Encyclopedia of GIS

2017 Edition
| Editors: Shashi Shekhar, Hui Xiong, Xun Zhou

3D Network Analysis for User Centric Evacuation Systems

  • Umit Atila
  • Ismail Rakip Karas
  • Yasin Ortakci
Reference work entry
DOI: https://doi.org/10.1007/978-3-319-17885-1_1549

Introduction

Research on evacuation of high-rise buildings in case of disasters such as fire, terrorist attacks, indoor air pollution incidents, etc., has become popular in the last decade. In case of such disasters, people inside the buildings should be evacuated out of the area as soon as possible. However, organizing a quick and safe evacuation is a difficult procedure due to the complexity of high-rise buildings and the huge number of people occupied inside such buildings. Besides, problems such as smoke inhalation, confluence, panic, and inaccessibility of some exits may arise during the evacuation procedure. Therefore, an efficient user-centric evacuation system should be developed for quick and safe evacuation from high and complex buildings.

Routing someone to an appropriate exit in safety can only be possible with a system that can manage the 3D topological transportation network of a building. Realizing an evacuation of a building in such systems also called navigation systems by guiding people in real time requires complex analysis on 3D spatial data.

Interest on 3D navigation systems has increased especially after the 9/11 attacks, and many researchers concentrated on how a safe and quick evacuation could be realized in case of such disasters (Lee 2007). Most of the navigation systems operate on 2D data to find and simulate the shortest path (which is lacking building environment) (Musliman and Rahman 2008). Therefore, there is a need for different approaches which use the 3D objects and eliminate the network analysis limitations on multilevel structures (Cutter et al. 2003; Pu and Zlatanova 2005; Kwan and Lee 2005; Zlatanova et al. 2004).

In a study conducted by Kwan and Lee (2005), relative accessibility of the emergency response between a disaster site and an emergency station in a building was measured. Their results showed that extending 2D-GIS to 3D-GIS representations of the interiors of high-rise buildings can improve the overall speed of the rescue process.

Most of the GIS researchers use graph networks for indoor routing and evacuation analysis (Karas et al. 2006; Jun et al. 2009). While most of the 3D visualization problems have been solved by CityGML, initial requirements, concepts, frameworks, and applications from a wide point of view have been represented by some other research such as (Pu and Zlatanova 2005; Musliman et al. 2006). However, there is still a lack of implementation of 3D network analysis and navigation specifically for evacuation purposes.

The objective of this study is to investigate and implement 3D visualization and navigation techniques and solutions for indoor spaces within 3D GIS. We explain how to perform 3D network analysis using Oracle Spatial and Graph within a Java-based 3D-GIS implementation. As an initial implementation, a GUI provides a 3D visualization of a building. A network model based on CityGML data stores spatial data in Oracle database and then performs network analysis under different constraints, such as avoiding nodes or links in the network model. All experiments highlighted in this chapter are performed on the 3D model of the Corporation Complex in Putrajaya, Malaysia.

Sections “Evacuation Process” and “Evacuation Systems” summarize evacuation process and evacuation systems, respectively. Section “Visualization of 3D Network Models for Evacuation” gives examples of visualization of 3D building and network models from the CityGML format. Section “Representing Network Model in Geo-DBMS” gives some information on storing spatial data and explains how to create Network Models in Oracle Spatial and Graph. Section “Network Analysis Tool” introduces a 3D network analysis tool and gives visualized results of 3D network analysis performed by our proposed 3D-GIS implementation. Section “Simulation of User Centric Evacuation” elaborates the routing engine integrated in the simulation module and presents a visualization sample.

