KR20240049860A - Multi disaster prevention system of railway facility based on digital twin, and method for the same - Google Patents

Abstract

Establish a digital twin for the interior of railway facilities such as railways, tunnels, and underground stations, and allow managers to easily detect abnormalities inside railway facilities, accurately identify indoor and outdoor positioning information through client terminals such as pedestrians and passers-by, By virtually simulating the optimal evacuation route in the event of a disaster, the optimal evacuation route can be guided in the event of an actual disaster. In addition, in order to build a complex disaster prevention system for digital twin-based railway facilities, digital twin-based railway facilities can be installed through a single platform. By applying indoor and outdoor positioning technologies together, disaster prevention and maintenance of railway facilities can be easily performed. In addition, during indoor and outdoor vertical positioning, altitude changes are measured and measured using a barometer built into the client terminal, a smartphone. , A complex disaster prevention system and method for digital twin-based railway facilities that can accurately perform vertical positioning by correcting real-time changes in air pressure are provided.

Description

Complex disaster prevention system and method for digital twin-based railway facilities {MULTI DISASTER PREVENTION SYSTEM OF RAILWAY FACILITY BASED ON DIGITAL TWIN, AND METHOD FOR THE SAME}

The present invention relates to complex disaster prevention of digital twin-based railway facilities. More specifically, building a digital twin inside railway facilities (or railway SOC) such as railways, tunnels, and underground stations, and preparing for the occurrence of disaster situations. It relates to a complex disaster prevention system and method for digital twin-based railroad facilities that can prepare for disasters within railroad facilities through simulation of disaster situations through virtual space and simulation that induces evacuation in disaster situations.

In general, Digital Twin is a concept representing the era of the 4th Industrial Revolution where everything is digitized, and refers to a model that exists in cyberspace for a real system, for example, manufacturing, power, aviation, and smart systems. It is used in various fields such as cities.

These digital twins are virtual models (e.g., discrete event models, physics-based models, statistical models, machine learning models, etc.) that maintain identity and consistency with physical assets (e.g., target systems) and depict dynamic properties. and

For example, based on inputs such as real-time operation data of the target system, comprehensive data of the same group, operation history, maintenance history, failure mode effect analysis (FMEA), finite element analysis (FEA) model, CAD model, etc. Simulation can be performed to monitor, diagnose, or predict the operation of a real target system, or to optimize operating conditions (parameters).

Figure 1 is a diagram for explaining the concept of a general conventional digital twin.

Referring to Figure 1, the characteristic of this digital twin is that it maintains constant synchronization with the actual system based on big data about the actual system collected through Internet of Things (IoT) sensors.

And through simulation of the synchronized model, the state of the actual system can be analyzed, optimally controlled, and future failures can be predicted.

In addition, the most important part of digital twin is to develop a simulation model according to the analysis purpose by well reflecting the characteristics of the actual system.

Methods for developing such simulation models largely include the traditional modeling and simulation (M&S) method and the machine learning method using big data, which has recently attracted attention.

Specifically, the M&S method is a systems science approach, which means creating an abstract model in the form of a mathematical equation or logical algorithm using the operating principles and knowledge of the target system, and the causal relationship between the input and output of the target system. can be expressed.

On the other hand, the machine learning method expresses the correlation between input and output by learning big data about the target's input and output through an artificial neural network.

As deep learning technology has recently developed, machine learning methods can develop models with high fidelity that reflect the characteristics of the system well, even without knowing the operation principles or knowledge of the target system, if there is sufficient input and output data about the system. Therefore, it is attracting attention.

However, there are limitations in applying it when it is difficult to collect data or when previously learned data is not applied, that is, when the configuration or operating laws of the system are changed.

In other words, in the case of the M&S method, even if it is difficult to secure data, a model can be developed effectively if the characteristics and operating principles of the target system can be identified. Additionally, unlike data-based models developed through machine learning, it can be used even when the system's configuration or operating laws change. However, there are always problems with reliability verification.

Accordingly, attempts have recently been made to combine M&S methods and machine learning methods, which have complementary strengths and weaknesses, in order to develop a highly reliable digital twin model.

Meanwhile, as prior art related to the above-described digital twin, an invention titled “Method for building a 3D digital twin using a drone” is disclosed in Republic of Korea Patent No. 10-2355769.

Specifically, a number of NFC tags are installed, sensors are installed to detect abnormalities in each section of the tunnel, data is generated by receiving detection signals from the sensors, a drone is flown inside the tunnel to take pictures, and an identification code is collected from the NFC tag. By receiving and displaying 3D images of the tunnel section, we are disclosing the construction of a digital twin within the tunnel through drone filming, NFC tag identification code reception, and anomaly detection sensors.

As another prior art, Republic of Korea Patent Publication No. 2017-40652 discloses an invention titled “Indoor positioning device and indoor positioning method based on sensor images.”

Specifically, the signal strength of the sensor measured at the user terminal is visualized as an icon, and the icon is mapped to the point where the sensor is located to create an image map.

An image conversion algorithm is used to distinguish visualization icons according to sensor data, and a configuration that detects the center point within the polygon formed by the icon as the coordinates of the terminal is being disclosed.

As another prior art, Republic of Korea Patent Publication No. 2019-121275 discloses an invention titled “Indoor positioning system, device, and method.”

Specifically, when an indoor map of an indoor space is parasitic, the map data and information of the parasitic indoor map are acquired, an image of the indoor space is acquired, and the map data of the parasitic indoor map of the indoor space is obtained. We are disclosing a system that performs indoor positioning of an indoor space based on the matching results by matching newly acquired image information.

As another prior art, Republic of Korea Patent Publication No. 2021-0096392 discloses an invention titled “Intelligent evacuation guidance system for disaster situations.”

Specifically, by using Internet of Things (IoT) sensors and location-based information, the shortest evacuation route and optimal evacuation route considering the number of people indoors and crowded areas can be identified in fire and other disaster situations, and representative guidance can be delivered to people in disaster situations. As a system, we are launching a simulation system that takes into account the evacuation route and expected evacuation time.

As another prior art, Republic of Korea Patent Publication No. 2022-71880 discloses an invention titled “Underground tunnel customized digital twin disaster management system,” which is explained with reference to FIGS. 2A and 2B.

Figure 2a is a configuration diagram to explain a digital twin disaster management system customized for underground tunnels according to conventional technology, and Figure 2b is a specific configuration diagram of the digital twin model management sub-device shown in Figure 2a.

