An experimental method for collecting behavioral data of infant care during high-rise residential building fire evacuation based on mixed reality technology

Abstract

The present invention belongs to the technical field of building safety and evacuation simulation, and discloses an experimental method for collecting baby care behavior data during fire evacuation in high-rise residences based on mixed reality technology. Step 1: Scan the internal environmental information of the experimental high-rise residential building and establish a three-dimensional model of the high-rise residential building; Step 2: Complete the construction of the mixed reality fire scene; Step 3: The subjects wear mixed reality headsets and wearable odor players to conduct the experiment; Step 4: Produce a human dynamic model through volumetric motion capture method; Step 5: To evaluate whether the baby evacuation is successful and mark the successful evacuation experiments; Step 6: Process the collected data and group and number the results for storage. The present invention designs an experiment for baby guardians to carry out fire emergency evacuation while taking care of babies, so as to obtain relevant evacuation behavior data for evacuating with babies in care, providing data support for subsequent simulation and research.

Description

Technical Field

[0001] The present invention belongs to the technical field of building safety and evacuation simulation, and particularly relates to an experimental method for collecting behavior data of infant care during high-rise residential building fire evacuation based on mixed reality technology. Background Technique

[0002] In case of emergencies, as a vulnerable group with relatively high passivity, infants are difficult to evacuate in a timely and safe manner and need to be evacuated under the care of guardians. However, there is a lack of research on the evacuation behaviors of infants and their guardians. Residential buildings, as the main living places of infants, are typical locations for the behavior of guardians caring for infants during evacuation. In recent years, with the development of cities, the number of high-rise residential buildings has gradually increased. High-rise residential buildings have a high population density, relatively narrow evacuation paths, limited evacuation channels, and high floors, making it difficult to escape. Once a fire occurs, it may endanger the lives of a large number of residents. For infants and their guardians in a vulnerable state, this threat is even greater. Therefore, it is urgent to study the behavior of caring for infants during high-rise residential building fire evacuation.

[0003] Existing evacuation experiments based on virtual reality technology completely isolate the subjects from the real physical environment, lacking the sense of interaction with the real space. In particular, for buildings with a compact space layout, there is a lack of simulation means for the physical interactions between evacuees and building walls, doors and windows, furniture, and other equipment and items, as well as the physical interactions between evacuees. Evacuation experiments based on mixed reality can intuitively superimpose and display high-precision fire scenes on the real environment, assisting the subjects in intuitively experiencing the virtual fire scene during the evacuation experiment, enhancing the immersion of the subjects while strengthening the connection between the subjects and the real world. Summary of the Invention

[0004] The present invention provides an experimental method for collecting behavior data of infant care during high-rise residential building fire evacuation based on mixed reality technology, aiming to design an experiment for infant guardians to carry out fire emergency evacuation while caring for infants, so as to obtain relevant evacuation behavior data for caring for infants during evacuation and provide data support for subsequent simulation and research.

[0005] The present invention is realized through the following technical solutions:

[0006] An experimental method for collecting behavior data of infant care during high-rise residential building fire evacuation based on mixed reality technology, the experimental method includes the following steps,

[0007] Step 1: Scan the internal environmental information of the experimental high-rise residential building, establish a three-dimensional model of the high-rise residential building, and simulate a fire scene including the ignition point, acoustic environment, thermal environment, visual information, and smoke flow information through pyrosim, where the internal environmental information includes internal environmental images and position information;

[0008] Step 2: After calibrating and matching the images of the internal environment of high-rise residential buildings in Step 1 with the location information, place the visual information of the simulated fire scene into the real building internal environment, and simultaneously simulate the sound environment and thermal environment in the fire scene to complete the construction of the mixed fire scene;

[0009] Step 3: Based on the fire scene constructed in Step 2, the subjects wear mixed reality headsets and wearable odor players to conduct experiments; among them, the experimental group subjects carry infant body models to evacuate, and the control group subjects evacuate without carrying infant body models, and collect the evacuation behavior data of the experimental group and the control group respectively;

[0010] Step 4: Process the human body posture data in the subjects' evacuation behavior data collected in Step 3 through volume motion capture methods and create a human dynamic model;

[0011] Step 5: Evaluate whether the infant evacuation is successful according to the collision data of the infant body model in the experimental group collected in Step 3 and the height of the subjects and the infant body model detected by the horizontal evacuation path infrared detector, and mark the experiments with successful evacuations;

[0012] Step 6: Process the data collected in Step 3, where the data includes various sensor data of the infant body model, the subjects' eye movement data, and the entire evacuation process video images; match the basic information of the subjects, the basic information of participating in the experiment, and the various data collected in the experiment, and group and number the results for storage.

