered or an emergency instruction is received, the system will automatically link and execute the emergency plan via voice broadcast and the graphical disaster evacuation route guidance.

[0078] The locally generated operation logs and data are recorded to local storage and synchronized to the ground dispatch platform when network conditions permit.

[0079] This embodiment aims to refine the processes of "real-time fusion analysis" and "generating dynamic security situation levels" in the core steps, and reveal the specific implementation methods based on quantitative models and feedback mechanisms.

[0080] In this embodiment, the "edge computing and data fusion unit" and its built-in "early warning model" are composed of the following specific quantitative elements:

[0081] Multi-source monitoring vector (X): defined as X = [X1, X2, ..., X...] n ] T , where n is the number of sensor types connected. X i This represents the real-time normalized value of the i-th type of physical quantity, such as gas concentration, roof pressure, and wind speed as described in claim 2. The normalization method is X. i =(V i -V i,min ) / (V i,max -V i,min ), where V i V is the original measured value. i,min and V i,max This refers to the historical safe value range for this type of sensor in this roadway.

[0082] Dynamic weight vector (W): defined as W(t) = [W1(t), W2(t), ..., W... n [(t)], where W i (t) represents the fusion weight of the i-th class of data at time t, and This weight is dynamically adjusted by the adaptive optimization mechanism of the early warning model.

[0083] Security Situation Index (S): A scalar value, calculated by a fusion algorithm based on X and...

[0084] The comprehensive risk assessment quantification result obtained by W is in the range of [0,1].

[0085] Early warning threshold vector (Θ): defined as Θ=[θ green ,θ yellow ,θ red ], where 0 < θ green <θ yellow <θ red ≤1. Used to map continuous S values ​​to discrete security status levels (green, yellow, red).

[0086] Security situation index fusion calculation:

[0087]

[0088] S(t): Represents the security situation index at time t.

[0089] n: Represents the number of physical quantities to be fused, and is an integer.

[0090] W i (t): Represents the dynamic fusion weight of the i-th type of physical quantity data at time t.

[0091] X i (t): represents the preprocessed normalized monitoring value of the i-th type of physical quantity at time t.

[0092] f i (·): Represents the risk contribution function specific to the i-th type of physical quantity. This function is usually non-linear; for example, an exponential function can be used for gas concentration. (k is the gain coefficient) to amplify the risk contribution of high concentration areas; for wind speed, a parabolic function can be used to emphasize the optimal ventilation range.

[0093] After calculating S(t), it is immediately compared with the warning threshold vector Θ, triggering the corresponding safety level determination and on-site response:

[0094] If S(t) ≤ θ green If the status is positive, the system is classified as green (safe). The human-machine interface unit displays a green status light, indicating that the system is in normal monitoring mode.

[0095] If θ green <S(t)≤θ yellow If the risk level is high, it will be classified as Yellow (Caution). The system will automatically trigger a yellow audible and visual alarm on-site and highlight the trend of key risk parameters on the display unit.

[0096] If S(t)>θ yellow If the condition is detected, it will be classified as a red (dangerous) level. The system will immediately trigger a red audible and visual alarm and automatically activate the emergency broadcast linkage and disaster evacuation guidance unit to execute emergency plan broadcasts and dynamic route planning.

[0097] Initial intervention / adjustment: Dynamic weight adaptation

[0098] When the system remains at the yellow or red level for T consecutive periods (e.g., T=60 minutes), or when the platform issues a parameter adjustment command, the initial adaptive adjustment of the weight W(t) is initiated. The adjustment strategy is based on historical early warning effectiveness analysis.

[0099]

[0100] : The adjusted weight of the i-th type of data.

[0101] Weights of the i-th class of data before adjustment

[0102] α: Learning rate factor (recommended value range 0.01-0.1), controls the adjustment step size and prevents oscillation.

[0103] E i The contribution confidence of the i-th type of data in the most recent M effective warnings is calculated by analyzing the proportion of times that the mutation of this type of data precedes the mutation of the comprehensive index S in historical warning events.

[0104] The average confidence level of all data contributions.