Referring to Figures 2a and 2b, the digital twin disaster management system customized for underground tunnels according to conventional technology is

It includes a sensor sub-unit 10, a disaster management sub-unit 20, a digital twin model management sub-unit 30, a digital archive sub-unit 40, and a network sub-unit 50.

The sensor sub-device 10 is provided in the underground tunnel and senses environmental information and image information of the underground tunnel.

The disaster management sub-device 20 has a control function of centrally monitoring by displaying information about the configuration installed in the underground utility tunnel on the situation board and recording the situation.

The digital twin model management sub-device 30 uses the image information provided by the sensor sub-device 10 to create and update a virtual space corresponding to the underground utility tunnel, inserts various attributes into the virtual space, searches tagging information, and disaster information. Predict spread and infer risk of management facilities.

Specifically, as shown in Figure 2b, the digital twin model management subunit 30,

A spatial model creation and management function processing unit 31 that generates a virtual space corresponding to an underground utility tunnel from image information provided by the sensor sub-device 10;

a spatial model update function processing unit 32 that detects whether the generated virtual space model is modified and updates the initial virtual space model when the virtual space model is added or changed;

A risk management and decision support function processing unit (33) that calculates the degree of risk of management objects provided in the underground utility tunnel by analyzing information collected on site and disaster spread prediction information; and

It may include a digital twin model creation and management function processing unit 34 that generates a disaster model to which disaster information is applied in a virtual space corresponding to the underground utility tunnel, and provides the generated disaster model to an external user terminal.

The digital archive sub-device 40 stores information produced and processed in the sub-device.

The network sub device 50 provides the created virtual space to the external inspector's user terminal.

According to the digital twin disaster management system customized for underground tunnels according to conventional technology, it provides real-time analysis of information on pre-disaster preparedness in underground tunnels, and in the event of a disaster, provides initial response to disasters and prevents the spread of disasters before rescue teams arrive. Spread prediction information can be obtained.

As a result, it is possible to prepare in advance to significantly reduce the incidence of disasters in normal times, respond effectively to disasters before rescue teams arrive when a disaster occurs, and minimize damage from disasters by sharing response information with each other.

In other words, as described above, in Republic of Korea Patent No. 10-2355769, an NFC tag is installed in the inner space of the tunnel,

You can build a digital twin by capturing video from a drone, receiving an identification code from an NFC tag, and displaying a 3D image.

In addition, in Republic of Korea Patent Publication No. 2017-40652 and Republic of Korea Patent Publication No. 2021-0096392, the signal of the sensor measured in the terminal is imaged as an icon, and the visualization icon is distinguished according to the sensor data,

The center point within the polygon formed by the icon can be detected with the coordinates of the terminal, and information on existing indoor image data or map data and newly acquired image data can be matched, and indoor positioning can be performed using artificial intelligence deep learning.

In addition, Republic of Korea Patent Publication No. 2019-121275 identifies the number of evacuees and the number of escape exits when a disaster situation occurs through location-based information, calculates the evacuation route and evacuation time, and calculates the evacuation route and evacuation time. Through simulation that considers bottlenecks, etc., the optimal evacuation route can be derived.

However, according to the prior art, individual technical configurations of the complex disaster management system have been disclosed, but there is a need for an integrated complex disaster management system that manages them in an integrated manner.

Republic of Korea Patent No. 10-2355769 (registration date: January 21, 2022), title of invention: “Method for building a 3D digital twin using a drone” Republic of Korea Patent Publication No. 2017-40652 (publication date: April 13, 2017), title of invention: “Indoor localization device and indoor localization method based on sensor images” Republic of Korea Patent Publication No. 2019-121275 (publication date: October 25, 2019), title of invention: “Indoor positioning system, device and method” Republic of Korea Patent Publication No. 2021-96392 (publication date: August 5, 2021), title of invention: “Intelligent evacuation guidance system for disaster situations” Republic of Korea Patent Publication No. 2022-71880 (publication date: May 31, 2022), title of invention: “Underground tunnel customized digital twin disaster management system” Republic of Korea Patent No. 10-2212510 (registration date: January 29, 2021), title of invention: “Virtual reality control system”

The technical task to be achieved by the present invention to solve the above-described problems is to build a digital twin for the inside of railway facilities such as railways, tunnels, and underground stations, and to detect abnormalities inside the railway facilities, allowing pedestrians and passers-by to A complex disaster prevention system and method for digital twin-based railway facilities that can virtually simulate the optimal evacuation route in the event of a disaster by identifying indoor and outdoor positioning information through one terminal and guide the optimal evacuation route in the event of an actual disaster. It is intended to provide.

Another technical task to be achieved by the present invention is to prevent and maintain disasters in railway facilities by converging and applying indoor and outdoor positioning technology to digital twin-based railway facilities through a single platform in order to build a complex disaster prevention system for digital twin-based railway facilities. The purpose is to provide a complex disaster prevention system and method for digital twin-based railway facilities that can perform management.

Another technical problem to be achieved by the present invention is to measure and determine altitude changes through a barometer built into a smartphone, which is a client terminal, during indoor and outdoor vertical positioning, and to accurately perform vertical positioning by correcting real-time air pressure changes. , It is intended to provide a complex disaster prevention system and method for digital twin-based railway facilities.

As a means of achieving the above-described technical problem, the digital twin-based complex disaster prevention system for railway facilities according to the present invention is a complex disaster prevention system for railway facilities including railways, tunnels, underground stations, above-ground spaces, and underground spaces. The railway facility digital twin platform generates a virtual space model from railroad facility data and acquires the space model data through sensors installed in the virtual space model; A disaster detection module that detects disasters for each railway facility and generates disaster detection data; A terminal owned by a passenger or occupant within a railroad facility, which performs indoor and outdoor vertical positioning to generate positioning information for the terminal when a disaster occurs in the railroad facility, and generates multiple complex signal images; And collecting the spatial model data and disaster detection data to determine whether a disaster has occurred in a railroad facility, and performing multi-complex signal machine learning on the multi-complex signal images received from each of the client terminals through an AI-based positioning network, It includes a complex disaster prevention server that generates disaster evacuation information for each of the client terminals according to learning results, wherein the client terminal and the complex disaster prevention server form an artificial intelligence (AI)-based location network, and the complex disaster prevention server forms an artificial intelligence (AI)-based location network. The prevention server is characterized in that it generates positioning information corresponding to the indoor and outdoor locations of the client terminal using signals that the client terminal can receive.

