koorio.com
海量文库 文档专家
当前位置:首页 >> >>

SimSITE The HLARTI Based Emergency Preparedness and Response Training Simulation_图文

SimSITE: The HLA/RTI Based Emergency Preparedness and Response Training Simulation
Ke Liu1, Xiaojun Shen1, Abdulmotaleb El Saddik2, Azzedine Boukerche3 and Nicolas D. Georganas1 1 Distributed and Collaborative Virtual Environments Research Laboratory (DISCOVER) 2 Multimedia Communication Research Laboratory (MCRLab) 3 PARAllel, Distributed and Interactive Simulation of LargE scale Systems and Wireless and Mobile Networking Research Laboratory (PARADISE) School of Information Technology and Engineering University of Ottawa, K1N 6N5 Canada 1E-mail: {kliu/shen/georganas}@discover.uottawa.ca 2E-mail: abed@mcrlab.uottawa.ca 3E-mail: boukerch@site.uottawa.ca Abstract
Many wake-up calls have been received for emergency response, due to natural disasters such as hurricanes, fires or man-made incidents. The emergency responders need to work in a coordinated, well-planned manner to best mitigate the impact of an emergency incident. Simulation systems provide a wider range of training at a much lower expense for emergency preparedness and response. It is identified as the only feasible approach when it is difficult to emulate real-life experiments. The presented research demonstrates the emergency training simulation for the SITE building at the University of Ottawa, where the real-time interaction and collaboration are achieved over HLA/RTI, the IEEE standard for distributed simulation and modeling. discrete event simulation tools for planning medical resources attending the affected population. Some of these simulation systems are widely used in several countries, for example, a chemical plant fire simulation system, a radiological accident modeling for emergency response system, a nuclear weapons simulation system, etc. Simulation systems as valuable tools provide a wider range of training at a much lower expense for emergency preparedness and response, and can be used for vulnerability assessment, organizing, educating and decision support. This is identified as the only feasible approach, when it is difficult to emulate real-life experiments. Emergency preparedness and response includes all the activities of identifying, detecting, planning, analyzing vulnerability and preparing for unanticipated events which may cause injury, loss of human lives, damage and destruction of critical infrastructure elements [2]. The goal of training simulations is to provide high-fidelity training to increase the diffusion of innovative and less-invasive procedures while decreasing the trainee’s learning curve, which leads to two key issues in the design and implementation of surgical simulations, verisimilitude – accuracy/realism of simulation and appropriate training modality to facilitate skills generation and transfer by performing emulated tasks in the testing environment. 3D Virtual Reality (VR) technology provides the simulations with immersion, interactivity and information intensity, which gives us the sense of being mentally immersed in the simulation by delivering interactivity and feedback to one or more modalities. With the innovative technologies in high-fidelity VR, simulations of the actual emergency environments and user interactions can be achieved to a great extent.

1. Introduction
Many wake-up calls have been received for emergency response, due to natural disasters such as hurricanes, fires or man-made incidents, for example: the oil and chemical spills, city bombings, and the terrorist attacks. The emergency responders need to work in a coordinated, well-planned manner to best mitigate the impact of an emergency incident. There are significant demands for the need of preparedness for emergency responses both for natural and man-made disasters or other catastrophic events. A great number of research efforts have been targeted at emergency response modeling and simulation tools development. These include planning tools such as virtual representations of city landscapes, human body representations for enhanced medical response, tele-haptic[1] simulators that perform remote operations to the local objects, war theatre simulation tools which can be used for homeland security applications and

The rest of this paper is organized as follows. In section 2, simulation design is discussed. Two key technologies, collision detection and automatic avatar design are presented in section 3 and section 4 respectively. Finally, the summary of the presented demonstration and future work is given in the last two sections.

