KR101853700B1 - Indoor localization system in disaster relief and localization method thereof - Google Patents

Abstract

The present invention relates to aindoor localization system and method during disaster relief, which can carry out exact localization by recognizing an indoor position if existing communication facilities are interrupted, and an indoor environment is changed (damaged) in a disaster situation. The indoor localization system comprises: a disaster victim terminal as a fixed node; a plurality of rescuer terminals as a plurality of mobile nodes, which can communicate with the disaster victim terminal; a localization server which divides a lower region with respect to the indoor environment, determines a position of the disaster victim terminal with the highest accuracy among positions reported from all the rescuer terminals as the final position of the disaster victim terminal; a database which divides the inside of a building into a latticed target region by the localization server, divides the target region into a lower region in accordance with the number and type of walls between the region and each latticed point, and stores information with respect to the indoor environment realized by separately forming the gathering of reference points (RP) with respect to each latticed point. According to the present invention, therefore, costs of an off-line step may be reduced in comparison to a conventional fingerprint-based method.

Description

[0001] INDOOR POSITION DISPLAY SYSTEM AND METHOD FOR DISPLACEMENT IN DISASTER RESONANCE [0002]

The present invention relates to a system and method for displaying an indoor location at the time of a disaster, and more particularly, it relates to a system and method for displaying an indoor location when a disaster situation is interrupted and an indoor environment is changed (damaged) And more particularly, to a system and method for displaying an indoor location at the time of disaster relief.

Generally, a smart terminal periodically exchanges paging signals with base stations to operate in a mobile communication system. Accordingly, the position of the smart terminal can be grasped roughly through the location of the base station communicating with the smart terminal in the mobile communication system. In order to provide the paging signal to the smart terminal, the location of the smart terminal can be known only by knowing the approximate position of the smart terminal.

In addition, there is no way to present information and safety evacuation information of the disaster area to each accident area when the power and information communication network are destroyed at the time of disaster. Exit guide lights and emergency broadcasting, There is a problem that it can induce the worst damage by guiding the person to the stairs and the corridor of the accident area by broadcasting only the emergency exit and broadcasting only the fire alarm siren.

Recently, a program for positioning a smart terminal has been installed in an emergency due to a disaster or an accident. That is, various technologies for locating smart terminals in an emergency due to a disaster or an accident are being developed.

In other words, the rescue request point is briefly described at the requesting side, and the accurate position can be determined by tracking the GPS position of the smart terminal while the rescue team is dispatched. However, when an emergency rescue service is requested in an area where reception of GPS signals is difficult, such as an indoor or underground space, it has been difficult to technically detect the location of the smart terminal in the past, so that an emergency structure is not efficiently provided .

For example, when a disaster such as a fire or an earthquake occurs, indoor localization technology is very useful for rescuing people in buildings. However, if a disaster causes a power outage, the communication facilities installed in the building may not work. In this case, it is not possible to use the indoor location indication technology using the infrastructure. Instead, the received signal strength (RSS) between the rescuer's device and the victim's device can be used for indoor localization.

One example of such a technique is disclosed in the following patent documents and non-patent documents.

For example, in the following Patent Document 1, a beacon is installed in a subway, a subway station, a bus, a department store, a building, and an apartment (home), and the beacon ID installed is registered in the disaster prevention system DB together with the detailed position information. And vehicle IDs of vehicles and buses, and transmits beacon IDs related to the disaster occurrence location to the disaster report app (App) according to the judgment of the situation room manager. At this time, And a disaster prevention system through an indoor location providing device that is sent to a created message or the like.

Patent Document 2 discloses a personnel DB in which personnel distribution data obtained by sensing real-time access information of a person from a sensor unit installed in each partition space in real time in real time is stored, characteristics of a building to be inputted to a simulation engine, A simulation DB for storing a evacuation path including the personnel distribution information as a simulation result value by using the distribution data of the number of stars and algorithm parameters and a simulation DB for storing the personnel distribution data of the personnel DB in case of actual disaster or fire and the simulation result value The Euclidean distance is used to calculate the personnel difference for each of the compartment spaces, and the evacuation route included in the simulation DB having the personnel difference value closest to 0 is extracted And a control unit for receiving the extracted evacuation route, Ray output is disclosed for the emergency evacuation system including a to.

In addition, Patent Document 3 discloses a method in which a national institution terminal transmits a location search request signal to a smart terminal of an emergency rescue requester, a smart terminal drives a wireless LAN receiver to scan a nearby access point, Generating positioning information based on the identification information and the signal strength information of the acquired access points; acquiring reporting information by the positioning server; searching the database for identification information of the access points included in the reporting information; Obtaining a geographical location of the access points scanned by the smart terminal, calculating the geographical location of the smart terminal based on the geographical location of the access points and the signal strength information contained in the reporting information, The server sends the calculated geographical location of the smart terminal to the national institution And providing the geographical location information provided by the national institution terminal to the emergency structure.

The following non-patent document 1 discloses a technique for improving the accuracy of position prediction by reducing the area to be positioned by positioning, estimating the relative position of the terminal using the center of gravity of the polygon composed of the APs receiving the signal of the terminal, Discloses a technique capable of reducing the area of a target polygon by selecting the closest AP as a reference AP for predicting the terminal position when the relative position of the terminal is predicted.

Korean Patent Laid-Open Publication No. 2015-0138790 (published Dec. 10, 2015) Korean Registered Patent No. 10-1442658 (Registered on Sep. 15, 2014) Korean Registered Patent No. 10-1513658 (Registered on Apr. 14, 2014)

 Lim, Yu - Jin, Park, Jae - Sung, "Location Positioning Technique using Laterality in Indoor Environment", Korea Information Processing Society, 2010, Vol. 17 No.3 pp.251 ~ 259.