Here, during indoor/outdoor vertical positioning of the client terminal, altitude changes are measured and positioned using a barometer built into the client terminal, and a relative barometric pressure map is constructed by applying an EM (Expectation and Maximization) algorithm to correct real-time air pressure changes. And, when the rate of change of the barometric pressure value of the barometer is greater than a threshold, vertical positioning is performed by estimating the number of barometer floors.

Here, the railway facility digital twin platform includes a spatial model generator that generates a spatial model for a virtual space corresponding to railway facility data; And it may include a spatial model data acquisition unit that acquires spatial model data from sensors installed according to the spatial model.

Here, the disaster detection module includes a tunnel disaster detection unit that detects tunnel disasters including tunnel earthquakes, gas leaks, and tunnel flooding; And it may include an underground station fire detection unit that detects an underground station fire.

Here, the tunnel disaster detection unit may detect a tunnel earthquake using a pre-installed vibration sensor and an acoustic sensor, detect a gas leak using a pre-installed gas sensor, and detect tunnel flooding using a pre-installed water level sensor.

Here, the underground station fire detection unit may detect an underground station fire from at least one sensor selected from CCTV images, temperature sensors/humidity sensors, smoke sensors, beacons, or sound sensors.

Here, the client terminal is a sensor that measures the location of the client terminal, and a sensor for generating positioning information that measures the location of the terminal in response to disaster information received from the complex disaster prevention server: and positioning according to the location of the client terminal. It may include a positioning information generator that generates information.

Here, the positioning information generator performs hash-based positioning on the location detected through the positioning information generation sensor installed in the terminal, and then sets the color, and then performs a series of data conversion processes to generate a multi-complex signal image. By performing , multiple composite signal images can be generated as indoor and outdoor positioning signals.

Here, the complex disaster prevention server forms an AI-based positioning network with the client terminal, transmits disaster information to the client terminal when a disaster occurs for railway facilities, and transmits multiple complex signal images for indoor and outdoor location measurement from the client terminal. AI-based location network forming unit that receives; A data collection unit that collects spatial model data for railway facilities from the digital twin platform and collects disaster detection data from the disaster detection module; A disaster determination unit for railway facilities that determines whether a disaster has occurred according to the spatial model data and disaster detection data collected by the data collection unit; a multi-complex signal machine learning unit that performs machine learning according to a learning algorithm established on the multi-complex signal image received from the AI-based positioning network forming unit; And it may include a disaster evacuation information generation unit that generates disaster evacuation information for each of the client terminals according to the learning results of the multi-complex signal machine learning unit.

Here, the learning algorithm of the multi-complex signal machine learning unit is preferably a PNN (Positioning Neural Network) algorithm.

Here, the disaster evacuation information generation unit may generate disaster evacuation information according to a disaster evacuation scenario established in connection with a disaster evacuation guidance system.

Meanwhile, as another means of achieving the above-described technical problem, the method for preventing complex disasters in digital twin-based railway facilities according to the present invention is to prevent complex disasters in railway facilities including railways, tunnels, underground stations, above-ground spaces, and underground spaces. In the disaster prevention method, a) the digital twin platform creates a spatial model by modeling a virtual space according to railway facility data, and obtaining spatial model data; b) a disaster detection module detecting a disaster within a railway facility and generating disaster detection data; c) a complex disaster prevention server collecting spatial model data for railway facilities detected through a virtual space sensor installed on the digital twin platform, and collecting disaster detection data detected by the disaster detection module; d) a complex disaster prevention server determining a disaster of a railroad facility according to the collected spatial model data and disaster detection data; e) If it is determined that a disaster has occurred at a railway facility, receiving multiple complex signal images from a client terminal through the AI-based positioning network of the complex disaster prevention server; f) a multi-complex signal machine learning unit of the complex disaster prevention server performing machine learning on the multi-complex signal image received from the client terminal according to a built-in learning algorithm; and g) generating disaster evacuation information for each client terminal according to the learning results of the multiple complex signal machine learning unit, wherein the client terminal and the complex disaster prevention server form an AI-based positioning network, Characterized by generating positioning information corresponding to the indoor and outdoor locations of the client terminal using signals that the client terminal can receive.

According to the present invention, a digital twin is built for the inside of railway facilities such as railways, tunnels, and underground stations, and managers can easily detect abnormalities inside the railway facilities, while providing indoor and outdoor positioning information through client terminals such as pedestrians and passersby. By accurately identifying and virtually simulating the optimal evacuation route in the event of a disaster, it is possible to guide the optimal evacuation route in the event of an actual disaster.

According to the present invention, in order to build a complex disaster prevention system for digital twin-based railway facilities, disaster prevention and maintenance of railway facilities are easily performed by converging and applying indoor and outdoor positioning technology to digital twin-based railway facilities through a single platform. can do.

According to the present invention, during indoor and outdoor vertical positioning, the altitude change is measured and positioned using a barometer built into a smartphone, which is a client terminal, and the vertical positioning can be accurately performed by correcting the real-time air pressure change.

Figure 1 is a diagram showing the configuration of a general conventional digital twin platform.
Figure 2a is a configuration diagram to explain a digital twin disaster management system customized for underground tunnels according to conventional technology, and Figure 2b is a specific configuration diagram of the digital twin model management sub-device shown in Figure 2a.
Figure 3 is a diagram schematically explaining the concept of a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention.
Figure 4 is a configuration diagram of a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention.
Figure 5 is a screen showing the digital twin platform in the complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention.
Figure 6 is a screen showing the disaster detection module of the complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention detecting a tunnel disaster situation.
FIG. 7 is a diagram illustrating detection of tunnel flooding and tunnel earthquake as the tunnel disaster situation shown in FIG. 6.
Figure 8 is a screen showing the disaster detection module of the complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention detecting a fire in an underground station.
Figure 9 is a screen illustrating the collection of indoor and outdoor positioning information in a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention.
Figure 10 is a diagram showing the concept of AI-based positioning for fusing multiple complex signals in a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention.
Figure 11 is a diagram showing a list of indoor and outdoor multiple composite signals in a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention.
Figure 12 is a diagram showing a multiple complex signal conversion process in a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention.
Figure 13 is a diagram illustrating a method for generating multiple complex signal images in a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention.
Figure 14 is a diagram showing multiple complex signal images generated in a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention.
Figure 15 is a diagram showing the structure of a PNN applied to machine learning of multiple complex signals in a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention.
Figure 16 is a diagram illustrating measuring air pressure in a pair of two layers to measure vertical positioning in a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention.
Figure 17 is a diagram illustrating an air pressure pair merging algorithm for measuring vertical positioning applied to a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention.
Figure 18 is a diagram illustrating optimization based on the EM (Expectation and Maximization) algorithm for correcting changes in air pressure during vertical positioning in the complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention.
Figure 19 is a diagram for explaining the operation of estimating the number of floors of the barometer during vertical positioning in the complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention.
Figure 20 is an operation flowchart showing a complex disaster prevention method for digital twin-based railway facilities according to an embodiment of the present invention.