Evacuation Process

One of the most dangerous disasters threatening the high-rise and complex buildings is fire in which most of the people may lose their lives due to smothering rather than burning. In case of fire disasters, extraordinary indoor air pollution (EIAP) incidents happen suddenly and cause fatal consequences such as airlessness, excessive temperature, explosions, and smoke and toxic gas leakages. Table 1 indicates the number of people died due to various reasons after a residential fire incident (Holborn et al. 2003). As it can be deduced from Table 1, the major death cause was breathing in smoke, followed by combination of burning and smothering.
3D Network Analysis for User Centric Evacuation Systems, Table 1

Number of people died due to various reasons after a residential fire incident (Holborn et al. 2003)

Reason of death

Number of people who lost their lives

Percentage

 

Inhalation

101

36

 

Smothering

8

3

 

Burned bronchus

8

3

 

Burning

53

19

 

Combination of burning and smothering

69

25

 

Others

20

7

 

Injuries due to heart attack stroke and falling

20

7

 

There are three main stages in extraordinary indoor air pollution incidents. In the first stage, occupants are not affected by smoke, gas, or temperature; therefore, this stage is the most appropriate stage for evacuation. In the second stage, the occupants are heavily exposed to smoke, toxic gas, and excessive temperature.

In previous studies, the behaviors of the occupants are analyzed in the two main stages discussed in the previous paragraph during a disaster (Purser and Bensilum 2001). The first stage is the premovement time or response time, and the second stage is the movement time or action time. Premovement time is defined as the period between the time alarm systems activates and the time people react to escape from the building. Table 2 compares the main factors that triggered occupant evacuation in buildings in England and the USA (Wood 1972; Bryan 1977). This indicates that the effect of alarm systems in initiating people to react is unexpectedly low.
3D Network Analysis for User Centric Evacuation Systems, Table 2

The main factors that triggered occupant evacuation in buildings (Wood 1972; Bryan 1977)

The main factors that triggered occupant evacuation

England %

USA %

 

Smoke

34.0

35.1

 

Shouting and voices

33.0

34.7

 

Flames

15.0

8.1

 

Noise

9.0

11.2

 

Alarm

7.0

7.4

 

Others

2.0

2.8

 
A study conducted by Purser and Bensilum (2001) in a shopping mall indicated that when occupants are informed by announcement system, most of the evacuation time procedure was realization of a need to evacuate, rather than movement time. Figure 1 indicates that the percentages of realization, response, and reaction times were 65%, 16%, and 19%, respectively. Therefore, premovement time is 81% of the total evacuation time.
3D Network Analysis for User Centric Evacuation Systems, Fig. 1

Occupant behavior time (Purser and Bensilum 2001)

The studies also indicated when an alarm system sounds, occupants spend the most critical time period to understand the reason of the alarm rather than evacuating the building. Also, studies indicated that the occupants give different responses based on the type and method of alarm system or content and time of the announce (Bryan 2002). Uncertainty and insufficient information during the event may delay the evacuation procedure.

The second stage of evacuation is movement time or action time. Movement time is the period between the time people react to escape from the building and the time they reach out of the building or some safe place in the building (Purser and Bensilum 2001). Movement time varies based on two main factors: exit preferences and smoke problems.

Current evacuation systems assume that occupants use the closest exit in a time of emergency evacuation. Table 3 indicates the results from a study where the preferences of the occupants were investigated in a building where there was one emergency exit door and one entrance door located in opposite locations to each other. As seen in the Table 3, most of the guests used the entrance which they were more familiar with (Mawson 1980), while almost all of the occupants use the emergency exit door. People use the closest exit only if they know the building well (Gwynne et al. 1999). When the guidance of the evacuation systems is insufficient, people consider various factors in choosing the evacuation path.
3D Network Analysis for User Centric Evacuation Systems, Table 3

Exit preference rates of people (Sime 1985)

Exit preference

Guest

Occupant

Total

 

Entrance door

37

1

38

 