[0013] Further, the specific content of Step 1 is as follows:

[0014] Build a building structure model through revit for fire scene simulation and mixed reality scene positioning;

[0015] Simulate the fire scene through pyrosim software, construct a virtual fire scene, and obtain fire scene data; export the temperature, smoke, and flame information in the virtual fire scene. The fire scene data specifically includes the location of the ignition point, flame temperature, fire heat release rate, smoke concentration, smoke layer height, and smoke spread speed;

[0016] Use blender software to build a dynamic model of the fire scene, set the particle effects of flames and smoke through the UE5 engine, and add combustion and explosion sound effects at the ignition point to render a high-precision dynamic fire scene and obtain virtual fire scene images.

[0017] Further, the specific content of Step 2 is as follows:

[0018] Calibrate the collected building point cloud images with the positioning data, match the environmental images with the location information, and determine the mapping relationship between the virtual fire scene images and the environmental images;

[0019] During the evacuation experiment, scan the real-time environmental images during evacuation through the built-in camera of the mixed reality headset, and use the display device of the mixed reality headset to display the visual information in the simulated fire scene in the real space;

[0020] Set up an audio and temperature control system in the high-rise residential building to simulate the acoustic environment and thermal environment in the fire scene. The temperature control device consists of an electric heater and an electronic controller;

[0021] Control the odor player to simulate the pungent smoke generated by the combustion at the fire scene, and perform real-time odor simulation based on the smoke concentration, smoke layer height, smoke spread speed and the subject's positioning information obtained in step 1;

[0022] Arrange high-definition infrared cameras and infrared detectors on the evacuation path to collect the evacuation images of the subjects and the height of the subjects' heads.

[0023] Furthermore, the specific content of step 3 is as follows:

[0024] Step 3.1. Recruit volunteers of different ages and genders with experience in taking care of infants as the experimental group, and recruit volunteers of the same age and gender as the experimental group as the control group. The subjects in the control group do not need experience in taking care of infants;

[0025] Group the ages of the subjects in the experimental group and the control group into 20 - 29 years old, 30 - 39 years old, 40 - 49 years old, 50 - 59 years old, and 60 - 69 years old;

[0026] Step 3.2. Conduct evacuation training for the subjects, including the purpose and method of the experiment, precautions for the experiment, method for judging the success of evacuation, method for scoring the evacuation experiment, and evacuation path of the experimental building;

[0027] Step 3.3. The subjects wear MR headset devices, wearable odor players, physiological state collection sensors and positioning tags. In the experimental group, the subjects carry the infant human model in the manner of holding an infant to conduct evacuation, and in the control group, the subjects conduct evacuation without carrying the infant human model; Each evacuation experiment is carried out separately for each subject, and experimental data is collected;

[0028] Step 3.4. After the experiment, the subjects fill out the physical strength scale and the immersion evaluation scale to obtain subjective evaluation data on exercise intensity and immersion.

[0029] Further, in step 3.3, baby mannequins of the same weight and volume as babies of different months are selected respectively. The baby mannequins are built-in with speakers, UWB positioning tags and gyroscopes. Acceleration sensors are respectively placed on the head, neck and spine of the baby, and clay is installed on the abdomen of the baby mannequin;

[0030] The built-in speaker is used to simulate the crying of a baby;

[0031] The UWB positioning tag is used to record the real-time position information of the baby mannequin;

[0032] The acceleration sensor is used to collect the data of shaking, impact and extrusion suffered by the baby mannequin, and is used to determine whether the baby replaced by the model encounters extrusion and bruising during the evacuation process;

[0033] The clay is used to determine whether the baby's abdomen is severely squeezed.

[0034] Further, the volume motion capture method in step 4 includes:

[0035] Step 4.1: Use a high-definition infrared camera to shoot the evacuation process of the subject;

[0036] Step 4.2: Use a 4DViews volumetric capture system to record the evacuation process of the subject, and output a 4DViews format animation file, which contains a dynamic 3D human model;

[0037] Step 4.3: Import the animation file in step 4.2 into Blender, use the Python API and the Iterative Closest Point alignment plugin to track the movement of the animated human body, generate an aligned object, and calculate the virtual marker position; output the tracking row-column file;

[0038] Step 4.4: Import the tracking row-column file in step 4.3 into the open-source biomechanics analysis software OpenSim, align the anatomical skeleton model from OpenSim with the imported marker tracking data through inverse kinematics, generate a dynamic skeleton model, calculate the angles of joint flexion and extension and visualize them, and establish a human dynamic model.

