Firefighter fire-fighting protective clothing position tracking and posture recognition method and system

Abstract

The application provides a firefighter fire extinguishing protective clothing position tracking and posture recognition method and system, comprising the following steps: multi-source sensor data acquisition, adaptive multi-source fusion positioning, mode switching and error constraint correction, posture solution and dynamic compensation, combat action type identification, data sending and safety warning. The method provided by the application realizes continuous high-precision positioning of the whole scene of the fire scene and accurate real-time perception of the whole body posture of the firefighter by constructing an adaptive multi-source fusion positioning model containing an extended Kalman filter and a mode switching strategy, and simultaneously realizes deep fusion and collaborative optimization of positioning, posture perception, action recognition and safety warning.

Description

Technical Field

[0001] This invention relates to the field of fire protection and intelligent sensing technology, specifically to a method and system for tracking the position and recognizing the posture of firefighters' firefighting protective clothing. Background Technology

[0002] Firefighters face multiple dangers during firefighting and rescue operations, including high temperatures, dense smoke, toxic gases, and building collapses. Real-time monitoring of firefighters' positions and postures is crucial for ensuring their safety and optimizing tactical deployment. However, the harsh environment of a fire scene, characterized by high temperatures, high humidity, heavy smoke, and signal obstruction, presents significant challenges to firefighter positioning and posture monitoring.

[0003] In existing technologies, the Global Positioning System (GPS) is unusable indoors and underground, Wi-Fi / Bluetooth positioning has low accuracy and is easily affected by environmental interference, and video surveillance is completely ineffective in dense smoke environments. Inertial navigation, while unaffected by the external environment, suffers from drift errors that accumulate over time, making it difficult to operate independently for extended periods. Ultra-wideband (UWB) positioning offers higher accuracy, but metal structures, wall obstructions, and multipath effects in fire scenes can lead to frequent signal loss.

[0004] Furthermore, existing technologies lack real-time sensing methods for the full-body posture of firefighters. While traditional inertial motion capture systems can acquire posture data, they do not consider the physical constraints of protective clothing on joint movement, leading to posture estimations that deviate from reality. Additionally, the functional modules of existing systems operate independently, failing to achieve deep integration and collaborative optimization of positioning, posture perception, motion recognition, and safety warning. Summary of the Invention

[0005] The main objective of this invention is to overcome the aforementioned deficiencies in the prior art. This invention provides a method for tracking the position and attitude recognition of firefighters' firefighting protective clothing. By constructing an adaptive multi-source fusion positioning model that includes an extended Kalman filter and a mode switching strategy, it achieves continuous high-precision positioning of the entire fire scene and accurate real-time perception of the firefighter's full-body posture. At the same time, it realizes deep integration and collaborative optimization of positioning, attitude perception, action recognition, and safety early warning.

[0006] This invention provides a method for tracking the position and posture of firefighters' firefighting protective clothing, comprising the following steps: S100: Acquire sensor data collected by the multi-source sensor components embedded in the protective suit body; S200: Construct an adaptive multi-source fusion localization model, which includes an extended Kalman filter framework and a mode switching strategy; S300: The acceleration and angular velocity data collected by the inertial measurement unit are input into the extended Kalman filter for dead reckoning. At the same time, the ranging information between the ultra-wideband positioning tag and multiple ultra-wideband base stations deployed inside the fire scene, as well as the air pressure data collected by the barometric altimeter, are input into the extended Kalman filter. The acceleration data, the angular velocity data, the ultra-wideband ranging information, and the air pressure data are fused after time synchronization and spatial alignment to generate the three-dimensional position information of the firefighter in the fire scene. S400: Real-time monitoring of the signal quality between the ultra-wideband positioning tag and the ultra-wideband base station. When the signal quality is lower than a preset threshold, the mode switching strategy switches the positioning mode from ultra-wideband assisted inertial positioning mode to pure inertial dead reckoning mode, and constrains and corrects the cumulative error of dead reckoning based on the firefighter's historical motion trajectory data and the bending data collected by the flexible bending sensor. S500: Based on the bending data collected by the flexible bending sensor, construct the bending change diagram of each flexible bending sensor, and combine the acceleration data and the angular velocity data to generate the firefighter's whole-body posture data through the attitude calculation algorithm; S600: Based on the full-body posture data, perform dynamic time warping and matching with a preset standard combat action library to identify the firefighter's current combat action type; S700: The three-dimensional position information, the whole-body posture data, and the current combat action type are transmitted to the remote command terminal via the wireless communication module.