However, in the conventional art as described above, there has been a problem that a preliminary measurement of the propagation characteristics between the types of walls in the target building is made in advance and a database for the indoor position is constructed, thereby increasing the construction cost.

If the existing communication facility is interrupted and the indoor environment is changed (damaged) in the event of a disaster, even if the database is constructed as described above, the existing localization plan may cause a large error. Was not. That is, if a Wi-Fi or an AP does not operate due to a power outage in the event of a disaster, there is a problem that it is impossible to locate the victim using the communication system.

Disclosure of Invention Technical Problem [8] The present invention has been made to solve the above problems, and it is an object of the present invention to provide a subarea division method that can effectively operate in a disaster relief situation, And to provide an indoor location display system and method.

It is another object of the present invention to provide a system and method for indicating an indoor location at the time of a disaster, in which a mobile node (rescuer) can estimate the location of a fixed node (disaster victim) based on a lower region database.

In order to achieve the above object, according to the present invention, an indoor location display system for a disaster relief system accurately displays an indoor location of a disaster victim for disaster relief even when an existing communication facility is interrupted and the indoor environment is changed A plurality of rescuer terminals capable of communicating with the disaster victim terminal as a plurality of mobile nodes, and a plurality of rescuer terminals capable of communicating with the disaster victim terminal, wherein the resident victim terminal is divided into a lower region for the indoor environment, A position display server for determining the position of the disaster victim terminal having the highest accuracy among the positions reported from all rescuer terminals as the final position of the disaster victim terminal; The number of walls between the region and each lattice point, And a database for storing information on the indoor environment, which is obtained by dividing the target area into lower areas and forming a set of reference points (RPs) separately for each lattice point according to the type of the indoor space.

Further, in the indoor location display system according to the present invention, when the disaster victim terminal receives a signal from the rescuer terminal, it reports RSS (Received Signal Strength) to the rescuer terminal, And estimating a location of the disaster victim terminal based on the lower-level domain database.

Further, in the indoor position display system according to the present invention, the rescue terminal includes an accelerometer and a gyroscope, and when the measured value of the accelerometer becomes a peak value, it is considered that a step event occurs, And estimating the moving direction by averaging the gyroscope values measured during the period.

Also, in the indoor location display system according to the present invention, the rescuer terminal uses Bluetooth for communication between the group leader and other group members, and the communication between the rescuer terminal and the disaster victim terminal uses a wifi hot spot .

In order to achieve the above object, according to the present invention, there is provided a method for displaying an indoor location of a disaster victim in a disaster situation even if an existing communication facility is interrupted and the indoor environment is changed (A) setting a disaster victim terminal as a fixed node in a building, and setting a plurality of rescuer terminals capable of communicating with the disaster victim terminal as a plurality of mobile nodes; (b) Receiving a signal from the mobile node, the fixed node reports RSS (Received Signal Strength) to the mobile node; (c) when the mobile node is in a fixed state based on the path loss model and the sub- Estimating the position of the node, (d) Estimating a probability that each fixed node, the yarn model is based be located at each lower region, it characterized in that it comprises the step of determining by the position of the fixed node to the probability.

According to another aspect of the present invention, there is provided a method for displaying an indoor location, the method comprising: dividing a building into a target area by a grid, dividing the target area into a lower area according to the number and type of walls between the area and each grid point, And is formed by forming a set of reference points (RP) separately for each lattice point, and the database is used as a location indication of a fixed node in an online step.

In the indoor location display method according to the present invention, the step (a) may include grouping the plurality of mobile nodes and selecting one of a plurality of mobile nodes as a group leader for displaying an accurate location of the mobile node, And the step (c) includes the step of correcting an estimation error using RSS between the group leader and another group member.

In the indoor location display method according to the present invention, the step (a) includes an off-line step and an on-line step. In the off-line step, a bottom area database is formed using the floor plan of the building, The location is characterized in that the mobile node calculates and estimates the current location based on Pedestrian Dead Reckoning (PDR) techniques.

Further, in the method of displaying an indoor location according to the present invention, in the step of correcting the estimation error, each mobile node determines whether correction is necessary by examining a specific condition of an estimated position, If the mobile node satisfies the condition, the mobile node requests correction to the group leader, the group leader executes a correction process on the position of the mobile node that needs to be corrected, and sends the corrected position to the mobile node .

Further, in the indoor location display method according to the present invention, each mobile node calculates the location of each fixed node according to the estimated location, the RSS from the fixed node, and the lower area database established at the offline stage, The location server reports the estimated location of each fixed node to the location indication server to find the optimal location of the node, and the location indication server reports the location of the fixed node with the highest accuracy among the locations reported from all mobile nodes, Is determined as the final position of the fixed node.

As described above, according to the indoor location display system and method at the time of the disaster relief structure according to the present invention, the lower area where the node is located is estimated using the probability based on the path loss model, In the offline stage, the target building is divided into several sub-areas based on the floor plan of the building, and since the RSS is not measured at the offline stage, the cost of the offline stage can be reduced compared to the conventional fingerprint-based method Effect is obtained.

In addition, according to the system and method for displaying the indoor location at the time of disaster relief according to the present invention, the mobile node corresponding to the rescue member moves around the building to find a fixed node such as the disaster victim, For example, since each location can be tracked through Pedestrian Dead Reckoning (PDR), which uses the sensing values of the accelerometer and magnetometer in mobile devices to track the location of the device, It is possible to obtain an effect of improving the position display accuracy of the display device.