Below, with reference to the attached drawings, embodiments of the present invention will be described in detail so that those skilled in the art can easily implement the present invention. However, the present invention may be implemented in many different forms and is not limited to the embodiments described herein. In order to clearly explain the present invention in the drawings, parts unrelated to the description are omitted, and similar parts are given similar reference numerals throughout the specification.

Throughout the specification, when a part is said to “include” a certain element, this means that it may further include other elements rather than excluding other elements, unless specifically stated to the contrary. Additionally, terms such as “… unit” used in the specification refer to a unit that processes at least one function or operation, and may be implemented as hardware, software, or a combination of hardware and software.

[Complex disaster prevention system for digital twin-based railway facilities (100)]

Figure 3 is a diagram schematically illustrating the concept of a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention, and Figure 4 is a diagram for complex disaster prevention of digital twin-based railway facilities according to an embodiment of the present invention. This is the configuration diagram of the system.

Referring to Figures 3 and 4, the complex disaster prevention system 100 for digital twin-based railway facilities according to an embodiment of the present invention is,

In the complex disaster prevention system for railway facilities including railways, tunnels, underground stations, above-ground and underground spaces, the railway facility digital twin platform (110), disaster detection module (120), client terminal (130), and complex disaster It is configured to include a prevention server 140, At this time, the complex disaster prevention server 140 generates integrated disaster evacuation information in connection with the disaster evacuation guidance system 200.

The railroad facility digital twin platform 110 creates a virtual space model from railroad facility data and acquires space model data through sensors installed in the virtual space model.

Specifically, the railway facility digital twin platform 110 includes a spatial model generator 111 that generates a spatial model for a virtual space corresponding to railway facility data; and

It may include, but is not limited to, a spatial model data acquisition unit 112 that acquires spatial model data from sensors installed according to the spatial model.

The disaster detection module 120 generates disaster detection data by detecting disasters for each railway facility, including railways, tunnels, underground stations, above-ground spaces, and underground spaces.

Specifically, the disaster detection module 120,

a tunnel disaster detection unit 121 that detects tunnel disasters including tunnel earthquakes, gas leaks, and tunnel flooding; and an underground station fire detection unit 122 that detects an underground station fire, but is not limited thereto.

For example, as shown in FIG. 3, the tunnel disaster detection unit 121 detects a tunnel earthquake by a pre-installed vibration sensor and an acoustic sensor, detects a gas leak by a pre-installed gas sensor, and detects a gas leak by a pre-installed gas sensor. Disaster detection data can be generated by detecting tunnel flooding using a water level sensor.

In addition, as shown in FIG. 3, the underground station fire detection unit 122 detects an underground station fire from at least one sensor selected from CCTV images, temperature sensors/humidity sensors, smoke sensors, beacons, or sound sensors to detect disasters. Sensing data can be generated.

The client terminal 130 is a terminal owned by a passerby or occupant of a railroad facility, and generates multiple complex signal images to generate positioning information for the terminal when a disaster occurs in the railroad facility.

Here, the method of generating the multiple composite signal image will be described later with reference to FIGS. 11 to 14.

Specifically, the client terminal 130 is a sensor that measures the location of the client terminal, and a sensor 131 for generating positioning information that measures the location of the terminal in response to disaster information received from the complex disaster prevention server 140. ): and a positioning information generator 132 that generates positioning information according to the location of the client terminal, but is not limited to this.

At this time, the positioning information generator 132 performs hash-based spatialization on the location detected through the positioning information generation sensor 131 installed in the terminal, and then sets the color. ), then a series of data conversion processes to generate multiple composite signal images can be performed to generate multiple composite signal images as indoor and outdoor positioning signals.

The complex disaster prevention server 140 collects the spatial model data and disaster detection data to determine whether a disaster has occurred at a railroad facility,

Perform multiple complex signal machine learning on multiple complex signal images received from each of the client terminals 130 through an AI-based positioning network, and generate disaster evacuation information for each of the client terminals 130 according to the learning results. do.

That is, in the case of a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention, the client terminal 130 and the complex disaster prevention server 140 form an AI-based positioning network, and the client terminal Positioning information corresponding to the indoor and outdoor locations of the client terminal 130 can be generated using signals that 130 can receive.

Specifically, the complex disaster prevention server 140 includes an AI-based positioning network forming unit 141, a data collection unit 142, a railway facility disaster determination unit 143, a multi-complex signal machine learning unit 144, and It may include a disaster evacuation information generation unit 145.

The AI-based positioning network forming unit 141 forms an AI-based positioning network with the client terminal 130, and transmits disaster information to the client terminal 130 when a disaster occurs at a railroad facility and the client terminal ( Receive multiple complex signal images for indoor and outdoor position measurement from 130).

The data collection unit 142 collects spatial model data for railway facilities from the railway facility digital twin platform 110 and collects disaster detection data from the disaster detection module 120.

The railway facility disaster determination unit 143 determines whether a disaster has occurred according to the spatial model data and disaster detection data collected by the data collection unit 142.

The multiple complex signal machine learning unit 144 performs machine learning on the multiple complex signal images received from the AI-based positioning network forming unit 141 according to a built-in learning algorithm.

Disaster evacuation information for each of the client terminals 130 is generated according to the learning results of the multi-complex signal machine learning unit 144.

At this time, the disaster evacuation information generation unit 145 may generate disaster evacuation information according to a disaster evacuation scenario established in connection with the disaster evacuation guidance system 200.

In other words, in the case of the complex disaster prevention system 100 for digital twin-based railway facilities according to an embodiment of the present invention,

Establish a digital twin for the inside of railway facilities such as railways, tunnels, and underground stations, and allow managers to easily detect abnormalities inside the railway facilities, while providing indoor and outdoor positioning information through client terminals 130 for pedestrians and passersby. It can accurately identify and virtually simulate the optimal evacuation route in the event of a disaster, allowing it to be linked to the optimal evacuation route in the event of an actual disaster.