[0039] Further, the establishment of the human dynamic model in step 4.4 is specifically as follows,

[0040] Step 4.4.1: Use the Python API and the Iterative Closest Point alignment plugin of Blender to develop a program to track the movement of the dynamic human model, use the Boolean operator of Blender to generate the intersection of the aligned object and the base mesh, and transform the aligned object to represent the body part; use the forward tracking and backward tracking features of the program, and use the Iterative Closest Point ICP algorithm to register each aligned object to the mesh in each frame of the animation;

[0041] Step 4.4.2: Each alignment object has two to three additional markers. When the alignment object is generated, the markers are automatically placed at the default positions and can be moved manually. When the alignment object is tracked to the animated mesh by the ICP algorithm, the markers will also be animated accordingly.

[0042] Step 4.4.3: The positions of the markers in the entire animation sequence are exported as a Track Row Column file.

[0043] Further, the Iterative Closest Point (ICP) algorithm in Step 4.4.1 is specifically as follows. Suppose there are two sets of 3D positions: X = {x1,..., x n}, Y = {y1,..., y n}; each x i and y i is a three-dimensional coordinate; to find the translation vector t and the rotation matrix R that minimize the sum of the squared errors E, the formula is:

[0044]

[0045] where the calculation methods of R and t are as follows:

[0046] Center each set of points at the origin by subtracting the centroid from each point

[0047]

[0048] Calculate the Singular Value Decomposition (SVD) of the N matrix A:

[0049] A = USV T

[0050] where S has the singular values and is diagonal, and both U and V are orthogonal matrices;

[0051] Therefore, the optimal R and t that minimize E are:

[0052] R = UV T

[0053]

[0054] Further, the evaluation criteria for the evacuation experiment results in Step 5 are as follows:

[0055] The detection results of the infrared detectors during the horizontal evacuation process are used to evaluate whether the subjects and the infant human models avoid the fire smoke layer during the horizontal evacuation process;

[0056] The collision degree of the infant human model is used to evaluate the safety of the simulated infant during evacuation; the judgment method is as follows: the synthesized acceleration of the chest acceleration sensor of the infant human model throughout the process does not exceed 55g, the synthesized acceleration from the abdomen to the head does not exceed 30g, and there should be no obvious indentation marks on the soil of the infant human model's abdomen.

[0057] Further, step 6 is specifically as follows.

[0058] Step 6.1: Match the data described in step 2, group and number the experimental data of each subject, and attach a scoring table and a subjective evaluation scale.

[0059] Step 6.2: Obtain the evacuation path of each subject, the average speed, maximum speed, and evacuation distance during horizontal evacuation, and the average speed, maximum speed, and evacuation distance during vertical evacuation according to the real-time positioning data.

[0060] Step 6.3: Store the data obtained above.

[0061] The beneficial effects of the present invention are as follows:

[0062] The present invention proposes an experimental method for collecting data on the behavior of caring for infants during fire evacuation in high-rise residential buildings based on mixed reality technology for the study of the behavior of guardians caring for infants in the event of a fire in high-rise residential buildings. By simulating a fire scene in a real environment through mixed reality technology, the fire evacuation behavior of guardians when caring for infants is recorded, and behavioral data is collected, providing an important basis for the establishment of an infant care evacuation behavior model, thereby accurately and effectively predicting the evacuation time and evacuation safety of infants and guardians in high-rise residential buildings. Description of the Drawings

[0063] Figure 1 is the flowchart of the method of the present invention.

[0064] Figure 2 is the system diagram of the present invention. Detailed Embodiments

[0065] In the following description, specific details such as specific system structures and technologies are presented for the purpose of illustration rather than limitation, so as to thoroughly understand the embodiments of the present application. However, those skilled in the art should clearly understand that the present application can also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid unnecessary details from interfering with the description of the present application.

[0066] It should be understood that when used in this specification and the appended claims, the term "comprising" indicates the presence of the described features, integers, steps, operations, elements and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or their groups.

[0067] It should also be understood that the terms used in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application. As used in the specification of this application and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include the plural forms.

[0068] The following combines the appendix of the specification of this application Figure 1-2 , and clearly and completely describes the technical solutions in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of this application.