In addition, according to the present invention, it is possible to eliminate the RSS measurement cost required for the fingerprinting technique and to accurately grasp the indoor position even when the indoor environment is changed (damaged) It is also possible to easily realize a disaster structure such as rescuing people with disabilities.

1 is a conceptual diagram of an indoor location display system at the time of a disaster relief structure according to the present invention,
FIG. 2 is a flowchart illustrating a method of displaying an indoor location in a disaster relief according to the present invention.
3 is a diagram for explaining the process of sub region division,
FIG. 4 is a top view of the experimental environment modeled on the third floor of 302, Seoul National University,
5 is a graph showing simulation results of a normal scenario,
6 is a graph showing simulation results of a scenario in which a wall is damaged,
7 is a graph comparing simulation results and experimental results.

These and other objects and novel features of the present invention will become more apparent from the description of the present specification and the accompanying drawings.

First, the basic configuration of the present invention will be described.

The fixed node and the mobile node applied to the present invention are set to two types of nodes set for the building. That is, a fixed node representing a disaster victim in a building does not know its location. Conversely, mobile nodes representing rescuers within a building know their respective locations. It is assumed that all mobile nodes know the initial location at the bottom because the rescuer can touch the floor through the building entrance or stairs.

Also, a plurality of mobile nodes are grouped and one of these mobile nodes is selected as a group leader for precise location indication of the mobile node. It then uses RSS (Received Signal Strength) between the group leader and other group members to correct the estimation error.

Each mobile node estimates the moving distance of the accelerometer and the direction of movement of the magnetometer and gyroscope. Then, each mobile node can estimate the position using the moving distance and the moving direction for a predetermined time. While each mobile node is tracking its location, the node sends a signal to a fixed node (disaster victim) near the mobile node.

When the fixed node receives a signal from the mobile node, it reports the RSS to the mobile node. Then, the mobile node estimates the location of the fixed node based on the path loss model and the subdomain database established at the offline stage. Instead of directly calculating the distance between nodes, the location server estimates the probability that each fixed node will be located in each sub-region based on the path loss model. Finally, the location (sub-region) of the fixed node is determined by this probability.

Therefore, in the present invention, a system having M mobile nodes and N fixed nodes in a single layer region is considered. In the present invention, the path loss model is used to estimate the probability that each fixed node will be located in each sub-region in the reported RSS. We also consider the wall attenuation factor (WAF) by the wall between the two nodes as well as the distance between the two nodes.

COST231 multi-wall model (G. Ding, Z. Tan, J. Zhang, and L. Zhang, "Regional propagation model based fingerprinting localization in indoor environemnts," IEEE PIMRC 2013, pp. 291-295, Sept. 2013.) , The path loss model with WAF (Wall Attenuation Factor) can be expressed by Equation (1) as follows.

…(1)

... (One)

, waf_w는 벽 w로 인한 신호 감쇄이다. 그리고,

는 경로 손실 지수이고, d₀는 대표적으로 1m로 가정되는 기준 거리이다.here,

, waf _w is the signal attenuation due to the wall w. And,

Is the pathloss index, and d ₀ is a reference distance, which is typically assumed to be 1 m.

D _ij is the distance between the mobile node i and the fixed node j, and when the mobile node i transmits the signal with the transmission power T _i (dB), RSS and R _ij at the fixed node j are expressed by the following equation (2) Given.

log(d_ij) - WAF + η_i …(2)R _ij = R ₀ - 10

log (d _ij ) - WAF + η _i ... (2)

log(d₀)가 거리 d₀에서 RSS이고, η_i 는 표준 편차 σ를 갖는 제로-평균 가우스 잡음이다.Here, R ₀ = T _i - PL (d ₀ ) + 10

log (d ₀ ) is RSS at distance d ₀ , and η _i Is a zero-mean Gaussian noise having a standard deviation sigma.

와 d_ij와 만날 때, C = R₀ - 10

log(d_ij) - WAF 이다.η _i ~ N (0, σ ² ), R _ij ~ N (C, σ ² )

And d _ij , C = R ₀ - 10

log (d _ij ) - WAF.

Therefore, if the measured RSS is r _ij using the accuracy function shown in the following equation (3), the accuracy of the given distance d can be calculated by comparing with the actual distance d _ij .

…(3)

... (3)

이고, ε는 이 함수의 정확도를 제어하는 상수이다.here,

And ε is a constant that controls the accuracy of this function.

In the present invention, the location display process is composed of two steps, an offline step and an online step. In the offline phase, the floor plan of the building is used to form a lower area database. In the online stage, the position of the fixed node is estimated according to the following steps. First, each mobile node calculates its current location based on, for example, PDR technology.

In order to increase the accuracy of the location indication of the mobile node, each mobile node determines whether the estimated location needs to be checked by checking a specific condition. If the mobile node satisfies the condition, the mobile node requests the group leader for correction. Next, the group leader executes a correction process on the position of the mobile node that needs to be corrected, and notifies the mobile node of the corrected position.

Next, each mobile node calculates the position of each fixed node according to the estimated position, the RSS from the fixed node, and the lower region database established at the offline stage. Since the positions of the estimated fixed nodes in different mobile nodes may be different from each other, each mobile node reports the estimated position of each fixed node to the location indication server to find the optimal position of the fixed node. Finally, the location server determines the position of the fixed node having the highest accuracy among the positions reported from all the mobile nodes as the final position of the fixed node as the final position of the fixed node.

Next, an indoor location display system and method at the time of a disaster relief according to the present invention will be described with reference to the drawings.