Accordingly, the digital twin-based complex disaster prevention system 100 for railway facilities according to an embodiment of the present invention is a digital twin-based integrated disaster management complex system for railway facilities and railway SOC through a single platform converged. Disaster prevention and maintenance can be easily performed.

Meanwhile, Figure 5 is a screen showing the digital twin platform in the complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention.

As shown in Figure 5, the complex disaster prevention system 100 for digital twin-based railway facilities according to an embodiment of the present invention is,

By building a digital twin platform for railways, underground stations, tunnels, underground spaces, or above-ground spaces, a virtual space model is built by collecting object-specific data needed to express railway facility objects such as railways, underground stations, tunnels, etc. However,

At this time, railroad facility data is input through spatial photography and image construction using drones or sensors. Afterwards, the location of data for each object is identified on the virtual space model, and then data for each object is generated based on the identified location. A digital twin virtual space can be built through connection to a data hub.

Meanwhile, Figure 6 is a screen showing that the disaster detection module of the complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention detects a tunnel disaster situation, and Figure 7 is the tunnel disaster situation shown in Figure 6. This is a diagram illustrating detection of tunnel flooding and tunnel earthquakes.

As shown in FIG. 6, the disaster detection module 120 of the complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention detects tunnel disasters including tunnel earthquakes, gas leaks, and tunnel flooding. You can,

For example, as shown in b) of FIG. 7, a tunnel earthquake is detected by a pre-installed vibration sensor and an acoustic sensor, a gas leak is detected by a pre-installed gas sensor, and also, as shown in a) of FIG. 7. As shown, disaster detection data can be generated by detecting tunnel flooding using a pre-installed water level sensor.

Meanwhile, Figure 8 is a screen showing the disaster detection module of the complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention detecting a fire in an underground station.

As shown in Figure 8, in the case of the disaster detection module 120 of the complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention,

Fires in underground stations can be detected, and for example, disaster detection data can be generated by detecting fires in underground stations from at least one sensor selected from CCTV images, temperature sensors/humidity sensors, smoke sensors, beacons, or sound sensors.

Meanwhile, Figure 9 is a screen illustrating the collection of indoor and outdoor positioning information in a complex disaster prevention system for a digital twin-based railway facility according to an embodiment of the present invention, and Figure 10 is a screen illustrating a digital twin-based railway facility according to an embodiment of the present invention. This is a diagram showing the concept of AI-based positioning for fusing multiple complex signals in a complex disaster prevention system.

Referring to Figure 9, in the case of a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention, optimized indoor and outdoor positioning information for railway facilities such as railways, underground stations, tunnels, underground spaces, and above-ground spaces is provided. can be collected,

At this time, as shown in FIG. 10, horizontal positioning and vertical positioning can be performed using only the client terminal 130, for example, a smartphone, without installing separate equipment using artificial intelligence.

In particular, high-precision positioning is possible using all available signals that can be received indoors, and in this case, multiple complex signals are fused based on artificial intelligence (AI). In other words, horizontal and vertical positioning are performed using all signals that the smartphone can receive.

At this time, by generating positioning information according to an artificial intelligence (AI)-based positioning network rather than the existing simple triangulation, indoor and outdoor locations are measured with a 1m level error, and location information is collected at low cost and applied to various environments. can do.

Meanwhile, Figure 11 is a diagram showing a list of indoor and outdoor multiple complex signals in the complex disaster prevention system of digital twin-based railway facilities according to an embodiment of the present invention, and Figure 12 is a diagram showing a list of indoor and outdoor multiple complex signals of digital twin-based railway facilities according to an embodiment of the present invention. This is a diagram showing the multiple complex signal conversion process in a complex disaster prevention system.

The digital twin-based complex disaster prevention system 100 for railway facilities according to an embodiment of the present invention can analyze multiple complex signals of railway facilities. At this time, some of the indoor signals can be received even outdoors,

Additionally, some of the outdoor signals may be received indoors. Here, Figure 11 shows a list of indoor and outdoor multiple composite signals, and the reception ranges of the signals can be distinguished.

In addition, accuracy can be guaranteed by treating indoor spatial information such as windows, doors, pillars, and elevators as important metadata when determining indoor location.

In addition, the complex disaster prevention system 100 for digital twin-based railway facilities according to an embodiment of the present invention,

As shown in Figure 11, multiple complex signals from railway facilities can be fused. At this time, various types of signals (indoor/outdoor radio signals, sensor measurements, etc.) available at a specific point in time are converted into signal images. .

Figure 12 shows the multiple composite signal conversion process, and one composite signal image can be obtained by arranging the data converted into several pixels by setting the position on one image.

Subsequently, the converted images are learned through deep learning, and then when indoor positioning is requested later, the n most similar images are found and the distance between each image is calculated to determine positioning.

For example, signals measured from Wi-Fi, Mobile Network, Bluetooth, GPS/GLONASS, geomagnetic sensor, acceleration sensor, and barometer can be used to create multiple composite signal images representing a specific location.

Meanwhile, FIG. 13 is a diagram illustrating a method for generating multiple complex signal images in a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention, and FIG. 14 is a diagram for explaining a method of generating multiple complex signal images for a digital twin-based railway according to an embodiment of the present invention. This is a diagram showing multiple complex signal images generated by the facility’s complex disaster prevention system.

As shown in Figure 13, in the case of a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention, multiple complex signal images can be generated through the following process.

Specifically, multiple complex signals are viewed as a pair of <Key, Value> and visualized. For example, the MAC of Wi-Fi is considered the key, and the reception strength (RSSI) value is considered the value.

At this time, the location of the key within the image can be determined using a hash-based spatialization technique.

In addition, the value can determine the color and intensity in the image through normalization and coloring techniques, and accordingly, a multi-complex signal image as shown in FIG. 14 can be created.

Meanwhile, Figure 15 is a diagram showing the structure of a PNN applied to machine learning of multiple complex signals in a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention.

In the case of a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention, machine learning of multiple complex signals is performed. For example, as shown in FIG. 15, artificial intelligence for learning multiple complex signals is used. As an intelligent indoor positioning model, it uses the PNN (Positioning Neural Network) algorithm.

For example, multiple composite signals received from the client terminal 130 can be converted into images (constellations), where Euclidean distance is a commonly used method for calculating the distance between two points.