[0105] The process of generating a "multi-source fusion risk trend curve" is broken down into the following steps:

[0106] Data preparation: Extract the historical security situation index sequence {S(tL), ..., S(t-1), S(t)} from the local storage unit for the past L time windows (e.g., the past 24 hours, with a window interval of 5 minutes).

[0107] Intermediate value calculation: Perform first-order difference calculation on the above sequence to obtain the risk change rate sequence {ΔS(t-L+1), ..., ΔS(t)}, where ΔS(t) = S(t) - S(t-1).

[0108] Graphical generation: Two superimposed curves are plotted on the display screen: the primary axis plots the risk index curve of S(t) over time; the secondary axis plots the rate of change curve of ΔS(t) over time. When ΔS(t) exceeds θ... green When the curve color gradually changes from green to yellow and then to red, when ΔS(t) > 0, the rate of change curve is marked with an upward arrow, which visually shows the accelerating trend of risk.

[0109] To evaluate the effectiveness of the aforementioned fusion early warning model under current operating conditions, a new dynamic evaluation index is defined—the model real-time effectiveness V(t).

[0110]

[0111] V(t): represents the model validity at time t, with a range of [-1, 1].

[0112] N: Dynamically evaluate the window size, such as the most recent 100 data points.

[0113] γ: Balance weighting factor (0 < γ < 1), used to balance the timeliness of level triggering and the sensitivity of the rate of change.

[0114] sign(·): The sign function, which is used when S(k) > θ. yellow If the assessment is timely, add 1; otherwise, subtract 1 to determine whether the assessment level was determined promptly.

[0115] |ΔS(k)|: The absolute value of the rate of change of risk.

[0116] σ: Standardized parameter for rate of change.

[0117] tanh(·): Hyperbolic tangent function, used to map the rate of change to the interval [-1,1] to assess whether the risk development trend has been captured.

[0118] Based on the secondary evaluation index V(t), establish a warning threshold θ yellow Closed-loop feedback optimization control forms an adaptive adjustment mechanism:

[0119]

[0120] Adjusted yellow warning threshold.

[0121] The yellow warning threshold before adjustment.

[0122] β: Adjust the gain coefficient (recommended range 0.005-0.02), which determines the magnitude of each adjustment.

[0123] V ref The validity reference benchmark is set to a value slightly greater than 0 (e.g., 0.2), representing the balance between warning sensitivity and accuracy that the system expects to maintain.

[0124] When V(t) > V ref This indicates that the model is currently too "sensitive" (potentially generating many false alarms), and the system will adjust θ. yellow The threshold for early warning should be raised appropriately, and the adjustment range should be adjusted in accordance with (V(t) - V). ref It is directly proportional to .

[0125] When V(t) < V ref This indicates that the model is currently "sluggish" (potentially missing a risk), and the system will use θ. yellow Appropriately lower the threshold for early warning, and adjust the range accordingly (V). ref -V(t) is directly proportional.

[0126] This closed-loop mechanism enables the early warning system to dynamically optimize its core judgment parameters based on its recent operational performance, thereby achieving continuous self-optimization.

[0127] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the flow or function according to the embodiments of the present invention is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via a wired network and / or wireless network. A computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. Available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media. Semiconductor media can be solid-state drives (SSDs).

[0128] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0129] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0130] In the several embodiments provided by this invention, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only one method, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0131] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0132] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0133] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0134] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A multi-source sensing data fusion and intelligent early warning system for coal mine roadways, characterized in that, This includes intrinsically safe intelligent control terminals deployed at the entrances of tunnels, as well as ground dispatch platforms and mobile clients; The intelligent control terminal is deployed at the entrance of the alleyway and configured to connect to multi-source sensors. The terminal collects various types of physical quantity monitoring data in real time through the sensing data aggregation unit; Locally on the terminal, the monitoring data is fused and analyzed in real time through edge computing and data fusion unit to generate a dynamic security status level. Based on the security status level, immediately activate the local sound, light, and display alarm devices on the terminal to respond on-site; When personnel approach, the system simultaneously identifies their identity and permissions, compares this information with the current security status level in real time, and executes on-site control actions based on preset strategies. When an alert is triggered or an emergency command is received, the system automatically links to execute voice broadcasts of emergency plans and graphical evacuation route guidance. Furthermore, it can take photos of the operators during the operation to record the operation, so as to leave a record of the work. It can also take photos when the infrared sensor gets close to the operator. The locally generated operation logs and data are recorded to local storage and synchronized to the ground dispatch platform when network conditions permit.