1 is a conceptual diagram of a system for displaying an indoor location in a disaster relief structure according to the present invention.

INDUSTRIAL APPLICABILITY According to the present invention, an indoor location display system for a disaster relief system is provided with an indoor location display system capable of accurately displaying an indoor location of a victim of a disaster for disaster relief even when an existing communication facility is interrupted and the indoor environment is changed 1, a disaster victim terminal 10 as a fixed node, a rescue terminal 20 as a mobile node, a location display server 30 for determining a final location of a fixed node, And a database 40 for storing data.

FIG. 1 shows one disaster victim terminal 10 and a rescuer terminal 20 for convenience of explanation. In the present invention, a plurality of disaster victim terminals 10 and a rescuer terminal 20 are provided. In addition, the disaster victim terminal 10 and the rescuer terminal 20 can be applied to a mobile terminal having a normal computing function such as a smart phone, a tablet PC, and a pellet, which are capable of communicating with each other. In particular, the disaster victim terminal 10 and the rescuer terminal 20 are mobile terminals to which a mobile application (or an app, an application) can be installed and executed.

When the disaster victim terminal 10 receives a signal from the rescuer terminal 20, the rescuer terminal 20 reports the RSS to the rescuer terminal 20. The rescuer terminal 20 analyzes the path loss model and the lower region database And estimates the location of the disaster victim terminal 10 based on the location information.

The rescue terminal 20 includes an accelerometer and a gyroscope. When the measured value of the accelerometer is a peak value, the rescue terminal 20 estimates a moving distance by considering that a step event occurs, and calculates the gyroscope value The moving direction can be estimated.

The rescuer terminal 20 uses any one of bluetooth, zigbee and WiFi, preferably Bluetooth, for communication between the group leader and other group members, Communication between the victim terminal 10 and the disaster victim terminal 10 uses any one of bluetooth, zigbee and wifi, preferably WiFi.

The location display server 30 divides the lower area of the indoor environment and determines the highest accuracy among the positions reported from the rescuer terminals 20 as all the mobile nodes as the final position of the disaster victim terminal 10, The disaster victim terminal 10 is determined as the final position of the disaster victim terminal 10 and is displayed on the rescuer terminal 20.

The information on the indoor environment is divided into a target area by a lattice by a location display server 30, a target area is divided into a lower area according to the number and kind of walls between the area and each lattice point, And is formed by forming a reference point (RP) group separately for the lattice points.

Next, a method for displaying the indoor position at the time of the disaster relief according to the present invention will be described with reference to FIG. 2 to FIG. 7 by using the system as described above.

FIG. 3 is a diagram for explaining a process of dividing a lower region. FIG. 3 (a) shows a lattice point in an area, and FIG. 3 3 (b) shows the lower region of the room 1 at the lattice point a, and Fig. 3 (c) shows the RPs at the lattice point a with respect to the entire target region.

First, in the method of displaying the indoor position at the time of the disaster relief according to the present invention, even if the existing communication facility is interrupted and the indoor environment is changed (damaged) in the disaster situation, the indoor position of the disaster victim can be accurately displayed As a display method, a disaster victim terminal 10 is set as a fixed node in a building, and a plurality of rescuer terminal 20 capable of communicating with the disaster victim terminal 10 is set as a plurality of mobile nodes (S10). Hereinafter, the disaster victim terminal 10 will be described as a fixed node and the rescuer terminal 20 will be described as a mobile node.

In addition, the lower region database 40 is constructed to store information on the indoor environment (S20). In step S20, the inside of the building is divided into a target area by the location display server 30, the target area is divided into sub areas according to the number and types of walls between the area and each lattice point, (RP) is formed separately for the reference points.

Thereafter, when a fixed node receives a signal from the mobile node in the building, the fixed node reports RSS (Received Signal Strength) to the mobile node (S30). According to the report in step S30, The location of the fixed node is estimated based on the loss model and the lower region database established at the offline stage (S40).

Instead of directly calculating the distance between the nodes, the location display server 30 estimates the probability that each fixed node is located in each sub-region based on the path loss model, and determines the location of the fixed node by this probability (S50 ).

In order to construct the lower region database 40 in the step S20, the target region is divided into grids having the same size as shown in Fig. 3 (a). Next, the target area is divided into a plurality of sub areas for each lattice point. Since RSS is mainly influenced by walls of the indoor environment, the target region is divided into sub regions according to the number and types of walls between the region and each grid point. Fig. 3 (b) illustrates the lower region for room 1 at lattice point a. For example, if a signal is transmitted at grid point a, it is attenuated by wall 1 when the signal is received at any point in the lower region A.

Likewise, the signal from the lattice point a is attenuated by the walls 2 and 3 when the signal is received from any point in the lower region B. Therefore, in this case, room 1 is divided into two parts, lower area A and lower area B. After dividing the lower region for the entire object region in a similar manner, the center of each lower region is selected as the reference point (RP). Figure 3 (c) shows the RPs at the lattice point a for the entire target area. In this way, you create separate sets of RPs for each grid point.

에 의해 주어지며, 여기서 n은 격자점 i와 RP j 사이의 벽의 개수이다. 각 격자점 i와 각 RP j에 대해 RP와 WAF_ij의 위치 정보를 데이터베이스에 저장한다. 이 데이터베이스는 온라인 단계에서 고정 노드의 위치 표시로 사용된다.WAF _ij represents the damping coefficient due to the wall between the grid points i and RP j. WAF _ij

, Where n is the number of walls between the grid points i and RP j. For each grid point i and each RP j, the location information of RP and WAF _ij is stored in the database. This database is used to indicate the location of fixed nodes in the online phase.