[0007] This invention also provides a firefighter's firefighting protective clothing position tracking and posture recognition system, comprising: The data acquisition unit is used for acceleration and angular velocity data, bending data, ranging information between the ultra-wideband positioning tag and multiple ultra-wideband base stations deployed inside the fire scene, and air pressure data. An adaptive multi-source fusion localization unit is used to construct an adaptive multi-source fusion localization model that includes an extended Kalman filter framework and a mode switching strategy. The three-dimensional position information acquisition unit is used to input the acceleration data and the angular velocity data into the extended Kalman filter for dead reckoning, and simultaneously input the ranging information and the air pressure data into the extended Kalman filter. The unit performs time synchronization and spatial alignment of the acceleration data, the angular velocity data, the ranging information and the air pressure data into multi-source data fusion to generate the three-dimensional position information of the firefighter in the fire scene. The mode switching and error constraint correction unit is used to monitor the signal quality between the ultra-wideband positioning tag and the ultra-wideband base station in real time. When the signal quality is lower than a preset threshold, the mode switching strategy switches the positioning mode from ultra-wideband assisted inertial positioning mode to pure inertial dead reckoning mode, and constrains and corrects the cumulative error of dead reckoning based on the firefighter's historical motion trajectory data and the bending data collected by the flexible bending sensor. The attitude calculation unit is used to construct a bending change map of each flexible bending sensor based on the bending data, and generate the firefighter's whole-body attitude data by combining the acceleration data and the angular velocity data through the attitude calculation algorithm. The motion recognition unit is used to perform dynamic time-adjusted matching between the full-body posture data and a preset standard combat motion library to identify the current combat motion type of the firefighter. The wireless communication module is used to send the three-dimensional position information, the whole-body posture data, and the current combat action type to the remote command terminal.

[0008] This invention also provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement a method for tracking the position and attitude of a firefighter's firefighting protective clothing.

[0009] This invention provides another computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of a method for tracking the position and attitude of a firefighter's firefighting protective suit.

[0010] As can be seen from the above description of the present invention, compared with the prior art, the present invention has the following beneficial effects: This invention proposes a method for tracking the position and attitude recognition of firefighters' firefighting protective suits, comprising the following steps: multi-source sensor data acquisition, adaptive multi-source fusion positioning, mode switching and error constraint correction, attitude calculation and dynamic compensation, combat action type recognition, data transmission and safety warning. The method provided by this invention achieves continuous high-precision positioning across the entire fire scene and accurate real-time perception of the firefighter's full-body attitude by constructing an adaptive multi-source fusion positioning model that includes an extended Kalman filter and a mode switching strategy. Simultaneously, it achieves deep fusion and collaborative optimization of positioning, attitude perception, action recognition, and safety warning. Attached Figure Description

[0011] Figure 1 A flowchart of a method for tracking the position and posture of a firefighter's firefighting protective suit, provided in an embodiment of the present invention; Figure 2A structural diagram of a firefighter's firefighting protective suit position tracking and posture recognition system provided in an embodiment of the present invention; Figure 3 A schematic diagram illustrating an embodiment of an electronic device provided by an embodiment of the present invention; Figure 4 This is a schematic diagram of an embodiment of a computer-readable storage medium provided in this invention. Detailed Implementation

[0012] The preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0013] In this embodiment, the protective suit body adopts a multi-layer composite structure. From the outside to the inside, the protective suit body includes at least a fire-retardant coating, an outer layer, a first thermal insulation and fire-resistant layer, a second thermal insulation and fire-resistant layer, and an inner layer. The first thermal insulation and fire-resistant layer contains multiple airbags evenly embedded within it, with adjacent airbags connected by air guide tubes. One end of the air guide tube is closed, and the other end connects to the inside of an inflation nozzle on the protective suit body. The second thermal insulation and fire-resistant layer contains a serpentine liquid guide tube embedded within it, with one end connecting to the inside of an injection valve and the other end connecting to the inside of an outlet valve. Separating layers are provided between the layers to effectively separate them, allowing each layer to perform its respective function while improving the overall integrity of the suit.