Emergency exit door

24

13

37

 
Previous studies reported that when occupants encounter a smoke problem, they keep moving through the smoke if the sight distance is more than 20 m; however, they hesitate and do not take the risk when sight distance is less than 20 m (Bryan 1995). Thus, smoke is a serious problem which affects the movement time in evacuation process. People slow down in smoke, and they cannot determine an optimum evacuation path or cannot follow a straight route due to diminished sight distance (Jin 1976). However, it can sometimes be necessary to pass through a smoke area for survival. Based on a previous study, Table 4 indicates the percentages of occupants returning back due to low sight distance in smoked zones (Bryan 1995).
3D Network Analysis for User Centric Evacuation Systems, Table 4

Percentages of occupants returning back due to low sight distance

Visibility (meter)

England (%)

USA (%)

 

0–2

29.0

31.8

 

3–6

37.0

22.3

 

7–12

25.0

22.3

 

13–30

6.0

17.6

 

31–36

0.5

1.3

 

37–45

1.0

0

 

46–60

0.5

4.7

 

> 60

1.0

0

 

Evacuation Systems

Traditional evacuation systems can be divided into three main groups: sensors to detect heat, smoke, or radiation; alarm system to alert people at the early stages of a disaster; and evacuation lighting to allow occupants to continue to navigate (Fig. 2). Traditional evacuation systems are not sufficient for safe and quick evacuation of today’s high-rise and complex buildings (Pu and Zlatanova 2005). These evacuation systems are not flexible due to their static predefined scenarios. This may guide people to block exits or places where there are gas leakages. Also, traditional evacuation systems become useless when sight distance is very low due to smoke and electricity cuts. They also provide insufficient evacuation information, especially for people who are not familiar with the building.
3D Network Analysis for User Centric Evacuation Systems, Fig. 2

The components of current evacuation systems (Pu and Zlatanova 2005)

Emergency incidents are not static, but they are dynamic and variable events. However, traditional evacuation instructions are generally insufficient in dynamic evacuation process. The stage in which people spend most of the time in case of emergency is the stage during which they do not react or take action but rather the stage of realizing the event before starting to move. Uncertainty at the time of the emergency and the lack of clear information about the incident are factors in delaying the evacuation of the building. Therefore, a system that can provide understandable and clear information to all users in real time and resolve their concerns will definitely shorten the evacuation process. Such an ideal system is a smart evacuation system that can avoid congestion by allocating traffic across the available routes or guide people away from areas of risk (smoky and dangerous) in case of necessity. Therefore, an ideal evacuation system allows people to progress rapidly without hesitation and without the need for determining the route themselves. To realize an ideal indoor evacuation system, a number of main functionalities should be addressed. These functionalities are listed as follows:
  • A spatial database for the management of the building and network models.

  • 3D-GIS-based routing engine centralized in an appropriate host.

  • Mobile-based navigation software for passing user-related data to the host and for presenting routing instructions to the user clearly.

  • An accurate 3D indoor positioning system.

  • Well-organized wireless communication and sensor network architectures inside the building.

In the rest of this topic, we will concentrate on formalization of 3D building and network models within 3D GIS needed to construct a dynamic evacuation system and present a shortest path analysis and various evacuation simulation examples.

Visualization of 3D Network Models for Evacuation

A Java-based 3D-GIS implementation has been developed that is able to visualize 3D building model and perform network analysis on the network model of building. The implementation uses citygml4j Java class library and API for facilitating work with the CityGML and JOGL Java bindings for the OpenGL graphic library to carry out visualization of 3D spatial objects.

CityGML is introduced as one of the international standards for representing and exchanging spatial data, making it easier to visualize, store, and manage 3D city models data efficiently. CityGML is able to represent 3D city models in five well-defined Level of Details (LOD), namely, LOD0 to LOD4. The accuracy and structural complexity of the 3D objects increases with the LOD level where LOD0 is the simplest LOD with a two-and-a-half-dimensional Digital Terrain Model, while LOD4 is the most complex LOD including architectural details with interior structures. LOD1 is the well-known blocks model comprising prismatic buildings with flat roofs. Differentiated from LOD1, LOD2 has roof structures. LOD3 denotes architectural models with detailed wall and roof structures and balconies (Gröger et al. 2008).