[0069] Many specific details are set forth in the following description in order to provide a thorough understanding of this application, but this application may also be implemented in other ways different from those described herein. Those skilled in the art can make similar extensions without departing from the connotation of this application. Therefore, this application is not limited by the specific embodiments disclosed below.

[0070] Embodiment 1

[0071] This embodiment provides an experimental method for collecting baby care behavior data during high-rise residential fire evacuation based on mixed reality technology. According to Figures 1 to 2 shown, the experimental method includes the following steps,

[0072] Step 1: Scan the internal environmental information of the experimental high-rise residential building, establish a three-dimensional model of the high-rise residential building, and simulate a fire scene including the fire origin, acoustic environment, thermal environment, visual information and smoke flow information through Pyrosim, where the internal environmental information includes internal environmental images and location information;

[0073] Further, the specific content of Step 1 is,

[0074] Establish a 1:1 structural model of the building through Revit for fire scene simulation and mixed reality scene positioning;

[0075] Simulate the fire scene through Pyrosim software, construct a virtual fire scene, and obtain fire scene data; export the temperature, smoke, and flame information in the virtual fire scene. The fire scene data specifically includes the location of the ignition point, flame temperature, fire heat release rate, smoke concentration, smoke layer height, and smoke spread speed.

[0076] Use Blender software to establish a dynamic model of the fire scene, set the particle effects of flames and smoke through the UE5 engine, and add combustion and explosion sound effects at the ignition point to render a high-precision dynamic fire scene and obtain virtual fire scene images.

[0077] Step 2: After calibrating and matching the images and location information of the internal environment of the high-rise residential building in Step 1, use mixed reality technology to place the visual information of the simulated fire scene into the real building internal environment. At the same time, use the audio system and temperature control system in the experimental high-rise residential building to simulate the sound environment and thermal environment in the fire scene to complete the construction of the mixed reality fire scene.

[0078] The sound, temperature control system, and odor player in the experimental building in Step 2 simulate the sound environment, thermal environment, and pungent odor dispersion in the fire scene in the real environment.

[0079] Further, Step 2 is specifically as follows:

[0080] Calibrate the collected building point cloud images and positioning data, match the environmental images and location information, and determine the mapping relationship between the virtual fire scene images and the environmental images.

[0081] During the evacuation experiment, use the built-in camera of the mixed reality headset to scan the real-time environmental images during evacuation, and use the display device of the mixed reality headset to display the visual information in the simulated fire scene in the real space.

[0082] Set up an audio and temperature control system in the high-rise residential building to simulate the sound environment and thermal environment in the fire scene. The temperature control device consists of an electric heater and an electronic controller.

[0083] Control the odor player through Bluetooth function to simulate the pungent smoke generated by combustion at the fire scene, and perform real-time odor simulation according to the smoke concentration, smoke layer height, smoke spread speed, and the positioning information of the subjects obtained in Step 1.

[0084] Arrange high-definition infrared cameras and infrared detectors on the evacuation path to collect the evacuation images of the subjects and the height of the subjects' heads.

[0085] Step 3: Based on the fire scene constructed in step 2, the subjects wore mixed reality headsets and wearable odor players to conduct experiments in high-rise residential buildings; the subjects in the experimental group carried baby mannequins for evacuation, while the subjects in the control group did not carry baby mannequins for evacuation, and the evacuation behavior data of the experimental group and the control group were collected respectively. The baby mannequin would produce crying sounds of different volumes according to the subject's behavior; the collected experimental data included the whole evacuation experiment video, the subject's evacuation behavior data, the baby mannequin's collision data, and the subjective evaluation of the immersiveness of the mixed reality evacuation experiment;

[0086] The mixed reality head-mounted display device in step 3 is used to place the simulated fire visual scene into the real environment, and the built-in eye tracker and positioning tag are used to collect the subject's eye movement information and positioning information;

[0087] The infant human model in step 3 is used by the subject to simulate assisting the infant in evacuation, and the built-in acceleration sensor and positioning tag are used to collect the external force and positioning information of the model in the experiment;

[0088] The high-definition infrared camera in step 3 is arranged on the evacuation path to perform unmarked volume motion capture; the infrared detector is used to determine whether the mouth and nose positions of the subject and the infant mannequin have entered the smoke layer;

[0089] The physiological state collection unit includes a blood oxygen sensor and a heart rate sensor. The physiological state collection unit is worn on the subject's body. The collection ends of the blood oxygen sensor and the heart rate sensor are placed at the blood vessel dense area of ​​the human wrist to collect blood oxygen saturation and real-time exercise heart rate during human exercise.