[0014] The multi-source sensor assembly includes multiple inertial measurement units located at the corresponding positions of the firefighter's major joints, multiple flexible bending sensors located at the corresponding positions of the firefighter's limb segments, an ultra-wideband positioning tag located on the protective suit body, and a barometric altimeter located on the protective suit body.

[0015] The sensor assembly is embedded between the first and second thermal insulation and fireproof layers. To prevent the sensors from being compressed after the airbags are inflated, the inertial measurement unit and the flexible bending sensor are located in the gap between adjacent airbags.

[0016] The sensor assembly includes: multiple inertial measurement units (IMUs) positioned at corresponding locations on the firefighter's major joints (shoulder, elbow, wrist, hip, knee, ankle, etc.); multiple flexible bending sensors positioned at corresponding locations on the firefighter's limbs (upper arm, forearm, thigh, calf, torso, etc.); an ultra-wideband (UWB) positioning tag positioned on the shoulder area of ​​the protective suit; a barometric altimeter mounted on the protective suit; and ambient temperature and humidity sensors mounted on the protective suit. Each sensor is communicatively connected to the signal processing module via flexible wiring.

[0017] The signal processing module, located on the back of the protective suit, includes a microprocessor, a data acquisition unit, a wireless communication module, a storage unit, and a power management unit. The wireless communication module supports 4G / 5G mobile communication and dedicated Mesh network communication, used to transmit processed data to a remote command center. The power management unit is electrically connected to a replaceable battery located inside the protective suit, providing power to the various sensors and the signal processing module.

[0018] like Figure 1 This is a flowchart of a method for tracking the position and posture of firefighters' firefighting protective clothing, specifically including: Step S100: Obtain sensing data collected by a multi-source sensor assembly embedded in the protective suit body. The multi-source sensor assembly includes multiple inertial measurement units set at the corresponding positions of the firefighter's major joints, multiple flexible bending sensors set at the corresponding positions of the firefighter's limb segments, an ultra-wideband positioning tag set on the protective suit body, and a barometric altimeter set on the protective suit body. In step S100, the sensing data collected by the multi-source sensor components embedded in the protective suit body is obtained.

[0019] The inertial measurement unit uses a six-axis IMU chip, integrating a three-axis accelerometer and a three-axis gyroscope. The sampling frequency is set to 100Hz, and the ranges are set to ±16g and ±2000° / s, respectively, to cover the violent movement scenarios that firefighters may experience in a fire.

[0020] The flexible bending sensor employs a long, flexible sensing element based on a conductive composite material, whose resistance changes monotonically with the bending angle. Each flexible bending sensor is installed at a corresponding location on the firefighter's upper arm, forearm, thigh, lower leg, and torso to collect bending data for each limb segment. The sensor length varies from 5-15 cm, selected to fit the length of the installed limb segment. The sampling frequency is set to 100 Hz and synchronized with the IMU.

[0021] The UWB positioning tags use the Decawave DW1000 chip solution, operating in the 3.5GHz-6.5GHz frequency band, with a communication distance of no less than 100 meters and a positioning accuracy of 10-30cm. Multiple UWB base stations are deployed inside the fire scene with known relative positions, and the base stations synchronize their clocks via wired or wireless means.

[0022] The barometric altimeter uses the MS5611 high-precision barometric sensor with a resolution of up to 10cm and a sampling frequency of 10Hz to assist in measuring the floor height information of firefighters.

[0023] An ambient temperature sensor and an ambient humidity sensor are placed between the outer layer of the protective suit and the first thermal insulation and fireproof layer, with a sampling frequency of 1Hz.

[0024] Steps S200-S300: S200: Construct an adaptive multi-source fusion localization model, which includes an extended Kalman filter framework and a mode switching strategy; S300: The acceleration and angular velocity data collected by the inertial measurement unit are input into the extended Kalman filter for dead reckoning. At the same time, the ranging information between the ultra-wideband positioning tag and multiple ultra-wideband base stations deployed inside the fire scene, as well as the air pressure data collected by the barometric altimeter, are input into the extended Kalman filter. The multi-source data fusion after time synchronization and spatial alignment of the acceleration data, the angular velocity data, the ultra-wideband ranging information, and the air pressure data generates the three-dimensional position information of the firefighter in the fire scene. In step S2, an adaptive multi-source fusion positioning model is constructed, including the extended Kalman filter framework and mode switching strategy.