The implemented system reads CityGML datasets from LOD0 to LOD2. 3D building models are represented in LOD2 described by polygons using the Building Module of CityGML (Fig. 3). Network models are represented as linear networks in LOD0 using CityGML’s Transportation Module (Fig. 4).
3D Network Analysis for User Centric Evacuation Systems, Fig. 3

Building model (textured viewing mode)

3D Network Analysis for User Centric Evacuation Systems, Fig. 4

Network model

Representing Network Model in Geo-DBMS

While CityGML is used to store and visualize 3D spatial objects, graph model managed in a geo-database management system (DBMS) is used to perform 3D network analysis. Oracle Spatial and Graph is one of the most powerful geo-DBMS, which offers a combination of geometry models and graph models (Murray 2009).

Oracle Spatial and Graph maintains a combination of geometry and graph models within the Network Data Model. A spatial network consists of nodes and links which are SDO_GEOMETRY objects representing points and lines, respectively (Kothuri et al. 2010).

Network support in the Oracle database is composed of the following elements:
  • A data model to store networks inside the database as a set of network tables: This is the persistent copy of a network.

  • SQL functions to define and maintain networks (i.e., the SDO_NET package).

  • Network analysis functions in Java programming language: The Java API works on a copy of the network loaded from the database. This is the volatile copy of the network.

  • Network analysis functions in PL/SQL (the SDO_NET_MEM package).

Figure 5 illustrates the relationship between the elements of the Oracle Network Model (Kothuri et al. 2010).
3D Network Analysis for User Centric Evacuation Systems, Fig. 5

Oracle network data model

To define a network in Oracle Spatial and Graph, at least two tables should be created. These are node and link tables. These tables should be provided with the proper structure and content to model the network. A node table (see Table 5) describes all nodes in the network. Each node has a unique numeric identifier (the NODE_ID column). A link table (see Table 6) describes all links in the network. Each link has a unique numeric identifier (the LINK_ID column) and contains the identifiers of the two nodes it connects (Kothuri et al. 2010).
3D Network Analysis for User Centric Evacuation Systems, Table 5

Example entry in node table in network model

NODE_ID

230

 

NODE_NAME

NODE-230

 

GEOMETRY

MDSYS.SDO_GEOMETRY(3001,

NULL,MDSYS. SDO_POINT_TYPE

(42.2019449799705,100.382921548

946, −3.7),NULL,NULL)

 

ACTIVE

Y

 
3D Network Analysis for User Centric Evacuation Systems, Table 6

Example entry in link table in network model

LINK_ID

15

 

START_NODE_ID

452

 

END_NODE_ID

455

 

LINK_NAME

Link-452-455-Corridor

 

GEOMETRY

MDSYS.SDO_GEOMETRY(3002,

NULL,NULL,MDSYS. SDO_ EL-

EM_INFO_ARRAY(1,2,1),MDSYS.

SDO_ORDINATE_ARRAY (115.30

6027729301,85.9775129777152,1.8, 115.306027729301,82.9483382781

573,1.8))

 

LINK_LENGTH

3,029174699557899

 

ACTIVE

Y

 

LINK_TYPE

Corridor

 

In this study, as we define a spatial network containing both connectivity and geometric information, we use SDO_GEOMETRY for representing points and lines.