[0090] Furthermore, the infant mannequin in step 3 is used to replace the infant for the experiment, and infant mannequins of the same weight and volume as infants of different months of age are selected respectively, the infant mannequin has a built-in speaker, a UWB positioning tag and a gyroscope, acceleration sensors are placed on the infant's head, neck and spine respectively, and clay is installed on the abdomen of the infant mannequin;

[0091] The built-in speaker is used to simulate the crying sound of a baby;

[0092] The UWB positioning tag is used to record the real-time position information of the baby mannequin;

[0093] The acceleration sensor is used to collect shaking, impact and squeezing data of the infant mannequin, so as to determine whether the infant represented by the mannequin is squeezed or injured during the evacuation process;

[0094] The clay is used to determine whether the baby's abdomen is severely squeezed.

[0095] Furthermore, after the evacuation experiment in step 3 begins, the Bluetooth controller causes the built-in speaker of the baby mannequin to emit a baby crying sound, and the crying sound is controlled at 60dB; when the shaking amplitude of the baby mannequin's head and neck is greater than 5 cm, the crying sound emitted by the built-in speaker is 80dB; when the synthetic acceleration of the acceleration sensor at the baby mannequin's head, neck or spine exceeds 50g, the crying sound emitted by the built-in speaker is 100dB.

[0096] Furthermore, the step 3 is specifically as follows:

[0097] Step 3.1. Recruit volunteers of different ages and genders who have experience in caring for infants as the experimental group, and recruit volunteers of the same age and gender as the experimental group as the control group. The control group subjects do not need experience in caring for infants;

[0098] The age of the subjects in the experimental group and the control group were divided into groups according to 20-29 years old, 30-39 years old, 40-49 years old, 50-59 years old, and 60-69 years old;

[0099] Step 3.2. Conduct evacuation training for the subjects, including: experimental purpose and method, experimental precautions, method for determining successful evacuation, evacuation experiment scoring method, and evacuation route of the experimental building;

[0100] Step 3.3. The subjects wore MR head-mounted display devices, wearable odor players, physiological state collection sensors and positioning tags. The experimental group of subjects carried the baby mannequin in the form of holding the baby for evacuation, and the control group of subjects did not carry the baby mannequin for evacuation. Each evacuation experiment was conducted individually for the subjects, and the experimental data were collected;

[0101] Step 3.4. After the experiment, the subjects filled out the physical strength scale and immersion evaluation scale to obtain the subjective evaluation data of exercise intensity and immersion;

[0102] Furthermore, in step 3.3, after collecting the subject's head height, it is used to generate feedback in the mixed reality scene. Specifically, when the high-definition infrared camera detects that the subject's head height is higher than the smoke layer of the virtual fire scene, the smoke concentration in the mixed reality head display screen changes according to the simulated fire scene in step 1, and the field of view of the screen displayed by the mixed reality head display becomes narrower, the viewing distance becomes shorter, and the picture clarity decreases;

[0103] The odor output by the mixed reality headset odor player changes according to the smoke flow conditions of the simulated fire scene described in step 1; the change is manifested in the change of the odor output effect as the subject adjusts his position and height. For example, when the subject is in a position with high smoke concentration in the simulated scene, the odor concentration output by the odor player increases with the simulation data.

[0104] Further, the data collected in step 3.3 includes the basic information of the subjects, including the subject number, gender, age, height, weight, and the duration of taking care of the baby;

[0105] The information of the subjects participating in the experiment, including the experiment date, order, experiment room information, whether to carry a baby model or not, and the weight of the carried baby model;

[0106] The real-time data of the pressure sensors, the built-in gyroscope data, and the real-time positioning data during the evacuation process of the baby human model;

[0107] The evacuation behavior data of the subjects, including the total evacuation duration of the subjects, the horizontal and vertical evacuation durations, the real-time positioning information during the evacuation process; eye movement data, including the position of the fixation point, the position of the eye, and the position of the head; the video images of the whole evacuation process of the subjects; the real-time head height of the subjects;

[0108] The subjective evaluation of the immersion degree and the fatigue degree evaluation level of the mixed reality experiment by the subjects after the experiment;

[0109] The information recorded in the experiment also includes the simulated fire scene information, specifically, the fire ignition point, the real-time image of the scene, the smoke concentration at different positions, the smoke height, the visibility data, the acoustic environment, and the thermal environment data.