Simulation Application
Avatar Model 3DS Loader Image Loader Map Loader

2. Simulation Design
2.1 Architecture The requirements of an overall system for emergency response should be multi-facetted, real-time, and synchronized. Therefore, no single simulation model or software system will be capable of representing all the above aspects. A number of simulation tools have to be integrated to address multiple aspects of a single disaster event. A few simulation model tools can be considered, e.g. explosion simulation, fire simulation, emergency response simulation, traffic flow and evacuation modeling, and so on. In older to apply the designed training simulator on different nodes of the local network, to involve other types of simulators, and to enable the interconnection of them into distributed environments in the near future, the system architecture is to be designed in such ways to provide mechanisms to coordinate the initiation, execution, and shutdown of distributed simulations, to enable data transfers from dispersed data sources, and to provide time synchronization. Therefore, the reusability and interoperability for multiple simulations (federates) drive the system structure to be finally designed by the HLA/RTI [3] infrastructure. Figure 1 illustrates the conceptual architecture of the simulation. Facing the challenges of designing a safety training program that could simulate real-life scenarios, this research developed a virtual environment, which represents the simulation of emergency evacuation in a university building. Figure 2 presents the actual application design architecture that involves various industry standard technologies.

OpenAL OpenGL API Operating System

OpenUI

Figure 2 Simulation Development 2.2 Graphics Modeling The static virtual environment is developed by 3DsMAX7.0 in the light of the SITE building at the University of Ottawa as the prototype. The reason the SITE building is chosen is because it has complex floor plans with multiple exits and evacuation routes, where students, professors, and supporting staffs occupy different floors. The variety with this building creates different realistic scenarios. Compared with hand drawing in vertex coordinates by pure OpenGL or other programming tools, 3DsMAX developing environment allows us to create any type of objects (especially for complicated objects) in a more intuitive and human way. The architecture of the 3D SITE building was studied. The designer went to SITE building at the University of Ottawa many times. Photos of the layout were taken, the geometry data was collected, the architecture was sketched and measurement was taken to compare the model with the real entities from different view angles. Based on the 2D floor map the security office provided, a simple 3D model was built first. The model has 5 stories as shown in Figure3-(a), which was built by mesh, polygon, NURBS, surface modifier and other predefined primitives modeling tools provided by 3DsMAX7.0.

Figure 1 Conceptual Architecture of the Simulation

(a) Front View

Alarm on

On the ground

Start to play User

Timer on

Avatars find the shortest way to exits

Avatars leave the building

On the first floor Avatars find the shortest way to the stairs

Move downstairs

(b) Ground View

Figure 4 Training Scenario 1 Senario2: Emergency Evacuation Training 1. Emergency starts. The alarm is triggered and the timer starts to count the simulation time. 2. A trainee will be asked to select the floor he/she is currently located on. This will lead the trainee into either exploring state or following state. 3. If the trainee chooses the first floor, he/she enters the First Floor Exploring State to bird view this shared virtual space. 4. Otherwise the trainee enters the Ground Floor Follow State. The trainee can manipulate one avatar. The camera view perspective can be set to the first person view, the third person view, or global view to see more of the environment. His/her purpose is to manipulate the avatar to leave the building. He/she can simply walk out of the building by following other auto-avatars.
Extends

(c) Interior View Figure 3 Graphics Modeling of the SITE Building 2.3 Training Scenarios The objective of the research is to design a multi-user distributed simulator to conduct the safety training. A virtual environment is developed where multiple avatars are created. With collaborating network setup, the simulation represents different scenarios of emergency evacuation in a high-rise building. Senario1: Automatic Evacuation 1. Emergency starts. The alarm is triggered and the timer starts to count the simulation time. 2. People (auto-avatars) located on the ground floor inside of the building automatically choose the direction of movement based on optimal path calculation algorithms. 3. People (auto-avatars) on the first floor look for the entrance to the ground floor, since, except the ground, all the doors, exits, and elevators on other floors of the building will stop operating once the alarm is triggered; people move down to the ground floor and will follow step2. 4. Those people who successfully find the path and avoid all obstacles reach the exit finally to escape.