For completing the network creation process, Oracle Spatial and Graph needs a metadata table called USER_SDO_NETWORK_METADATA (see Table 7) to ensure the table structures are consistent with the metadata. The metadata table USER_SDO_NETWORK_METADATA describes the elements that compose a network such as names of the tables and names of the optional columns.
3D Network Analysis for User Centric Evacuation Systems, Table 7

Example entry in USER_SDO_NETWORK_METADATA view

NETWORK

CORPORATION_

PUTRAJAYA

 

NETWORK_CATE-

GORY

SPATIAL

 

GEOMETRY_TYPE

SDO_GEOMETRY

 

NO_OF_HIERARCHY-LEVELS

1

 

NO_OF_PARTITIONS

1

 

LINK_DIRECTION

UNDIRECTED

 

NODE_TABLE_NAME

CORP_NETWORK_NODE

 

NODE_GEOM_

COLUMN

GEOMETRY

 

NODE_COST_CO-

LUMN

NULL

 

LINK_TABLE_NAME

CORP_NETWORK_LINK

 

LINK_GEOM_

COLUMN

GEOMETRY

 

LINK_COST_COLUMN

LINK_LENGTH

 

PATH_TABLE_NAME

CORP_NETWORK_

PATH

 

PATH_LINK_TABLE_

NAME

CORP_NETWORK_PATH_

LINK

 

PATH_GEOM_COLUMN

GEOMETRY

 

There are two choices to create a network. One can either prefer to create network automatically using CREATE_SDO_NETWORK function of SDO_NET package or prefer to create network manually. CREATE_SDO_NETWORK function creates all the structures of a network, but it is not flexible as it gives very little control over the actual structuring of the tables. Sample code given below illustrates creation of CORPORATION_PUTRAJAYA network with explicit table and column names.

$$\displaystyle{ \begin{array}{*{20}{l}} \mbox{ SQL $>$ }&\mbox{ BEGIN} \\ &\mbox{ SDO}{\_}\mbox{ NET.CREATE}{\_}\mbox{ SDO}{\_}\mbox{ NETWORK} (\ \\ &\mbox{ NETWORK $ =>$ `CORPORAHON}\_ \mbox{PUTRAJAYA}' , \\ &\mbox{ NO}{\_}\mbox{ OF}{\_}\mbox{ HIERARCHY-LEVELS $ =>$ 1,} \\ &\mbox{ IS}\_\mbox{DIRECTED $ =>$ FALSE,} \\ &\mbox{ NODE}{\_}\mbox{ TABLE}{\_}\mbox{ NAME $ =>$ `CORP}{\_}\mbox{ NETWORK}\_\mbox{NODE}', \\ &\mbox{ NODE}{\_}\mbox{ GEOM}{\_}\mbox{ COLUMN $ =>$ `GEOMETRY}', \\ &\mbox{ NODE}\_\mbox{COST}{\_}\mbox{ COLUMN $ =>$ NULL,} \\ &\mbox{ LINK}{\_}\mbox{ TABLE}{\_}\mbox{ NAME $ =>$ `CORP}{\_}\mbox{ NETWORK}\_\mbox{LINK} ', \\ &\mbox{ LINK}{\_}\mbox{ GEOM}{\_}\mbox{ COLUMN $ =>$ `GEOMETRY}', \\ &\mbox{ LINK}{\_}\mbox{ COST}{\_}\mbox{ COLUMN $ =>$ `LINK}{\_}\mbox{ LENGTH}' \\ &\mbox{ );} \\ &\mbox{ END;}\\ \end{array} }$$

The alternative way is to create the network tables manually. When defining network manually, one has to create all needed tables and insert proper data into tables using SQL statements. Manual creation gives total flexibility over the table structures, but one must ensure that the table structures are consistent with the metadata.