2. The system according to claim 1, characterized in that, The monitored physical quantities include at least two or more of the following: gas concentration, roof pressure, wind speed, equipment operating status, personnel positioning information, oxygen concentration, water pressure and temperature for exploration and drainage, micro-vibration, ground sound, and drainage volume; the sensing data aggregation unit accesses the data through one or more of the following: industrial Ethernet, RS485, CAN bus, or wireless communication interface.

3. The system according to claim 1, characterized in that, The early warning model built into the edge computing and data fusion unit has an adaptive optimization mechanism, which can fine-tune the early warning threshold and fusion algorithm weight parameters in real time based on the historical monitoring data of the roadway, the early warning history record, and the parameter adjustment instructions issued by the external platform. The early warning model uses a deep neural network model to optimize the weight allocation and uses an LSTM model to analyze time series data to dynamically correct the gas mutation threshold.

4. The system according to claim 1, characterized in that, The data fusion algorithm includes, but is not limited to, DS evidence theory or adaptive weighted fusion algorithm. Its output is a dynamic security status level, which is graphically displayed on the screen of the human-computer interaction and display unit in the form of red, yellow and green status lights and key parameter curves.

5. The system according to claim 1, characterized in that, The personnel identification and authorization unit integrates RFID and UWB card reading modules. It synchronously obtains personnel identity information, safety training validity period and current electronic work ticket with the ground dispatch platform or through local caching. Combined with the comprehensive security situation level generated in real time by the edge computing and data fusion unit, it automatically executes control actions such as release, warning or prohibition.

6. The system according to claim 1, characterized in that, The emergency broadcast linkage and disaster evacuation guidance unit is configured as follows: Receive and parse structured emergency instruction packets from ground dispatch platforms or emergency broadcast systems, wherein the emergency instruction packets contain at least the warning level, the affected area, and the recommended action instructions; When a local warning is triggered or the emergency instruction package is received, the system automatically matches and broadcasts the pre-stored voice emergency plan content that corresponds to the current risk type and location. Based on real-time personnel location information and risk source locations, the mine's disaster avoidance route map leading to designated safety exits or refuge chambers is dynamically generated and highlighted on the screen of the human-computer interaction and display unit.

7. The system according to claim 1, characterized in that, The local storage and offline working unit ensures that the core functions of the sensing data aggregation unit, edge computing and data fusion unit, personnel identity and permission recognition unit, and emergency broadcast linkage and disaster avoidance guidance unit can continue to operate independently when network communication is interrupted, and data synchronization will be performed after the network is restored.

8. The system according to claim 1, characterized in that, The intelligent control terminal includes: a sensing data aggregation unit, an edge computing and data fusion unit, a human-computer interaction and display unit, a personnel identity and permission recognition unit, an emergency broadcast linkage and disaster avoidance guidance unit, and a local storage and offline working unit; The data aggregation unit receives physical quantity monitoring data in real time; the processing and fusion unit runs the fusion algorithm to generate security assessment results; the human-computer interaction unit displays the assessment results; and the identity recognition unit verifies personnel information.

9. The system according to claim 5, characterized in that, The control actions include: issuing voice and graphic warnings based on the safety situation level when the personnel's work permit area does not match the risk level or when safety training has expired, and sending alarm information to the remote platform.

10. The system according to claim 3, characterized in that, It also includes algorithm iteration and optimization: the processing and fusion unit periodically uses local historical data from the roadway to adjust algorithm parameters through supervised learning training, thereby improving the accuracy of early warnings for specific working conditions.

Images