When the mobile node is located at a specific lattice point in the on-line stage, the position of the fixed node is determined to be one of the RPs corresponding to the lattice point. Each RP represents the possible location of the fixed node, and WAF _ij represents the signal attenuation due to the wall between the two nodes. The algorithm according to the invention then calculates the probability that the stationary node will be located at each possible location.

Next, the location indication of the mobile node will be described.

The present invention can be used in conjunction with various methods for tracking the location of a moving device within a building. Among them, PDR technology (W. Kang and Y. Han, "SmartPDR: Smartphone-based pedestrian dead reckoning for indoor localization," IEEE Sensors J. vol. 15, no. 5, pp. 2906-2916, May 2015.) Can be simply applied to track the mobile node. The PDR estimates the moving distance and the moving direction using the measured values of the sensor. The moving distance is estimated by calculating the number of step events. The step event is an event indicating whether or not the mobile node moves.

In the present invention, when the measured value of the accelerometer becomes the peak value, it is assumed that the step event occurs. It is assumed that all step events have the same step length l. Therefore, the travel distance can be calculated by D = l x N _S. Here, N _S represents the number of step events during a specific period. The moving direction is estimated by averaging gyroscope values measured over a certain period of time. Since the average value of the gyroscope is used for the estimation of the moving direction, the error in the moving direction is not large.

Since the initial position of the mobile node is known, the position of the mobile node can be estimated by the movement distance and the movement direction. However, this simple PDR technique can have a large estimation error due to the error of the accumulated detection value over time. The mobile nodes are grouped, and the position of the node is corrected using RSS between the nodes of the group. In the present invention, a similar correction technique by grouping is used.

, 이동 방향

및 시간 t-1에서의 그 위치

를 사용하여 추정된 위치

를 계산한다.This correction process will be described in more detail. At every location indication time t, the group leader sends a location indication request message to other mobile nodes in the group. When receiving the location indication request message, each mobile node i measures the RSS of this message r _il ,

, Direction of movement

And its position at time t-1

Using the estimated location

.

가 어떤 임계 값

보다 큰 값인지를 검사한다.

이면, 이동 노드의 이동 방향이 변화하고 있다. 그렇지 않은 경우, 이 차이는 감지 값의 오류로 인해 발생할 수 있으며 추정 오류가 발생할 수 있다. 이 경우, 각 이동 노드의 추정된 위치는 정정이 필요하다. 시간 t-1과 시간 t에서 노드의 이동 방향 간의 차이가

보다 작게 하기 위해 이동 노드 집합을 S_u로 한다. S_u에서 각 이동 노드 i는 시간 t-1에서의 최종 추정 위치

, 시간 t에서의 계산된 위치

및 리더에 대한 r_il을 보고하여 정정을 요청한다. 다음에 그룹 리더 ℓ은 S_u에서 이동 노드의 위치를 정정한다.Then, the mobile node i

Lt; RTI ID = 0.0 >

It is checked whether the value is larger.

, The moving direction of the mobile node changes. Otherwise, this difference may be due to an error in the sense value and an estimation error may occur. In this case, the estimated position of each mobile node needs to be corrected. The difference between the direction of movement of the node at time t-1 and time t

The mobile node set S _u is set to be smaller. In S _u , each mobile node i has a final estimated location at time t-1

, The calculated position at time t

And r _il for the leader to request correction. Next, the group leader ℓ corrects the position of the mobile node in S _u .

이다. 다른 하나는 그룹 리더 ℓ에 의해 계산된 위치

이다. 리더는 시간 t-1에서

에 위치한 이동 노드 i가

의 방향으로 이동 노드 i에 의해 보고된

의 거리를 이동시킨다는 가정하에

을 계산한다.

는 방향

를 사용하여 계산되고,

는 방향

를 사용하여 계산된다.The correction process is as follows. For each mobile node i in S _u , two candidate positions are considered in the correction process. One is the location reported by mobile node i at time t

to be. The other is the position calculated by group leader l

to be. At time t-1, the reader

RTI ID = 0.0 > i <

Lt; RTI ID = 0.0 > i < / RTI >

On the assumption that the distance of

.

Direction

, ≪ / RTI >

Direction

≪ / RTI >

로서

m과

중 높은 평균 정확도를 갖는 것을 선택한다. 즉, dis(a,b)가 점 a와 점 b 사이의 거리를 나타낼 때, 하기 식(4)와 같이,On the other hand, the group leader may be at S _u . In this case, the reader's position must be corrected. Group leader l is the last calibration position

as

m and

Among those having a high average accuracy. That is, when dis (a, b) represents the distance between point a and point b,

…(4)

... (4)

는 하기 식 (5)와 같다., Where n is the number of mobile nodes in S _u , excluding the leader. Similarly, the position of the mobile node in S _u is corrected using the final estimated position of the reader,

(5). &Quot; (5) "

…(5)

... (5)

After correcting the position of all the mobile nodes in S _u , the reader notifies the mobile node of the corrected position.

Next, the location indication of the fixed node will be described.

When the location indication of the mobile node is completed, all mobile nodes estimate the location of each fixed node (S40). The mobile node i estimates the position of the fixed node j as follows. First, the mobile node i obtains the RPs and the wall decay rate WAF _ir corresponding to each RPr from the database constructed at the offline stage. It is assumed that each mobile node is located on the closest lattice point from the estimated position of the mobile node since the estimated position of the mobile node i is an arbitrary position in the area.