Network Analysis Tool

In this section, we will present our implementation that performs network analysis with its network analysis tool based on a Java API provided by the Network Data Model of Oracle Spatial and Graph. The Java API which is put in a package called oracle.spatial.network is very rich and provides a range of analysis functions. Our network analysis tool allows conducting most common 3D network analysis supported by the Oracle Spatial and Graph. With this network analysis tool, it is also possible to perform common 3D network analysis with full functionality including constraints and to see the results on a 3D graphical screen. In this section, a shortest path example will be presented. Figure 6 shows a UML diagram that summarizes network analysis process.
3D Network Analysis for User Centric Evacuation Systems, Fig. 6

UML diagram summarizing network analysis process

The analysis functions are provided by the methods of the NetworkManager class. These methods operates on volatile copy of the network. Therefore, the first step in the network analysis tool is to load network from database. The following loads the complete network named CORPORATION_PUTRAJAYA from database in read only mode.

$$\displaystyle{ \begin{array}{l} \mbox{ Network corporation}{\_}\mbox{ Putrajaya = NetworkManager.readNetwork} \\ ~~~~~~~~~\mbox{(dbConnection, \textquotedblleft CORPORATION}\_\mbox{PUTRAJAYA\textquotedblright );}\\ \end{array} }$$

If we want to define a set of constraint for any of analysis methods to limit the search space, we simply define a SystemConstraint object. The SystemConstraint class allows to define constraints such as MaxCost, MaxDepth, MaxDistance, MaxMBR, MustAvoidNodes, and MustAvoidLinks. Once we definethe SystemConstraint object, we can pass it as last parameter to any of the analysis methods of the NetworkManager class. The following sets a constraint to avoid use of link identified by 6012, 6013, and 6014 in the network.

$$\displaystyle{ \begin{array}{l} \mbox{ Vector avoidLinks = new Vector();} \\ \mbox{ avoidLinks.add(\textquotedblleft 6012\textquotedblright );} \\ \mbox{ avoidLinks.add(\textquotedblleft 6013\textquotedblright );} \\ \mbox{ avoidLinks.add(\textquotedblleft 6014\textquotedblright );} \\ \mbox{ SystemConstraint myConstraint = new SystemConstraint} \\ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ (corporation_Putraj aya) ; \\ \mbox{ myConstraint.SetMustAvoidLinks(avoidLinks) ;}\\ \end{array} }$$

A fundamental operation on a network is to find the shortest path between two nodes. The shortestPath() method returns the best path between two nodes in a network. This method takes network object on which we perform analysis and the start and end nodes. The best path between two nodes is the one with smallest cost. Cost of a node or a link is defined in tables with numeric values. The cost can represent anything such as length of a link or time to travel along that link. If there is no cost column in tables, then all links are considered to have a cost of 1 and nodes have a cost of 0. As stated in the previous section, in this study we use link lengths as costs.

The shortestPath() method returns a Path object. We have a number of methods to extract various pieces of information from a path object such as the cost of path, number of the links, and array of Link objects to extract further information. The following finds the shortest path between nodes 3059 and 3368 on the network applying the constraint defined and then prints various information on found path.

$$\displaystyle{ \begin{array}{l} \mbox{ Path foundPath = NetworkManager.shortestPath(corporation}{\_}\mbox{ Putrajaya,} \\ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~\mbox{3059, 3368, myConstraint);} \\ \mbox{ System.out.println(\textquotedblleft Path cost is \textquotedblright + foundPath.getCost());} \\ \mbox{ System.out.println(\textquotedblleft Number of links \textquotedblright + foundPath.getNoOfLinks());} \\ \mbox{ Link }[\,]\mbox{ linkArray = foundPath.getlinkArray();} \\ \mbox{ Node}[\,]\mbox{ nodeArray = foundPath.getNodeArray();}\\ \\ \mbox{ for (int i = 0; i $ <$ linkArray.length; i++)} \\ ~~~~~~\mbox{System.out.println (\textquotedblleft Link \textquotedblright + linkArray[i].getID() + \textquotedblleft $\backslash$t\textquotedblright} \\ ~~~~~~\mbox{+ linkArray[i].getName() +\textquotedblleft $\backslash$t\textquotedblright + linkArray[i].getCost());}\\ \end{array} }$$
Figure 7 shows the shortest path analysis result on a graphical screen without any constraint. The found path follows nodes 3059-3067-3066-3366-3367-3359-3365-3368. Figure 8 shows how the shortest path is updated after links associated with one of elevators are avoided, shown by red lines which means that elevator is not in use any more. Updated path to destination follows nodes 3059-3065-3070-3069-3369-3370-3365-3368.
3D Network Analysis for User Centric Evacuation Systems, Fig. 7