This distance can be used to define Euclidean space.

Specifically, in the complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention, the characteristics of multiple complex signal images and the specificity of PNN implementation accordingly require a learning algorithm including an error back-propagation function. It can be implemented very simply.

In addition, the PNN algorithm actually has a three-dimensional transfer function with an output that decreases depending on the size and difference between the two input values, and the connection strength or weight of the connections that make up the neural network itself are all fixed at the same value. , an approach is possible by applying parameters that tune the transfer function itself.

The transfer function of the existing artificial neural network uses fixed non-linear functions such as Sigmoid and relu, and learning can be implemented by controlling the connection weight of each neural node through learning. However, in the case of the PNN algorithm, the connection weight of each node is Since everything is fixed, it can be implemented by defining a 3D transfer function with four parameters and using an error back-propagation function that tunes this transfer function according to the learning results.

Meanwhile, Figure 16 is a diagram illustrating measuring air pressure in a pair of two layers to measure vertical position in a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention, and Figure 17 is This is a diagram illustrating an air pressure pair merging algorithm for measuring vertical positioning applied to a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention.

In the case of the digital twin-based complex disaster prevention system for railway facilities according to an embodiment of the present invention, vertical positioning can be achieved through the following process by reflecting the optimized positioning of the client terminal with respect to the railway facility.

At this time, compared to the typical scenario shown in b) of FIG. 16, it is possible to collect air pressure pairs at the lowest cost according to the best scenario shown in a) of FIG. 16.

Specifically, as shown in FIG. 17, multiple composite signal images of railway facilities are collected using a client terminal 130, for example, a smartphone, and then the collected multiple composite signal images are themselves Separately from the location, it also has a location provided to the Android system from that location,

This information is recorded as a collection of pixels that can be restored in a specific area of each multiple composite signal image, so that location information, that is, positioning information, where the multiple composite signal image was created can be known.

At this time, positioning is possible through the client terminal 130 by normalizing multiple complex signals of various types and imaging them. In particular, 100% vertical accuracy is achieved by constructing a relative pressure map based on the pressure pair. Vertical positioning is possible.

Meanwhile, Figure 18 is a diagram illustrating EM algorithm-based optimization for correcting air pressure changes during vertical positioning in a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention, and Figure 19 is an implementation of the present invention. This is a drawing to explain the floor number estimation operation of the barometer during vertical positioning in the complex disaster prevention system of digital twin-based railway facilities according to an example.

The barometer built into a typical smartphone can measure altitude changes with a precision of 30 cm or more, but since the barometric pressure changes depending on the weather, correction for changes in barometric pressure is necessary.

In the case of a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention, as shown in FIG. 18, it can be optimized according to the EM (Expectation and Maximization) algorithm to correct real-time atmospheric pressure changes.

especially. As shown in FIG. 19, the number of barometer floors can be estimated when the rate of change of the barometric pressure value of the smartphone built-in barometer is above the threshold value based on the relative pressure map.

Accordingly, in the case of a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention, during indoor and outdoor vertical positioning, the altitude change is measured and positioned using a barometer built into the smartphone, which is a client terminal,

By constructing a relative barometric pressure map by applying the EM algorithm to correct real-time barometric pressure changes, accurate vertical positioning is possible by estimating the number of barometer floors when the barometric pressure value change rate of the smartphone's built-in barometer is above the threshold.

In other words, in the case of a complex disaster prevention system for digital twin-based railroad facilities according to an embodiment of the present invention, spatial photography and image construction of railroad facilities are performed using drones or sensors, etc.

After identifying the location of data for each object on the spatial model, a virtual space with a digital twin can be built by connecting the data for each object and the data hub based on the identified location.

In addition, in the case of a digital twin-based complex disaster prevention system for railway facilities according to an embodiment of the present invention, various signals from inside and outside railway facilities such as railways, subway stations, and tunnels are collected, and various types of multiple complex signals are converted into images. And

Afterwards, the converted images are subjected to deep learning learning through the PNN artificial intelligence indoor positioning model, and the indoor location can be determined by estimating the distance through comparison with the collected multi-complex signal images and previously collected images.

In particular, in the case of vertical positioning, altitude changes are measured and determined through the smartphone's built-in barometer, and a relative pressure map is constructed by applying an EM algorithm to correct real-time air pressure changes, so that the rate of change in the barometric pressure value of the smartphone's built-in barometer is When it is above the threshold, vertical positioning is performed by estimating the number of barometer floors.

In addition, in the case of a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention, in the virtual space created by digital twin for railway facilities such as railways, subway stations, and tunnels,

In conjunction with the above-mentioned horizontal and vertical positioning technologies, it is possible to predict and simulate the spread of disasters such as earthquakes and fires. At this time, the number of occupants or pedestrians is identified, and the number of pedestrians, shortest escape route, time required for escape, and escape Integrated simulation can be performed considering path stability, etc.

In addition, in the case of a complex disaster prevention system for digital twin-based railway facilities according to an embodiment of the present invention, it is possible to predict and simulate the spread of disaster situations by constructing a virtual space by converting railway facilities into digital twins.

Through indoor positioning technology, the location of indoor occupants or passers-by can be accurately determined, and through integrated linkage with the disaster evacuation guidance system, the optimal evacuation route in the event of an actual disaster can be predicted through simulation of various evacuation routes in disaster situations, and ultimately, disaster relief. The situation can be prevented.

[How to prevent complex disasters in digital twin-based railway facilities]

Figure 20 is an operation flowchart showing a method for preventing complex disasters in digital twin-based railway facilities according to an embodiment of the present invention.

Referring to Figure 20, the complex disaster prevention method for digital twin-based railway facilities according to an embodiment of the present invention is a complex disaster prevention system for railway facilities including railways, tunnels, underground stations, above-ground spaces, and underground spaces. ,

First, the railway facility digital twin platform 110 models the virtual space according to the railway facility data, creates a spatial model, and acquires the spatial model data (S110).

Next, the disaster detection module 120 detects a complex disaster within the railroad facility, for example, a tunnel disaster situation or an underground station fire, and generates disaster detection data (S120).

Next, the complex disaster prevention server 140 provided in the comprehensive control room shown in FIG. 3 collects spatial model data for the railroad facility detected through the virtual space sensor installed on the railroad facility digital twin platform 110, Disaster detection data detected by the disaster detection module 120 is collected (S130).

Next, the complex disaster prevention server 140 determines the disaster of the railroad facility according to the collected spatial model data and disaster detection data (S140).