은 고정 노드 j가 RPr에 위치하는 조건하에서 이동 노드 i와 고정 노드 j 사이의 거리를 나타낸다. 다음에 식 (3)의 정확도 함수로

의 정확도를 계산할 수 있다. 고정 노드 j의 실제 위치가 RPr에 가까울 경우,

의 정확도가 높아진다. 그러므로,

의 정확도는 고정 노드 j가 이동 노드 i의 관점에서 RPr에 위치할 확률일 수 있다.Next, the mobile node i measures RSS, r _ij from the fixed node j.

Represents the distance between the mobile node i and the fixed node j under the condition that the fixed node j is located at the RPr. Next, as an accuracy function of equation (3)

Can be calculated. When the actual position of the fixed node j is close to RPr,

The accuracy of the image is increased. therefore,

May be the probability that the fixed node j will be located at the RPr in terms of the mobile node i.

보다 높게 설정된 RP의 집합으로 식 (6)과 같이

를 정의한다.However, since the mobile node is usually located at a random location other than the grid point, it may not be accurate to select the RP with the highest probability. Thus, the position indication result at time t-1 is additionally used in the proposed technique. We assume that Stij is the probability that the computed probability is at a certain threshold

As a set of higher RPs,

.

…(6)

... (6)

= 0일 때, 이동 노드 i는

를

만큼 낮게 하고,

> 0일 때까지 다시 계산한다. 그 후, 이동 노드 i는 이전 결과를 고려하여 형성된 RPs 세트인 하기 식 (7)

에서 가장 높은 가능성을 가진 RP로서 고정 노드 j의 위치를 추정한다. Here, S is a set of all RPs.

= 0, the mobile node i

To

As a result,

Calculate again until> 0. After that, the mobile node i calculates the following equation (7), which is a set of RPs formed considering the previous result,

The position of the fixed node j is estimated as the RP having the highest probability.

(7)

(7)

의 세트에서 선택할 수 있다. 그러나

= 0일 때, 이동 노드는

에서 가장 높은 확률을 갖는 RP로서 고정 노드의 위치를 추정한다. 이동 노드 i에서 고정 노드 j의 위치 표시 단계는 하기 표 1의 알고리즘으로 주어진다.This is for the following reasons. If the mobile node (i.e., rescuer) is a pedestrian, it is expected that the travel speed is not fast and the travel distance during the location update interval is small. For this reason, the estimated position of the fixed node at time t is not significantly different from the predicted node at time t-1. Therefore, the position of the fixed node j is

Can be selected. But

= 0, the mobile node

The position of the fixed node is estimated as the RP having the highest probability. The location indication step of the fixed node j at the mobile node i is given by the algorithm of Table 1 below.

, 고정 노드 j의 예상 위치

및 각 고정 노드 r_ij의 RSS를 위치 표시 서버로 전송한다.

는 하기 식 (8)에 의해 모든 이동 노드들로부터 고정 노드 j의 추정된 위치들의 세트를 나타낸다.In order to further improve the accuracy of the location indication, the location display server collects the location indication results of all mobile nodes and selects one of the results as the final location of the fixed node. If there is more than one mobile node in the group,

, The expected position of the fixed node j

And the RSS of each fixed node r _ij to the location display server.

Represents a set of estimated positions of fixed node j from all mobile nodes by the following equation (8).

(8)

(8)

로부터 가장 높은 평균 정확도를 갖는 RP를 고정 노드 j의 최종 추정 위치로 결정한다(S50).Then, the location display server uses the following equation (9) as in equation (4)

The RP having the highest average accuracy is determined as the final estimated position of the fixed node j (S50).

(9)

(9)

The evaluation of the simulation according to the present invention is as follows.

= 0.85,

= π/18 rad, ε = 18 및

= 2m이다. 위치 표시 업데이트 간격은 1초로 설정한다. 여기서, 이 지역에 10개의 고정 노드를 무작위로 배포한다. 시뮬레이션에서 각각의 이동 노드는 1 m/s의 속도로 움직이며, 제안된 기법의 최종 목표인 고정 노드 위치 추정의 정확성을 집중적으로 나타내기 위해 그 위치는 정확히 알려져 있는 것으로 가정한다.The performance of the position recognition technique according to the present invention is evaluated through simulation. The total plane is 33.0m × 34.8m. FIG. 4 is a plan view of an experimental environment modeled on the third floor of the university building 302 of Seoul National University, and shows an example of the lower region database at the lattice point a (reference point?). Here, R0 = -23dB and the standard deviation sigma = 6dB of the new drawing effect are used. It was also assumed that all the walls had wall attenuation coefficients of 18 dB each. These parameters are obtained through empirical methods. For example, the RPs and the set of wall attenuation coefficients at grid point a are shown in FIG. Other parameters include an accuracy threshold

= 0.85,

=? / 18 rad,? = 18, and

= 2m. The location update interval is set to 1 second. Here, 10 fixed nodes are randomly distributed in this area. In the simulation, each mobile node moves at a speed of 1 m / s. It is assumed that its position is known accurately in order to focus on the accuracy of the fixed node location estimation, which is the final goal of the proposed technique.

In order to evaluate the performance, a conventional fingerprinting based nearest neighbor algorithm (hereinafter referred to as 'SD-NN') combined with a sub-region decision algorithm was compared with the scheme proposed in the present invention. In the SD-NN algorithm, the number of RPs in each sub-region affects not only the cost of the offline step but also the performance of the algorithm. Therefore, the performance of the SD-NN algorithm was evaluated for two cases: 12 RPs in each sub-region and 2 RPs in each sub-region.