[0025] Time synchronization and spatial alignment: Before performing multi-source data fusion, the data from each sensor are first synchronized in time and aligned in space. For time synchronization, the timestamps of the UWB ranging information are interpolated and aligned with the sampling time of the IMU. Since the data update rate of UWB (approximately 10-20Hz) is lower than that of the IMU (100Hz), a linear interpolation method is used to synchronize the UWB ranging information to the sampling time of the IMU.

[0026] In terms of spatial alignment, a rotation matrix transformation is performed between the global coordinate system of the UWB base station and the carrier coordinate system of the IMU to unify the data of each sensor under the same spatiotemporal reference.

[0027] Extended Kalman filter fusion localization: The acceleration and angular velocity data collected by the IMU are input into the extended Kalman filter (EKF) for dead reckoning. At the same time, the UWB ranging information and the barometric pressure data collected by the barometric altimeter are input into the EKF for fusion correction.

[0028] The state vector of EKF is defined as: x = [p^T, v^T, q^T, b_a^T, b_g^T]^T; Where p is the three-dimensional position vector, v is the velocity vector, q is the attitude quaternion, and b_a and b_g are the zero biases of the accelerometer and gyroscope, respectively.

[0029] In the prediction step, the state vector is updated over time based on the acceleration and angular velocity data from the IMU. In the update step, when UWB ranging information is available, the state vector is corrected using the distance between the UWB base station and the positioning tag as the observation value; simultaneously, the altitude component is corrected using the barometric altimeter's barometric altitude as the observation value.

[0030] Step S400: Monitor the signal quality between the ultra-wideband positioning tag and the ultra-wideband base station in real time. When the signal quality is lower than a preset threshold, the mode switching strategy switches the positioning mode from ultra-wideband assisted inertial positioning mode to pure inertial dead reckoning mode, and constrains and corrects the cumulative error of dead reckoning based on the firefighter's historical motion trajectory data and the bending data collected by the flexible bending sensor.

[0031] In step S400, the signal quality between the UWB positioning tag and the UWB base station is monitored in real time. The signal quality is comprehensively evaluated by the Received Signal Strength Indication (RSSI) and the Signal-to-Noise Ratio (SNR). When the RSSI is lower than a preset threshold (e.g., -85dBm) or the SNR is lower than a preset threshold (e.g., 10dB), the signal quality is determined to be lower than the preset threshold.

[0032] When the signal quality falls below a preset threshold, the mode switching strategy switches the positioning mode from ultra-wideband assisted inertial positioning mode to pure inertial dead reckoning mode. In pure inertial dead reckoning mode, position is recursively calculated using only IMU data, while the cumulative error of dead reckoning is constrained and corrected based on the firefighter's historical motion trajectory data and bending data collected by flexible bending sensors.

[0033] Constraint correction includes the following two aspects: First, a gait cycle model of firefighters is established based on historical movement trajectory data. The gait phase of a firefighter is identified by the bending change map of a flexible bending sensor—when the bending data from the flexible bending sensor located on the lower leg shows periodic changes, and a specific trough characteristic appears in the curvature change curve, it is determined to be the foot contact phase. When the foot contact phase is identified, the velocity of the IMU is zeroed, and zero-velocity correction (ZUPT) is performed on the three-dimensional position information.

[0034] Second, the movement patterns of firefighters' limbs are identified based on the bending change maps of flexible bending sensors. When the bending sensor data of each limb segment presents a specific combination pattern, it is determined that the firefighter is in a non-walking movement mode such as crawling or climbing, and the corresponding motion constraint model is switched to constrain the movement mode.

[0035] Step S500: Based on the bending data collected by the flexible bending sensors, construct bending change diagrams for each flexible bending sensor. Combine the acceleration data and the angular velocity data, and generate the firefighter's full-body posture data using an attitude calculation algorithm. Attitude calculation algorithm: The bending change graphs of each flexible bending sensor are constructed based on the bending data collected by the flexible bending sensors. These graphs typically refer to calibration curves or response characteristic curves. They visually demonstrate how the physical output signal (such as resistance or voltage) of the sensor changes with the degree of bending (such as angle or curvature) when the sensor is bent. In this embodiment, the raw resistance values ​​collected by each flexible bending sensor are filtered (using a low-pass filter with a cutoff frequency of 10Hz), and then the resistance values ​​are converted into bending angle values ​​according to a pre-calibrated resistance-bending angle mapping relationship.