First floor explore Alarm on Timer on
Extends

Chose storey Ground floor follow
Extends

Go to exit

Trainee

Trainee Lost
Uses Uses

Check Evacuation plan

Guided by leader

Figure 5 Scenario 2 5. If he/she gets lost in the building, which means there are no auto-avatars that can be followed, the trainee has two options: ? Look at the building plan, by pressing action key “L” to load the building map on the screen (or click the Load Map button on GUI). The requested floor map will be displayed on the screen

depending on which floor the trainee is located. The map will tell the trainee his/her evacuation routes.
?

will still be applicable. We only need to change the 2D maps for corresponding structures. Here, we take the SITE ground and first floor as an example. Based on the floor construction, the information about the two floors is saved in two text files respectively. These two files correspond to our 3D virtual space projected on the XZ plane. Figure 6 shows the Map_Ground.txt file that holds the actual ground floor data: symbol "d" stands for door to the goal or exit; ‘j’ indicates free space which the avatars can go through; ‘w’ shows where the walls (obstacles) are located; We set up a two-dimensional matrix called Map Matrix in the program that contains 18x39 units. It covers the whole area of the map. Each unit has the size of 10, corresponding to the avatar’s moving steps; each avatar needs about 10 steps to walk across one unit. The data will be read from the text files and put into this matrix for manipulation later on. The collision will be checked before the user navigates the avatar. The next possible position of the avatar will be sent with a small distance range to check if the collision will happen. The small distance range should be smaller than one avatar step. If the next cell is a wall (‘w’), the collision is detected. The avatars cannot be moved, and the user has to choose other ways to go.

Wait to be guided by the leader who is managed by another player in the network.

6. The trainee will escape from the building and finishes his/her training task, either by following avatars or by map checking or leader guiding.

3. Collision Detection
Collision detection is a key issue in physically-based modeling, geometric modeling, and computer animation. Our 3D virtual environment creates a computer-generated environment filled with virtual objects. Such an environment should give the user a feeling of presence, which includes solid feeling for both surrounding objects and moving avatars. For example, the avatars should not pass through each other and other static objects. In SimSITE, interactions between moving objects are modeled by dynamic constraints and contact analysis. The avatars' motions are constrained by various interactions, including collisions. To make the system more realistic, more than thousands of objects in the virtual space need to check for possible collisions as the user moves. For that purpose, an extremely complex algorithm is needed with lots of time and memory consumption. This is not practical for running a big simulation in real time over the network. Therefore, a fast and simple interactive collision detection algorithm is used in our simulation. Since the avatars always walk on the ground (no matter on which storey), and the main obstacles (walls) are always located at fixed positions, we only need a 2D design instead of the 3D design. Ignoring the height (Y axis), we will define 2D floor maps for traveling avatars. The number of levels in the building determines how many of those 2D maps are needed. Even if the structure of the virtual environment is changed later on, all the algorithms such as path choosing and collision detection Figure 6 Ground Floor 2D Map

(a) Avatar Moving Step1

(b) Avatar Moving Step2

(c) Avatar Moving Step3 Figure 7 Collision Detection From Figure7, we can see that after taking the next step, the avatar will not move out of the cell; by adding the small distance range, no collision will happen; therefore, it will proceed to the next position; Step2 is the same as step1, the next position will be approached; in Step3, even if the next step won’t cause any collision (just arrive at the edge), by adding the small distance range, we can see that the collision will happen after Step 3; therefore, there is no point to continue moving by this direction. The avatar should be guided toward a new direction from the current position.