Shortest path between two nodes without any constraint

3D Network Analysis for User Centric Evacuation Systems, Fig. 8

Recalculated shortest path considering avoided elevators in a part of building

Simulation of User Centric Evacuation

Our implementation has an instruction engine which is integrated into the simulation module to produce voice commands and visual instructions for assisting users dynamically on the way to their destination. In the simulation stage, a floating cursor moves over the path in order to simulate a walking person with respect to the given orders. In this procedure, first, the turns and descending and ascending ways are calculated between nodes and floors, and instructions are defined with respect to the calculations. Then, according to the path and calculations, moving person is simulated on the screen by a floating cursor over the path line. Guiding instructions are spoken by computer by using the text speech algorithms and written on the screen (Figs. 9, 10 and 11).
3D Network Analysis for User Centric Evacuation Systems, Fig. 9

Routing simulation process of the instruction engine-Scene-1 (The red point is the user)

3D Network Analysis for User Centric Evacuation Systems, Fig. 10

Routing simulation process of the instruction engine-Scene-2

3D Network Analysis for User Centric Evacuation Systems, Fig. 11

Routing simulation process of the instruction engine-Scene-3

Conclusions

The modern buildings are designed higher and more complex than ever before, which makes them vulnerable to many potential disasters such as terrorist bombings, fire, and toxic gas leakage. Considering the complexity of modern buildings and the great numbers of people inside, it is rather difficult to organize such a quick emergency evacuation.

Many evacuation systems have been developed to minimize losses in such disasters. 3D geo-information has been widely used in all the disaster management phases such as mitigation, preparedness, and recovery phases. However, it hasn’t really been applied to the response phase under extraordinary circumstances.

In this study, some samples of performing 3D network analysis with visualized results supporting both graph-based and geometric constraints applied were presented. It has been showed how Oracle Spatial and Graph can be a powerful geo-DBMS for realizing 3D network analysis and developing evacuation systems that provide dynamic, specific, and accurate evacuation guidance based on indoor geo-information.

We also presented a simulation module that produces voice commands and visual instructions for assisting people dynamically on the way to the destination. The instruction engine presented in this study for simulating evacuation is intended to be the infrastructure of a voice-enabled mobile navigation system for indoor spaces in our work currently in progress. In our future study, we intend to design an intelligent user-centric evacuation model based on neural networks for high-rise building fires in which we will consider the physical conditions of the environment and the properties of the person that requests to be evacuated and produce the personalized instructions in real time.

Notes

Acknowledgements

This study was supported by TUBITAK-The Scientific and Technological Research Council of Turkey research grant [grant number: 112Y050]. We are indebted for its financial support.