의 영향을 받는다. 따라서

가 알려진 이상적인 경우와

가 알려지지 않은 현실적인 경우인 두 가지 경우에 대해 본 발명에 따른 구성의 성능을 평가하였다.As shown in equation (3), the accuracy of the proposed algorithm is based on the path loss model,

. therefore

Is the ideal known case

The performance of the configuration according to the present invention was evaluated for two cases, in which it is unknown that it is a real case.

는 [2.0, 6.0] 사이에 균일하게 분산된 값으로 설정된다. 이상적인 경우(알려진

)에 대해 식 (3)의 계산에 대해

와 동일한 값을 사용한다. 반대로 현실적인 경우에

의 값을 모르므로, 식 (3)의 계산에 대해

= 4.0이라고 가정한다.In each simulation run, the path loss index

Is set to a value uniformly distributed between [2.0, 6.0]. Ideal case (known

) For the calculation of equation (3)

The same value is used. Conversely, in a realistic case

(3) is not known,

= 4.0.

We compared the average position error and average hit rate of fixed nodes. The hit ratio is defined as the ratio of the number of fixed nodes where the sub area is accurately determined to the number of all fixed nodes under test. Also, one of the goals of the configuration according to the present invention is robustness to environmental changes in disaster relief situations, so the algorithms are compared in two scenarios, a general scenario and a wall damage scenario. In normal scenarios there is no environmental change between off-line and on-line. In a scenario where a wall is damaged, some walls in the area are damaged during the online phase. 200 simulations were performed for each configuration of the two scenarios.

The results of the normal scenario are as follows.

In the normal scenario, the average position error and the average hit rate were compared for the number of mobile nodes used for location indication. The number of mobile nodes may be set to 2 to 5 since the number of rescue personnel may not be large for a single layer in a disaster relief situation. The SD-NN algorithm uses fixed and known location APs for location indication. Thus providing a fixed and known location for each mobile node that can be considered an AP in the SD-NN algorithm.

FIG. 5 is a graph showing simulation results of a normal scenario, and FIGS. 5 (a) and 5 (b) show an average position display error and an average hit ratio at a fixed node in a normal scenario, respectively. The method according to the present invention and the SD-NN algorithm show that the average position error of the fixed nodes decreases as the number of mobile nodes increases. This is because the RSS measure that can be used increases as the mobile node gets more information. Therefore, since the effect of the measurement noise is lowered, the average position display error decreases. If the number of mobile nodes is small, the SD-NN algorithm uses many RPs per sub-area, but the method according to the present invention performs more accurate positioning. Although there are five mobile nodes, the method according to the present invention provides similar performance compared to the SD-NN algorithm with many RPs per sub-region. However, if the existing algorithm uses many RPs for each sub-region, the cost of creating fingerprints for the RPs at the offline stage is high. The method according to the present invention is much more efficient at location indication accuracy per building cost since it is much less expensive at the offline stage. On the other hand, if the SD-NN algorithm uses less RPs to lower training costs in the off-line phase, the method according to the present invention provides more accurate locator results.

The results of the wall damaged scenario are as follows.

In the case of a wall damaged scenario, five mobile nodes were used for each setting. Instead, the average position error and the average hit rate are compared with the damage rate in the corresponding area. The damage rate is defined as the ratio of the number of walls damaged by a disaster to the number of all walls in the area.

FIG. 6 is a graph showing a simulation result of a scenario in which a wall is damaged, and FIGS. 6 (a) and 6 (b) show an average position display error and an average hit rate of fixed nodes in a wall damage scenario, respectively. The method according to the present invention and the SD-NN algorithm show that the average hit ratio decreases as the average position error of the fixed node increases and the area damage ratio increases. This is because the data pre-investigated at the off-line level is less reliable as the damage rate of the area increases. However, if the environment changes at the online level, the change in RSS is much greater than the change in wall attenuation.

In addition, the method according to the present invention confirms the probability that the RSS is within a certain range, so that there is a larger tolerance than the SD-NN algorithm which directly uses the fingerprint for the display position. Thus, the method according to the present invention has less performance degradation than the SD-NN algorithm in a wall damaged scenario.

The performance of the position indicating method according to the present invention was evaluated through experiments. In order to actually verify the method according to the present invention, experimental results and the above-described simulation results were compared. The experiment was conducted at the 3rd floor of the 302, Seoul National University. Four fixed nodes were deployed in the laboratory. The mobile node roams through the floor corridor. If the mobile node is too far away from the fixed node, the experiment is performed at one part of the floor because the signal from the mobile node can not be detected at the fixed node. The test environment is 33.0 m x 34.8 m in size, and Fig. 4 shows the layout. In the present invention, the same parameters as those used in the simulation for the experiment were used. Further, the present invention uses the same database constructed by using only the wall attenuation coefficient of 18 dB for each wall in building layout and simulation. The method according to the present invention was implemented as an Android application, and experiments were conducted using Samsung Galaxy S4, a commercial smartphone, and Samsung Galaxy Note 10.1. The mobile node also uses Bluetooth for communication between the group leader and other group members. Since the distance between the group leader and other group members is short, the range of Bluetooth communication is short, but using Bluetooth is sufficient. In addition, the mobile node uses a Wi-Fi hotspot for communication between the mobile node and the fixed node.

We compared the average position error and the average hit rate of fixed nodes with the number of mobile nodes used for location indication. As in the simulation setup, the number of mobile nodes is changed from 2 to 5. Experiments were run for 2 minutes for each case.

The results of the experiment are shown in FIG.