[0036] The attitude calculation algorithm uses gradient descent to fuse acceleration and angular velocity data for attitude quaternion updates. Based on this, using the bending change map from a flexible bending sensor as joint angle observations, an attitude observation equation is constructed. Kalman filtering is then applied to correct the attitude quaternion update results, outputting the corrected full-body attitude data.

[0037] Dynamic compensation steps: Obtain the constraint parameter matrix L = [l1, l2, ..., ln]^T for the range of motion of each joint of the protective suit, where n is the number of joints, and each l_i represents the maximum movable angle vector of the i-th joint in each direction of movement. The constraint parameter matrix is ​​predetermined based on the structural dimensions, fabric properties, and airbag inflation status of the protective suit.

[0038] The method for determining the constraint parameter matrix is ​​as follows: First, with the protective suit in its uninflated state, the maximum joint range of motion of the test personnel wearing the protective suit in each direction is measured using a 3D motion capture system to establish the basic constraint parameter matrix. Then, a mapping relationship between airbag inflation pressure and joint range of motion is established—the greater the airbag inflation pressure, the stronger the constraint of the protective suit on the joints, and the smaller the maximum range of motion.

[0039] The whole-body posture data is constrained and corrected based on the constraint parameter matrix. When the posture angle of a certain joint exceeds the maximum movable angle of that joint in the current direction of movement, the posture angle is truncated to the maximum movable angle, and a protective suit constraint compensation signal is generated.

[0040] Step S600: Perform dynamic time warping matching based on the full-body posture data and the preset standard combat action library to identify the current combat action type of the firefighter; In step S600, dynamic time warping (DTW) matching is performed based on the full-body posture data and the preset standard combat action library to identify the firefighter's current combat action type.

[0041] The standard operational maneuver library includes at least climbing, valve closing, decontamination, carrying, and crawling maneuvers. Each maneuver template consists of multiple key attitude frame sequences.

[0042] The specific method of dynamic time warping matching is as follows: Align the real-time full-body posture data sequence X = [x1, x2, ..., xm] with the action template sequence Y = [y1, y2, ..., yn], and find the optimal alignment path through dynamic programming to minimize the cumulative distance between the two sequences. Calculate the DTW distance between the full-body posture data and each action template, and identify the action type corresponding to the action template with the smallest distance as the current combat action type.

[0043] Step S700: The three-dimensional position information, the whole-body posture data, and the current combat action type are sent to the remote command terminal via the wireless communication module.

[0044] The three-dimensional position information p = [x, y, z]^T, the whole-body attitude data, and the current combat action type are transmitted to the remote command terminal via a wireless communication module. The data transmission frequency is set to 1Hz (position and action type) and 10Hz (attitude data).

[0045] Safety warning steps include: Acquire ambient temperature and humidity data. Determine the spatial distance D between the firefighter and the fire source based on 3D location information. Determine the firefighter's current activity intensity level using a metabolic equivalent algorithm based on full-body posture data. Determine the heat load level using a heat index algorithm based on ambient temperature and humidity.

[0046] Based on spatial distance D, activity intensity level, and heat load level, the real-time safety factor S is calculated according to the preset weighted fusion rules; The tiered alarm mechanism includes three levels of alarms: a level one warning is triggered when S is below the first safety threshold but above the second safety threshold; a level two emergency alarm is triggered when S is below the second safety threshold, with different frequencies and intensities of audible and visual signals.

[0047] The virtual-to-real mapping steps include: Acquire Building Information Modeling (BIM) and video surveillance data from the fire scene to construct a 3D spatial model of the fire scene. Based on 3D position information and full-body posture data, drive a virtual avatar of the firefighter in real time within the 3D spatial model, with a refresh rate of no less than 20Hz.

[0048] Based on three-dimensional position information and full-body posture data, combined with the firefighter's body shape data, a dynamic human body model of the firefighter wearing protective clothing is generated in a three-dimensional space model to simulate the changes in the range of motion of each joint and the distribution of the thickness of the air layer under the clothing.