References

  1. Bryan J (1977) Smoke as a determinant of human behaviour in fire situations: project people. NBS GCR 77–94, National Bureau of Standards, Washington, DCGoogle Scholar
  2. Bryan J (1995) Behavioural response to fire and smoke. In: SFPE, handbook of fire protection engineering, vol 3, 2nd edn. National Fire Protection Association, Quincy, pp 241–262Google Scholar
  3. Bryan J (2002) Human behaviour in fire. Fire Prot Eng 16:4–16Google Scholar
  4. Cutter S, Richardson DB, Wilbanks TJ (eds) (2003) The geographical dimensions of terrorism. Routledge, New York/London, pp 75–117Google Scholar
  5. Gröger G, Kolbe TH, Czerwinski A, Nagel C (2008) OpenGIS city geography markup language (CityGML) encoding standard: OGC 08-007r1. Open Geospatial Consortium Inc., 7–10, 22–25, 56–62, 77–79Google Scholar
  6. Gwynne S, Galea ER, Lawrence PJ, Filippidis L (1999) A review of the methodologies used in the computer simulation of evacuation from the built environment. Build Environ 34:741–749CrossRefGoogle Scholar
  7. Holborn PG, Nolan PF, Golt J (2003) An analysis of fatal unintentional dwelling fires investigated by London Fire Brigade between 1996 and 2000. Fire Saf J 38(1):1–42CrossRefGoogle Scholar
  8. Jin T (1976) Visibility through fire smoke – Part 5, allowable smoke density for escape from fire. Report No. 42, Fire Research Institute of JapanGoogle Scholar
  9. Jun C, Kim H, Kim G (2009) Developing an indoor evacuation simulator using a hybrid 3D model. In: Lee J, Zlatanova S (eds) 3D Geo-information science. Springer, Berlin, pp 173–178Google Scholar
  10. Karas IR, Batuk F, Akay AE, Baz I (2006) Automatically extracting 3D models and network analysis for indoors. In: Abdul-Rahman A, Zlatanova S, Coors V (eds) Innovation in 3D-geo information system. Springer, Berlin, pp 395–404CrossRefGoogle Scholar
  11. Kothuri R, Godfrind A, Beinat E (2010) Pro Oracle spatial for Oracle database 11g. Apress, New YorkGoogle Scholar
  12. Kwan MP, Lee J (2005) Emergency response after 9/11: the potential of real-time 3D GIS for quick emergency response in micro-spatial environments. Comput Environ Urban Syst 29:93–113CrossRefGoogle Scholar
  13. Lee J (2007) A three-dimensional navigable data model to support emergency response in microspatial built-environments. Ann Assoc Am Geogr 97(3):512–529CrossRefGoogle Scholar
  14. Mawson A (1980) Is the concept of panic useful for scientific purposes? In: Second international seminar on human behavior in fire emergencies. National Bureau of Standards, Washington, DCGoogle Scholar
  15. Murray C (2009) Oracle spatial developer’s guide, 11g release 1 (11.1). Oracle. pp 4–8Google Scholar
  16. Musliman IA, Rahman AA (2008) Implementing 3D network analysis in 3D GIS. In: International archives of ISPRS, Beijing, vol 37, Part B, Comm. 4/4Google Scholar
  17. Musliman IA, Rahman AA, Coors V (2006) 3D navigation for 3D-GIS – initial requirements. In: Abdul-Rahman A, Zlatanova S, Coors V (eds) Innovations in 3D geo information systems. Springer, Berlin/New York, pp 125–134Google Scholar
  18. Pu S, Zlatanova S (2005) Evacuation route calculation of inner buildings. In: van Oosterom PJM, Zlatanova S, Fendel EM (eds) Geo-information for disaster management. Springer, Heidelberg, pp 1143–1161CrossRefGoogle Scholar
  19. Purser DA, Bensilum M (2001) Quantification of a behaviour for engineering design standards and escape time calculations. Saf Sci 38:157–182CrossRefGoogle Scholar
  20. Sime JD (1985) Movement towards the familiar person and place affiliation a fire entrapment setting. Environ Behav 17:697–724CrossRefGoogle Scholar
  21. Wood PG (1972) The behaviour of people in fires. Fire research note, No. 953, Building Research Establishment. Fire Research Station, Borehamwood, p 113Google Scholar
  22. Zlatanova S, van Oosterom P, Verbree E (2004) 3D technology for improving disaster management: geo-DBMS and positioning. In: Proceedings of the XXth ISPRS congress, IstanbulGoogle Scholar

Copyright information

© Springer International Publishing AG 2017

Authors and Affiliations

  • Umit Atila
    • 1
  • Ismail Rakip Karas
    • 1
  • Yasin Ortakci
    • 1
  1. 1.Department of Computer EngineeringKarabuk UniversityKarabukTurkey