FIG. 7 is a graph comparing simulation results with experimental results, and FIGS. 7 (a) and 7 (b) show average position error and average hit rate of fixed nodes. The experimental results show that the accuracy is about 10% lower than the simulation results. This is because we assume that we know the exact location of each mobile node in the simulation to see the location indication performance of the fixed nodes. Conversely, since the position of the mobile node in the experiment can have some estimation error, the accuracy of the position indication of the fixed node also decreases.

In addition, performance may be degraded due to the inaccuracy of the path loss model. It is sufficient to actually demonstrate the feasibility of the method according to the invention even if the performance is slightly degraded in the experiment. In addition, since the method according to the present invention requires only a simple path loss model of a building plan and an off-line step, it can be applied to various buildings more easily than a fingerprint-based approach that requires high cost to actually measure RSS for building a fingerprint database have. Therefore, the present invention provides a very practical location marking method which is inexpensive especially in emergency situations such as disaster relief. Positioning accuracy will improve if the rescuer uses a specialized mobile device for disaster relief.

As described above, the present invention proposes a new indoor location display technique based on a floor plan. By performing sub-region partitioning according to the number of walls between nodes, the algorithm according to the present invention has more tolerance for environmental changes in the online phase. Further, the method according to the present invention does not measure any RSS in advance in order to reduce the cost of building a database. The simulation results show that the accuracy of position indication is improved compared to the conventional fingerprinting method under normal circumstances. In addition, the method according to the present invention provides a significant performance improvement in a disaster situation. Experimental results in addition to simulations have proved that other methods according to the invention are actually possible.

Although the present invention has been described in detail with reference to the above embodiments, it is needless to say that the present invention is not limited to the above-described embodiments, and various modifications may be made without departing from the spirit of the present invention.

The cost of the off-line step can be reduced compared to the conventional fingerprint-based method by using the indoor location display system and method according to the present invention.

10: Disaster victim terminal
20: rescue terminal
30: Location display server
40: Database

Claims (10)

An indoor location display system for accurately indicating an indoor location of a disaster victim for disaster relief even when an existing communication facility is interrupted and an indoor environment is changed in a disaster situation,
A disaster victim terminal as a fixed node,
A plurality of rescuer terminals capable of communicating with the victim terminal as a plurality of mobile nodes,
Wherein the location of the disaster victim terminal having the highest accuracy among the positions reported from all rescuer terminals as the final position of the disaster victim terminal is determined as the final position of the disaster victim terminal Display server,
The inside of the building is divided into the target area by the location display server, the target area is divided into the lower area according to the number and type of the wall between the area and each lattice point, and the reference point RP And a database for storing information on the indoor environment formed by forming a set of reference points. The method of claim 1,
The rescue terminal reports RSS (Received Signal Strength) to the rescuer terminal when the disaster victim terminal receives a signal from the rescuer terminal. The rescuer terminal reports the path loss model to the rescuer terminal, And estimating the position of the indoor location. The method of claim 1,
The rescue terminal includes an accelerometer and a gyroscope. When the measured value of the accelerometer is a peak value, the rescue terminal estimates a moving distance by considering that a step event occurs and averages the gyroscope values measured during a predetermined period And estimating a moving direction of the indoor space. The method of claim 1,
Wherein the rescuer terminal uses Bluetooth for communication between the group leader and other group members,
Wherein the communication between the rescuer terminal and the disaster victim terminal uses a wifi hot spot. There is provided an indoor location indication method for accurately indicating an indoor location of a disaster victim for disaster relief even when an existing communication facility is interrupted and the indoor environment is changed in a disaster situation,
(a) setting a disaster victim terminal as a fixed node in a building, and setting a plurality of rescuer terminals capable of communicating with the disaster victim terminal as a plurality of mobile nodes,
(b) the fixed node reports a received signal strength (RSS) to the mobile node when the fixed node receives a signal from the mobile node,
(c) estimating the location of the fixed node based on the path loss model and the lower region database established in the offline step,
(d) the location indication server estimates the probability that each fixed node will be located in each sub-area based on the path loss model instead of directly calculating the distance between the nodes, and determining the position of the fixed node by this probability and,
The lower region database divides the interior of the building into a target region by a lattice, divides the target region into sub regions according to the number and type of walls between the region and each lattice point, Reference Points) set,
Wherein the database is used as a location indicator of a fixed node in an online step. delete The method of claim 5,
Wherein the step (a) includes grouping the plurality of mobile nodes and selecting one of a plurality of mobile nodes as a group leader for displaying an accurate position of the mobile node,
Wherein the step (c) includes the step of correcting an estimation error using RSS between the group leader and another group member. The method of claim 5,
The step (a) comprises an off-line step and an on-line step,
In the off-line step, a bottom area database is formed using the floor plan of the building,
Wherein the position of the fixed node in the on-line step is calculated by calculating the current position based on a pedestrian dead reckoning (PDR) technique of the mobile node. 8. The method of claim 7,
Wherein the step of correcting the estimation error comprises the steps of: determining, in order to increase the accuracy of the location indication of the mobile node, each mobile node examining a specific condition to determine if correction is necessary; Wherein the group leader requests the group leader for correction, the group leader performs a correction process on the position of the mobile node that needs to be corrected, and notifies the mobile node of the corrected position. The method of claim 9,
Each mobile node calculates the position of each fixed node according to the estimated position, the RSS from the fixed node, and the lower region database constructed in the offline step,
Each mobile node reports the estimated location of each fixed node to the location indication server to find the optimal location of the fixed node,
Wherein the location display server determines the position of the fixed node having the highest accuracy among the positions reported from all the mobile nodes as the final position of the fixed node as the final position of the fixed node.

Images