[0049] Based on the current combat action type and three-dimensional position information, the movement trajectory of firefighters within a future preset time window is predicted in a three-dimensional spatial model, and the predicted trajectory is displayed as a phantom.

[0050] Specific examples Before entering the fire scene, firefighters don fire-fighting protective suits integrated with multi-source sensor components. The signal processing module is powered on and activated, and the sensors begin collecting data. The UWB positioning tag establishes communication connections with multiple UWB base stations deployed inside the fire scene, completing the initial positioning.

[0051] After firefighters entered the fire scene, the IMU collected data from each joint at 100Hz, the flexible bending sensor collected data from each limb segment at 100Hz, the UWB acquired distance measurement information at 10-20Hz, and the barometric altimeter collected air pressure data at 10Hz.

[0052] The EKF in the signal processing module performs state prediction and updates at 100Hz. In areas with good UWB signal, positioning accuracy can reach within 30cm. When the UWB signal is blocked, the mode switching strategy automatically switches to pure inertial dead reckoning mode, and uses a flexible bending sensor for zero-velocity correction and motion constraints.

[0053] Meanwhile, the attitude calculation algorithm generates full-body attitude data at 100Hz, and the dynamic compensation step performs constraint correction based on the airbag inflation status. The combat action recognition module identifies the current action type at 1Hz.

[0054] The aforementioned data is transmitted in real time to the remote command terminal via a wireless communication module. The remote command terminal then drives the virtual avatar in real time within the 3D spatial model. The safety early warning module continuously calculates the safety factor and triggers tiered alarms. Data reconstruction is automatically performed in the event of sensor failure.

[0055] like Figure 2 A firefighter's firefighting protective suit position tracking and posture recognition system, comprising: The data acquisition unit 210 is used to acquire acceleration data and angular velocity data collected by the multiple inertial measurement units, bending data collected by the multiple flexible bending sensors, ranging information between the ultra-wideband positioning tag and multiple ultra-wideband base stations deployed inside the fire scene, and air pressure data collected by the barometric altimeter. The adaptive multi-source fusion localization unit 220 is used to construct an adaptive multi-source fusion localization model that includes an extended Kalman filter framework and a mode switching strategy. The three-dimensional position information acquisition unit 230 is used to input the acceleration data and the angular velocity data into the extended Kalman filter for dead reckoning, and simultaneously input the ranging information and the air pressure data into the extended Kalman filter. After time synchronization and spatial alignment of the acceleration data, the angular velocity data, the ranging information and the air pressure data, the multi-source data are fused to generate the three-dimensional position information of the firefighter in the fire scene. The mode switching and error constraint correction unit 240 is used to monitor the signal quality between the ultra-wideband positioning tag and the ultra-wideband base station in real time. When the signal quality is lower than a preset threshold, the mode switching strategy switches the positioning mode from ultra-wideband assisted inertial positioning mode to pure inertial dead reckoning mode, and constrains and corrects the cumulative error of dead reckoning based on the historical motion trajectory data of the firefighter and the bending data collected by the flexible bending sensor. The attitude calculation unit 250 is used to construct a bending change map of each flexible bending sensor based on the bending data, and generate the firefighter's whole-body attitude data by combining the acceleration data and the angular velocity data through the attitude calculation algorithm. The motion recognition unit 260 is used to perform dynamic time-adjusted matching between the full-body posture data and a preset standard combat motion library to identify the current combat motion type of the firefighter. The wireless communication module 270 is used to send the three-dimensional position information, the whole-body posture data and the current combat action type to the remote command terminal.

[0056] In another aspect, the present invention provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement a method for infrastructure security analysis based on a large model.

[0057] like Figure 3 As shown, an embodiment of the present invention provides an electronic device 300, including a memory 310, a processor 320, and a computer program 311 stored in the memory 310 and executable on the processor 320. When the processor 320 executes the computer program 311, it implements an infrastructure security analysis method based on a large model provided by the embodiment of the present invention.

[0058] In practical implementation, when the processor 320 executes the computer program 311, it can achieve... Figure 1 Any of the corresponding implementation methods in the embodiments.

[0059] Since the electronic device described in this embodiment is a device used to implement a data processing device in the embodiment of the present invention, those skilled in the art can understand the specific implementation method and various variations of the electronic device in this embodiment based on the method described in the embodiment of the present invention. Therefore, how the electronic device implements the method in the embodiment of the present invention will not be described in detail here. Any device used by those skilled in the art to implement the method in the embodiment of the present invention is within the scope of protection of the present invention.