[0110] Table 1 Subjective Physical Strength Scale:

[0111] Table 2 Immersion Evaluation Scale:

[0112]

[0113]

[0114] Step 4: Process the human body posture data in the evacuation behavior data of the subjects collected in step 3 by using the volumetric motion capture method, and create a human dynamic model;

[0115] Further, the volumetric motion capture method in step 4 includes:

[0116] Step 4.1: Use a high-definition infrared camera to shoot the evacuation process of the subjects;

[0117] Step 4.2: Use the 4DViews volumetric capture system to record the evacuation process of the subjects, and output a 4DViews format (.4ds) animation file, which contains a dynamic 3D human model;

[0118] Step 4.3: Import the animation file from Step 4.2 into Blender, use the Python API and the Iterative Closest Point (ICP) alignment plugin to track the movement of the animated human body, generate an aligned object, and calculate the virtual marker positions; output a tracking row-column (.trc) file.

[0119] Step 4.4: Import the tracking row-column file described in Step 4.3 into the open-source biomechanics analysis software OpenSim. Align the anatomical skeleton model from OpenSim with the imported marker tracking data through inverse kinematics to generate a dynamic skeleton model, calculate and visualize the angles of joint flexion and extension, and establish a human dynamic model.

[0120] Specifically, Step 4.4 is as follows:

[0121] Step 4.4.1: Use the Python API of Blender and the Iterative Closest Point (ICP) alignment plugin to develop a program to track the movement of the dynamic human body model. Utilize the Boolean operator of Blender to generate the intersection of the aligned object and the base mesh, and transform the aligned object to represent body parts; use the forward tracking and backward tracking features of the program to register each aligned object to the mesh at each frame of the animation; this program achieves this by registering the aligned object with the mesh using the Iterative Closest Point (ICP) algorithm. The Iterative Closest Point (ICP) algorithm aims to find the transformation that aligns a set of points with a 3D surface or another set of points.

[0122] Specifically, assume there are two sets of 3D positions: X = {x1,..., x n}, Y = {y1,..., y n}. Each x i and y i is a three-dimensional coordinate; to find the translation vector t and the rotation matrix R that minimize the sum of the squared errors E, the formula is:

[0123]

[0124] where the calculation methods of R and t are as follows:

[0125] Center each set of points at the origin by subtracting the centroid from each point

[0126]

[0127] Calculate the singular value decomposition (SVD) of the N matrix A:

[0128] A = USV T

[0129] where S has singular values and is diagonal, and both U and V are orthogonal matrices; therefore, the optimal R and t that minimize E are:

[0130] R = UV T

[0131]

[0132] Step 4.4.2: Each alignment object has two to three additional markers. When generating the alignment object, the markers are automatically placed in the default positions and can be moved manually. When the alignment object is tracked to the animated mesh by the ICP algorithm, the markers will also be animated accordingly.

[0133] Step 4.4.3: The positions of the markers in the entire animation sequence are exported as a Track Row Column (.trc) file.

[0134] Step 5: Evaluate whether the evacuation of the infant is successful based on the collision data collected by the infant human model in Step 3 and the height of the subject detected by the infrared detector for the horizontal evacuation path, and mark the successful evacuation experiments.

[0135] Further, the evaluation criteria for the evacuation experiment results in Step 5 are as follows:

[0136] The detection result of the infrared detector during the horizontal evacuation is used to evaluate whether the subject and the infant human model avoid the fire smoke layer during the horizontal evacuation.

[0137] The collision degree of the infant human model is used to evaluate the safety degree of the simulated infant during the evacuation. The determination method is as follows: The total synthesized acceleration of the acceleration sensor on the chest of the infant human model does not exceed 55g, the total synthesized acceleration along the vertical direction from the abdomen to the head does not exceed 30g, and there should be no obvious depression marks on the soil of the abdomen of the infant human model.

[0138] Step 6: Process the data collected in Step 3, where the data includes various sensor data of the infant human model (acceleration sensor data, built-in gyroscope data, real-time positioning data), the eye movement data of the subject, and the video images of the entire evacuation process; match the basic information of the subject, the basic information of participating in the experiment with the various data collected in the experiment, and store the results in groups with numbers.

[0139] Further, Step 6 is specifically as follows:

[0140] Step 6.1: Match the various data described in Step 2, group and number the experimental data of each subject, and attach a scoring table and a subjective evaluation scale.