Next, when it is determined that a disaster has occurred at a railroad facility, multiple complex signal images are received from the client terminal 130 through the AI-based positioning network of the complex disaster prevention server 140 (S150).

Here, the AI-based location network is created by the AI-based location network forming unit 141 of the complex disaster prevention server 140, and the client terminal 130 through the AI-based location network,

For example, location information of passers-by, occupants, or workers is collected through smartphones. At this time, multiple complex signal images are generated using all signals that the smartphone can receive, and the complex disaster prevention server 140 collects the multiple complex signal images.

Specifically, the client terminal 130 generates multiple composite signal images by performing indoor and outdoor vertical positioning to generate positioning information for the terminal when a disaster occurs at a railroad facility,

At this time, during indoor and outdoor vertical positioning of the client terminal 130, the altitude change is measured and positioned using a barometer built into the client terminal 130, and an EM (Expectation and Maximization) algorithm is applied to correct the real-time air pressure change. Vertical positioning is achieved by constructing a relative pressure map and estimating the number of floors of the barometer when the rate of change of the barometric pressure value of the barometer is greater than a threshold.

Next, the multiple complex signal machine learning unit 144 of the complex disaster prevention server 140 performs machine learning according to a built-in learning algorithm for the multiple complex signal image received from the client terminal 130. Perform (S160).

Next, according to the learning results of the multiple complex signal machine learning unit 144, disaster evacuation information for each of the client terminals 130 is generated in connection with the disaster evacuation guidance system (S160) (S170).

Ultimately, according to an embodiment of the present invention, a digital twin is built for the inside of railway facilities such as railways, tunnels, and underground stations, and the manager can easily detect abnormalities inside the railway facilities, while client terminals such as pedestrians and passers-by are monitored. Through this, it is possible to accurately determine indoor and outdoor positioning information and virtually simulate the optimal evacuation route in the event of a disaster, leading to the optimal evacuation route in the event of an actual disaster.

In addition, in order to build a complex disaster prevention system for digital twin-based railway facilities, disaster prevention and maintenance of railway facilities can be easily performed by converging and applying indoor and outdoor positioning technology to digital twin-based railway facilities through a single platform. .

In addition, during indoor and outdoor vertical positioning, the altitude change is measured and positioned using a barometer built into the smartphone, which is a client terminal, and the vertical positioning can be performed accurately by correcting the real-time air pressure change.

The description of the present invention described above is for illustrative purposes, and those skilled in the art will understand that the present invention can be easily modified into other specific forms without changing the technical idea or essential features of the present invention. will be. Therefore, the embodiments described above should be understood in all respects as illustrative and not restrictive. For example, each component described as single may be implemented in a distributed manner, and similarly, components described as distributed may also be implemented in a combined form.

The scope of the present invention is indicated by the claims described below rather than the detailed description above, and all changes or modified forms derived from the meaning and scope of the claims and their equivalent concepts should be construed as being included in the scope of the present invention. do.

100: Complex disaster prevention system for railway facilities
200: Disaster evacuation guidance system
110: Railway facility digital twin platform 120: Disaster detection module
130: Client terminal (smart phone) 140: Complex disaster prevention server
111: Spatial model creation unit 112: Spatial model data acquisition unit
121: Tunnel disaster detection unit 122: Underground station fire detection unit
131: Sensor for generating positioning information 132: Positioning information generating unit
141: AI-based positioning network formation unit 142: Data collection unit
143: Railway facility disaster determination unit 144: Multi-complex signal machine learning unit
145: Disaster evacuation information generation unit

Claims (19)