[0060] Please see Figure 4 , Figure 4 This is a schematic diagram of an embodiment of a computer-readable storage medium provided in this invention.

[0061] like Figure 4 As shown, this embodiment provides a computer-readable storage medium 400 on which a computer program 411 is stored. When the computer program 411 is executed by a processor, it implements an infrastructure security analysis method based on a large model provided in this embodiment of the invention. In practical implementation, when the computer program 411 is executed by the processor, it can achieve the following: Figure 1 Any of the corresponding implementation methods in the embodiments.

[0062] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0063] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0064] This invention proposes a method for tracking the position and attitude recognition of firefighters' firefighting protective suits, comprising the following steps: multi-source sensor data acquisition, adaptive multi-source fusion positioning, mode switching and error constraint correction, attitude calculation and dynamic compensation, combat action type recognition, data transmission and safety warning. The method provided by this invention achieves continuous high-precision positioning across the entire fire scene and accurate real-time perception of the firefighter's full-body attitude by constructing an adaptive multi-source fusion positioning model that includes an extended Kalman filter and a mode switching strategy. Simultaneously, it achieves deep fusion and collaborative optimization of positioning, attitude perception, action recognition, and safety warning.

[0065] The above are merely specific embodiments of the present invention, but the design concept of the present invention is not limited thereto. Any non-substantial modifications made to the present invention using this concept shall be considered as infringing upon the protection scope of the present invention.

Claims

1. A method for tracking the position and recognizing the posture of firefighters' firefighting protective clothing, characterized in that, Includes the following steps: S1: Acquire sensor data collected by the multi-source sensor components embedded in the protective suit body; S2: Construct an adaptive multi-source fusion localization model, which includes an extended Kalman filter framework and a mode switching strategy; S3: The acquired sensor data, including acceleration and angular velocity data, is input into the extended Kalman filter for dead reckoning. Simultaneously, the ranging information between the ultra-wideband positioning tag and multiple ultra-wideband base stations deployed inside the fire scene, as well as the acquired sensor data, including air pressure data, are input into the extended Kalman filter. The multi-source data, including acceleration data, angular velocity data, ultra-wideband ranging information, and air pressure data, is fused after time synchronization and spatial alignment to generate the three-dimensional position information of the firefighter in the fire scene. S4: Monitor the signal quality between the ultra-wideband positioning tag and the ultra-wideband base station in real time. When the signal quality is lower than a preset threshold, the mode switching strategy switches the positioning mode from ultra-wideband assisted inertial positioning mode to pure inertial dead reckoning mode, and constrains and corrects the cumulative error of dead reckoning based on the firefighter's historical motion trajectory data and the bending data collected by the flexible bending sensor. S5: Based on the bending data collected by the flexible bending sensor, construct the bending change diagram of each flexible bending sensor, and combine the acceleration data and the angular velocity data to generate the firefighter's whole-body posture data through the attitude calculation algorithm; S6: Based on the full-body posture data, perform dynamic time warping and matching with the preset standard combat action library to identify the firefighter's current combat action type; S7: The three-dimensional position information, the whole-body posture data, and the current combat action type are sent to the remote command terminal via the wireless communication module.

2. The method according to claim 1, characterized in that, In step S4, the cumulative error of dead reckoning is constrained and corrected based on historical motion trajectory data and bending data. Specifically, this includes: establishing a gait cycle model of the firefighter based on the historical motion trajectory data; identifying the firefighter's gait phase through the bending change diagram of the flexible bending sensor; and when the foot touches the ground phase, reducing the velocity of the inertial measurement unit to zero and performing zero-velocity correction on the three-dimensional position information.

3. The method according to claim 1, characterized in that, The attitude calculation algorithm in step S5 includes: using gradient descent to update the attitude quaternion of the acceleration data and the angular velocity data, using the bending change map of the flexible bending sensor as the joint angle observation value, constructing the attitude observation equation, and performing Kalman filtering correction on the attitude quaternion update result to generate the whole body attitude data.