[0141] Step 6.2: Obtain the evacuation path of each subject, the average speed, maximum speed, and evacuation distance during horizontal evacuation, and the average speed, maximum speed, and evacuation distance during vertical evacuation based on the real-time positioning data.

[0142] Step 6.3: Store the data obtained above.

Claims

1. An experimental method for collecting data on infant care behavior during high-rise residential fire evacuation based on mixed reality technology, characterized in that: The experimental method comprises the following steps, Step 1: Scan the internal environment information of the experimental high-rise residential building, establish a three-dimensional model of the high-rise residential building, and simulate the fire scene including the fire point, acoustic environment, thermal environment, visual information and smoke flow information through pyrosim, wherein the internal environment information includes the internal environment image and location information; Step 2: After calibrating and matching the image and location information of the interior environment of the high-rise residential building in step 1, the simulated fire scene visual information is placed in the real building interior environment, and the acoustic and thermal environments in the fire scene are simulated to complete the construction of the mixed reality fire scene; Step 3: Based on the fire scene built in step 2, the subjects wore mixed reality headsets and wearable odor players to conduct the experiment; the subjects in the experimental group evacuated with a baby mannequin, while the subjects in the control group evacuated without a baby mannequin, and the evacuation behavior data of the experimental group and the control group were collected respectively; The step 3 is specifically as follows: Step 3.

1. Recruit volunteers of different ages and genders who have experience in caring for infants as the experimental group, and recruit volunteers of the same age and gender as the experimental group as the control group. The control group subjects do not need experience in caring for infants; Step 3.

2. Conduct evacuation training for the subjects; Step 3.

3. The subject wears the MR head-mounted display device, the wearable odor player, the physiological state collection sensor and the positioning tag; Step 3.

4. After the experiment, the subjects filled out the physical strength scale and immersion evaluation scale to obtain the subjective evaluation data of exercise intensity and immersion; In step 3.3, infant mannequins of the same weight and volume as infants of different months of age are selected respectively, wherein the infant mannequins are equipped with built-in speakers, UWB positioning tags and gyroscopes, acceleration sensors are placed on the infant's head, neck and spine respectively, and clay is installed on the abdomen of the infant mannequins; The built-in speaker is used to simulate the crying sound of a baby; The UWB positioning tag is used to record the real-time position information of the baby mannequin; The acceleration sensor is used to collect shaking, impact and squeezing data of the infant mannequin, so as to determine whether the infant represented by the mannequin is squeezed or injured during the evacuation process; The clay is used to determine whether the baby's abdomen is severely squeezed; Step 4: Process the human body posture data in the evacuation behavior data of the subjects collected in step 3 by using the volumetric motion capture method, and create a human body dynamic model; Step 5: Evaluate whether the infant evacuation is successful based on the collision data of the infant mannequin of the experimental group collected in step 3 and the height between the subject and the infant mannequin detected by the infrared detector of the horizontal evacuation path, and mark the experiment with successful evacuation; Step 6: Process the data collected in step 3, including the sensor data of the infant mannequin, the eye movement data of the subjects, and the video images of the entire evacuation process; match the basic information of the subjects and the basic information of the experiment participants with the various data collected in the experiment, and store the results by grouping and numbering.

2. The experimental method for collecting data on infant care behavior during fire evacuation in high-rise residential buildings according to claim 1 is characterized in that: The step 1 specifically comprises: Use Revit to build a building structure model for fire scene simulation and mixed reality scene positioning; Use pyrosim software to simulate fire scenes, build virtual fire scenes, and obtain fire scene data; export temperature, smoke and flame information in virtual fire scenes. Fire scene data specifically includes the location of the fire point, flame temperature, fire heat release rate, smoke concentration, smoke layer height and smoke spread speed; Use blender software to build a dynamic model of the fire scene, set the particle effects of flame and smoke through the UE5 engine, add burning and explosion sound effects at the fire point, render high-precision dynamic fire scenes, and obtain virtual fire scene images.