In the complex disaster prevention system for railway facilities including railways, tunnels, underground stations, above-ground spaces and underground spaces,
A railroad facility digital twin platform 110 that generates a virtual space model from railroad facility data and acquires space model data through sensors installed in the virtual space model;
A disaster detection module 120 that detects disasters for each railway facility and generates disaster detection data;
A terminal owned by a passenger or occupant within a railroad facility, which performs indoor and outdoor vertical positioning to generate positioning information for the terminal when a disaster occurs in the railroad facility, and generates a multiple complex signal image (client terminal 130); and
Collect the spatial model data and disaster detection data to determine whether a disaster has occurred at a railroad facility, and perform multi-complex signal machine learning on the multi-complex signal images received from each of the client terminals 130 through an AI-based positioning network. and a complex disaster prevention server 140 that generates disaster evacuation information for each of the client terminals 130 according to learning results,
The client terminal 130 and the complex disaster prevention server 140 form an AI-based positioning network, and the complex disaster prevention server 140 uses a signal that the client terminal 130 can receive to A complex disaster prevention system for digital twin-based railway facilities characterized by generating positioning information corresponding to the indoor and outdoor positions of the client terminal 130. According to paragraph 1,
During indoor/outdoor vertical positioning of the client terminal 130, the altitude change is measured and positioned using a barometer built into the client terminal 130, and an EM (Expectation and Maximization) algorithm to correct real-time air pressure changes is applied to determine relative positioning. A complex disaster prevention system for digital twin-based railway facilities, characterized in that vertical positioning is achieved by building a barometric pressure map and estimating the number of floors of the barometer when the rate of change of the barometric pressure value of the barometer is greater than a threshold. According to paragraph 1,
The railway facility digital twin platform 110 is,
A spatial model generator 111 that generates a spatial model for a virtual space corresponding to railway facility data; and
A complex disaster prevention system for a digital twin-based railway facility including a spatial model data acquisition unit 112 that acquires spatial model data from sensors installed according to the spatial model. According to paragraph 1,
The disaster detection module 120,
a tunnel disaster detection unit 121 that detects tunnel disasters including tunnel earthquakes, gas leaks, and tunnel flooding; and
A complex disaster prevention system for digital twin-based railway facilities that includes an underground station fire detection unit (122) that detects underground station fires. According to paragraph 4,
The tunnel disaster detection unit 121 detects a tunnel earthquake using a pre-installed vibration sensor and an acoustic sensor, detects a gas leak using a pre-installed gas sensor, and detects tunnel flooding using a pre-installed water level sensor. A complex disaster prevention system for railway facilities based on digital twin. According to paragraph 4,
The underground station fire detection unit 122 is a complex disaster of a digital twin-based railway facility, characterized in that it detects an underground station fire from at least one sensor selected from CCTV video, temperature sensor/humidity sensor, smoke sensor, beacon, or sound sensor. prevention system. According to paragraph 1,
The client terminal 130,
A sensor for measuring the location of a client terminal, a sensor 131 for generating positioning information that measures the location of the terminal in response to disaster information received from the complex disaster prevention server 140: and
A complex disaster prevention system for digital twin-based railway facilities that includes a positioning information generator 132 that generates positioning information according to the location of the client terminal. In clause 7,
The positioning information generator 132 performs hash-based spatialization on the location detected through the positioning information generation sensor 131 installed in the terminal, and then sets the color. A complex disaster prevention system for digital twin-based railway facilities is characterized by generating multiple complex signal images as indoor and outdoor positioning signals by performing a series of data conversion processes to generate multiple complex signal images. According to paragraph 1,
The complex disaster prevention server 140,
Forms an AI-based positioning network with the client terminal 130, transmits disaster information to the client terminal 130 when a disaster occurs in a railroad facility, and multiple complex signal images for indoor and outdoor location measurement from the client terminal 130. AI-based positioning network forming unit 141 that receives;
A data collection unit 142 that collects spatial model data for railway facilities from the railway facility digital twin platform 110 and collects disaster detection data from the disaster detection module 120;
A railway facility disaster determination unit 143 that determines whether a disaster has occurred according to the spatial model data and disaster detection data collected by the data collection unit 142;
a multi-complex signal machine learning unit 144 that performs machine learning according to a learning algorithm established on the multi-complex signal image received from the AI-based positioning network forming unit 141; and
Complex disaster prevention of digital twin-based railway facilities including a disaster evacuation information generation unit 145 that generates disaster evacuation information for each of the client terminals 130 according to the learning results of the multi-complex signal machine learning unit 144. system. According to clause 9,
The learning algorithm of the multiple complex signal machine learning unit 144 is a digital twin-based railway facility complex disaster prevention system, characterized in that it is a PNN (Positioning Neural Network) algorithm. According to clause 9,
The disaster evacuation information generation unit 145 is a complex disaster prevention system for digital twin-based railway facilities, characterized in that it generates disaster evacuation information according to a disaster evacuation scenario established in connection with the disaster evacuation guidance system 200. In the method of preventing complex disasters for railway facilities including railways, tunnels, underground stations, above-ground spaces and underground spaces,
a) the railroad facility digital twin platform 110 models the virtual space according to the railroad facility data to create a spatial model and acquires the spatial model data;
b) the disaster detection module 120 detects a disaster within a railway facility and generates disaster detection data;
c) The complex disaster prevention server 140 collects spatial model data for railroad facilities detected through the virtual space sensor installed on the railroad facility digital twin platform 110, and the disaster detected by the disaster detection module 120. collecting sensing data;
d) the complex disaster prevention server 140 determining a disaster of a railroad facility according to the collected spatial model data and disaster detection data;
e) when it is determined that a disaster has occurred at a railroad facility, receiving multiple composite signal images from the client terminal 130 through the AI-based positioning network of the composite disaster prevention server 140;
f) the multi-complex signal machine learning unit 144 of the complex disaster prevention server 140 performing machine learning on the multi-complex signal image received from the client terminal 130 according to a built-in learning algorithm; and
g) generating disaster evacuation information for each of the client terminals 130 according to the learning results of the multi-complex signal machine learning unit 144,
The client terminal 130 and the complex disaster prevention server 140 form an AI-based positioning network, and use signals that the client terminal 130 can receive to determine the indoor and outdoor locations of the client terminal 130. A complex disaster prevention method for digital twin-based railway facilities characterized by generating corresponding positioning information. According to clause 12,
In step e), the client terminal 130 performs indoor and outdoor vertical positioning to generate positioning information for the terminal when a disaster occurs in the railway facility, thereby generating a multiple composite signal image. How to prevent disaster. According to clause 13,
During indoor/outdoor vertical positioning of the client terminal 130, the altitude change is measured and positioned using a barometer built into the client terminal 130, and an EM (Expectation and Maximization) algorithm to correct real-time air pressure changes is applied to determine relative positioning. A complex disaster prevention method for a digital twin-based railway facility, characterized in that vertical positioning is achieved by constructing a barometric pressure map and estimating the number of floors of the barometer when the rate of change of the barometric pressure value of the barometer is greater than a threshold. According to clause 12,
The disaster detection module 120 of step b) includes a tunnel disaster detection unit 121 that detects tunnel disasters including tunnel earthquakes, gas leaks, and tunnel flooding; and a complex disaster prevention method for digital twin-based railway facilities including an underground station fire detection unit (122) that detects underground station fires. According to clause 15,
The tunnel disaster detection unit 121 detects a tunnel earthquake using a pre-installed vibration sensor and an acoustic sensor, detects a gas leak using a pre-installed gas sensor, detects tunnel flooding using a pre-installed water level sensor, and detects tunnel flooding using a pre-installed water level sensor. The station fire detection unit 122 is a digital twin-based complex disaster prevention method for railway facilities, characterized in that it detects an underground station fire from at least one sensor selected from CCTV footage, temperature sensor/humidity sensor, smoke sensor, beacon, or sound sensor. . According to clause 12,
The complex disaster prevention server 140,
Forms an AI-based positioning network with the client terminal 130, transmits disaster information to the client terminal 130 when a disaster occurs in a railroad facility, and multiple complex signal images for indoor and outdoor location measurement from the client terminal 130. AI-based positioning network forming unit 141 that receives;
A data collection unit 142 that collects spatial model data for railway facilities from the railway facility digital twin platform 110 and collects disaster detection data from the disaster detection module 120;
A railway facility disaster determination unit 143 that determines whether a disaster has occurred according to the spatial model data and disaster detection data collected by the data collection unit 142;
a multi-complex signal machine learning unit 144 that performs machine learning according to a learning algorithm established on the multi-complex signal image received from the AI-based positioning network forming unit 141; and
Complex disaster prevention of digital twin-based railway facilities including a disaster evacuation information generation unit 145 that generates disaster evacuation information for each of the client terminals 130 according to the learning results of the multi-complex signal machine learning unit 144. method. According to clause 17,
A method for preventing complex disasters in digital twin-based railway facilities, wherein the learning algorithm of the multiple complex signal machine learning unit 144 is a PNN (Positioning Neural Network) algorithm. According to clause 12,
A complex disaster prevention method for a digital twin-based railway facility, characterized in that disaster evacuation information for each of the client terminals (130) is generated in connection with the disaster evacuation guidance system (S160) in step g).