4. The method according to claim 1, characterized in that, Step S5 is followed by a dynamic compensation step: obtaining the limitation parameter matrix of the protective suit body for the range of motion of each joint, the limitation parameter matrix being predetermined based on the structural dimensions, fabric properties and airbag inflation status of the protective suit body; constraining and correcting the whole-body posture data according to the limitation parameter matrix, when the posture angle of a certain joint exceeds the maximum movable angle of the joint in the current action direction, the posture angle is truncated to the maximum movable angle, and a protective suit limitation compensation signal is generated.

5. The method according to claim 1, characterized in that, It also includes safety warning steps: Acquire ambient temperature data collected by an ambient temperature sensor and ambient humidity data collected by an ambient humidity sensor located on the protective suit body; The spatial distance between the firefighter and the fire source is determined based on the three-dimensional position information. The current activity intensity level of the firefighter is determined by the metabolic equivalent algorithm based on the whole body posture data. The heat load level of the environment where the firefighter is located is determined by the heat index algorithm based on the ambient temperature data and the ambient humidity data. Based on the spatial distance, the activity intensity level, and the heat load level, a real-time safety factor is calculated according to a preset weighted fusion rule. When the real-time safety factor is lower than a preset safety threshold, a warning signal containing the three-dimensional location information is sent to the remote command terminal through the wireless communication module, and a graded alarm is triggered on-site through the audible and visual alarm device installed on the protective suit.

6. The method according to claim 1, characterized in that, It also includes the virtual-to-real mapping step: Obtain building information models and video surveillance data from the fire scene, and construct a three-dimensional spatial model of the fire scene; Based on the three-dimensional position information and the full-body posture data, the virtual avatar of the firefighter is driven in real time in the three-dimensional space model. The position and posture of the virtual avatar are updated synchronously with the real-time position and posture of the firefighter at a refresh rate of not less than 20Hz. Based on the three-dimensional position information and the whole-body posture data, combined with the firefighter's body shape data, a dynamic human body model of the firefighter wearing the protective suit is generated in the three-dimensional space model. The dynamic human body model is used to simulate the coordination relationship between the protective suit and the human body when the firefighter performs different combat actions, including the changes in the range of motion of each joint and the distribution of the thickness of the air layer under the suit.

7. The method according to claim 6, characterized in that, The virtual-real mapping step further includes: predicting the firefighter's movement trajectory within a future preset time window in the three-dimensional space model based on the current combat action type and the three-dimensional position information, and displaying the predicted trajectory as a virtual image in the three-dimensional space model for the remote command terminal to make situation prediction.

8. A system for tracking the position and posture of firefighters' protective clothing, characterized in that, include: The data acquisition unit is used for acceleration and angular velocity data, bending data, ranging information between the ultra-wideband positioning tag and multiple ultra-wideband base stations deployed inside the fire scene, and air pressure data. An adaptive multi-source fusion localization unit is used to construct an adaptive multi-source fusion localization model that includes an extended Kalman filter framework and a mode switching strategy. The three-dimensional position information acquisition unit is used to input the acceleration data and the angular velocity data into the extended Kalman filter for dead reckoning, and simultaneously input the ranging information and the air pressure data into the extended Kalman filter. The unit performs time synchronization and spatial alignment of the acceleration data, the angular velocity data, the ranging information and the air pressure data into multi-source data fusion to generate the three-dimensional position information of the firefighter in the fire scene. The mode switching and error constraint correction unit is used to monitor the signal quality between the ultra-wideband positioning tag and the ultra-wideband base station in real time. When the signal quality is lower than a preset threshold, the mode switching strategy switches the positioning mode from ultra-wideband assisted inertial positioning mode to pure inertial dead reckoning mode, and constrains and corrects the cumulative error of dead reckoning based on the firefighter's historical motion trajectory data and the bending data collected by the flexible bending sensor. The attitude calculation unit is used to construct bending change diagrams of each flexible bending sensor based on the bending data, and generate the firefighter's full-body attitude data by combining the acceleration data and the angular velocity data through the attitude calculation algorithm. The motion recognition unit is used to perform dynamic time-adjusted matching between the full-body posture data and a preset standard combat motion library to identify the current combat motion type of the firefighter. The wireless communication module is used to send the three-dimensional position information, the whole-body posture data, and the current combat action type to the remote command terminal.

9. An electronic device, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the method according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the method described in any one of claims 1-7.

Images