3. The experimental method for collecting data on infant care behavior during fire evacuation in high-rise residential buildings according to claim 2 is characterized in that: The step 2 is specifically as follows: Calibrate the collected building point cloud images and positioning data, match the environmental images and location information, and determine the mapping relationship between the virtual fire scene image and the environmental image; During the evacuation experiment, the real-time environment image during evacuation was scanned by the built-in camera of the mixed reality headset, and the visual information of the simulated fire scene was displayed in the real space using the display device of the mixed reality headset; An audio and temperature control system is installed in high-rise residential buildings to simulate the acoustic and thermal environment in a fire scene. The temperature control equipment consists of an electric heater and an electronic controller. Control the odor player to simulate the pungent smoke produced by burning at the fire scene, and perform real-time odor simulation based on the smoke concentration, smoke layer height, smoke spreading speed and subject location information obtained in step 1; High-definition infrared cameras and infrared detectors were arranged on the evacuation path to collect the subjects’ evacuation images and the subjects’ head heights.

4. The experimental method for collecting data on infant care behavior during fire evacuation in high-rise residential buildings according to claim 1 is characterized in that: The volume motion capture method in step 4 comprises: Step 4.1: Use a high-definition infrared camera to record the evacuation process of the subjects; Step 4.2: Use a 4DViews volume capture system to record the evacuation process of the subject, and output a 4DViews format animation file, wherein the animation file includes a dynamic 3D human body model; Step 4.3: Import the animation file of step 4.2 into Blender, use Python API and iterative closest point alignment plug-in to track the movement of the animated human body, generate alignment objects, and calculate the virtual marker position; output the tracking row and column file; Step 4.4: Import the tracking row and column files described in step 4.3 into the open source biomechanical analysis software OpenSim, align the anatomical skeleton model from OpenSim with the imported marker tracking data through inverse kinematics, generate a dynamic skeleton model, calculate and visualize the angles of joint flexion and extension, and establish a dynamic model of the human body.

5. The experimental method for collecting data on infant care behavior during fire evacuation in high-rise residential buildings according to claim 4 is characterized in that: The step 4.3 is specifically as follows: Step 4.3.1: Using Blender’s Python API and the Iterative Closest Point Alignment plugin, develop a program to track the motion of the dynamic human model, using Blender’s Boolean operators to generate the intersection of the alignment objects and the base mesh, and transform the alignment objects to represent body parts; using the program’s forward tracking and backward tracking features, register each alignment object to the mesh using the Iterative Closest Point (ICP) algorithm on each frame of the animation; Step 4.3.2: Each alignment object has two or three markers attached to it. When the alignment object is generated, the markers are automatically placed in default positions, and they can be moved manually. When the alignment object is tracked to the animation mesh through the ICP algorithm, the markers will also be animated accordingly. Step 4.3.3: The positions of the markers throughout the animation sequence are exported as a Track Row Column file.

6. The experimental method for collecting data on infant care behavior during fire evacuation in high-rise residential buildings according to claim 5 is characterized in that: The specific step 4.3.1 of iterating the closest point ICP algorithm is as follows: assuming there are two sets of 3D positions: X = {x1, ..., x n }, Y={y1,…,y n }; Each x i and i are all three-dimensional coordinates; to find the translation vector t and the rotation matrix R so that the sum of the squared error E is minimized, the formula is: The calculation methods of R and t are as follows: By subtracting each point from the centroid, both sets of points are centered at the origin, Perform singular value decomposition on the covariance matrix A: A=USV T Among them, S is a diagonal matrix, its diagonal elements are singular values, and U and V are both orthogonal matrices; Therefore, the optimal R and t that minimize E are: R=UV T 7. The experimental method for collecting data on infant care behavior during fire evacuation in high-rise residential buildings according to claim 3 is characterized in that: The evaluation criteria for the evacuation test results in step 5 are: The infrared detector test results during horizontal evacuation were used to evaluate whether the subjects and the infant mannequin avoided the fire smoke layer during horizontal evacuation; The impact degree of the infant mannequin is used to evaluate the safety of the simulated infant during evacuation; The determination method is as follows: the total composite acceleration of the infant mannequin's chest acceleration sensor does not exceed 55g, the vertical composite acceleration from the abdomen to the head does not exceed 30g, and there should be no obvious signs of depression in the soil on the infant mannequin's abdomen.

8. The experimental method for collecting data on infant care behavior during fire evacuation in high-rise residential buildings according to claim 3 is characterized in that: The step 6 is specifically as follows: Step 6.1: Match the data described in step 2, group and number the experimental data of each subject, and attach a scoring sheet and a subjective evaluation scale; Step 6.2: Obtain the evacuation path, average speed, maximum speed and evacuation distance of each subject during horizontal evacuation, and average speed, maximum speed and evacuation distance during vertical evacuation based on the real-time positioning data; Step 6.3: Store the above data.

Images