4. Autonomic Avatar Design
The SimSITE is a multi-avatar evacuation training simulation. Almost all the avatars should have “intelligence” to find the optimal ways to escape form the building when an emergency happens, except one that can only be controlled by the user (trainee). These programmed avatars are called “auto-avatars”, since they can calculate the shortest path from where they randomly are to one of the safe exits when emergency occurs; also they can choose an optimal path and automatically travel there with collision avoidance. Shortest Path Calculation The Map Matrix represents the 2D map described in the previous section. Now we want to find the shortest path between any free space (f) to the goal (d). Since we have many obstacles randomly sitting in the map, the axiom “the shortest distance between two points is a straight line” is not true any more. We assume the moving paths cannot be diagonals. Sometimes to walk through from a maze, the easiest way is to start from the exit. Based on this logic, the shortest path calculation algorithm is designed, using a recursive method starting from the goal. ? The problem is: Find the shortest distance and path from start point N to destination G (goal/exit) across n different hops in the emergency evacuation training scenarios ? The base case is: If N is the same as G, then the shortest distance is 0; nothing has to be done ? Reduction (a smaller case): Now suppose the number of hops between N and G is at least one. Ignore the very last step - the path between (n-1)th to (n)th, there are n – 1 hops between N and G. ? Recursion: The smaller case can be solved by a recursive performance to reach the base case eventually. ? Building up the solution: If the shortest path from the destination to the stage n-1, in addition to the known path between (n-1)th to (n)th (We can find the shortest path between two neighbor stages/hops easily), the whole path between vertex N to the destination G is able to be obtained. Similarly, how to find the path between the destination and vertex n-1? As long as we can find the path between the destination to vertex n-2, and adding the path between (n-2)th to (n-1)th; after that, it is to find the path between the destination to stage n-3. We finally reach the stage just before the base - the destination to the first vertex, which is the neighborhood of the destination, unit (x+1, z), unit (x-1, z), unit (x, z+1), and unit (x, z-1); the distance between the destination to these four units is the shortest path from the goal G to the

stage1. Eventually, by traversing the path back we got the shortest path between N and G.

Figure 8 Path Matrix Same as Map Matrix, the Path Matrix used in program contains 18x39 units. The Path Matrix keeps the calculation result of the shortest path between each vertex to the exit. For example, as shown in Figure 8, we want to find the shortest path between Point N1 and Exit G1. Starting from the exit, we first calculate the shortest path from the stage1 to the exit. The exit positioned at (x, z), the four neighboring units (x+1, z), (x-1, z), (x, z+1), and (x, z-1) in direction North, South, West and East form the first stage. Unit (x, z+1) is the wall and the other three are the free spaces. They have the shortest distance to the exit as 1. Each passed node in the map as a base will keep calculating four directions; the traversing nodes will only stop under three conditions: reach the goals, hit the wall, or meet the cell with the smaller distance value that was calculated by itself or other nodes from other branches. Next step, we want to find the shortest path between stage 2 and stage1. 8 units (x+2, z), (x+1, z-1), (x, z-2), (x-1, z-1), (x-2, z), (x-1, z+1), (x, z+2), (x+1, z+1) form stage2. Except units (x, z+2), (x, z-2), and (x-1, z+1) which are walls, all other unites have the shortest distance to the goal as 2. Same idea, the third stage contains 13 units, except wall units, (x+3, z), (x+2, z-1), (x+1, z-2), (x-1, z-2), (x-2, z-1), (x-3, z), and (x+2, z+1) have the shortest distance to the goal as 3. We continue calculation like this and ignore any wall units and the units that are out of the range; we finally find the shortest distance between each cell to the goal. By putting these distances values into the Path Matrix, the shortest path can be traversed by following the smallest descendent distance number marked on the map. For instance, from S1 to G1, two alternative paths can be chosen by following the descendent number 8, 7, 6, 5, 4, 3, 2, 1, 0.

Multi-Goal Selection Since we have more than one exit (goal) on the map, the same algorithm can be applied several times based on different exits. The cell value of the Path Matrix will be updated when the distance to the current exit is smaller than the previous one. Therefore, by picking up the smallest distance number on the map, the avatars can evacuate from any location (but has to be in free space) through the nearest exit. Path Choosing After updating all the values in Path Matrix, each avatar is assigned to a specific position. The avatar is able to decide which way it should go when it moves to the center of the cell (This is important, since the avatar’s step is much smaller than the size of each cell. After the avatar starts to move, the direction turning only happens when it reaches the center of the cell.) As shown in Figure 9, the initial position is marked value 8, which means from this cell to the nearest goal an avatar needs to cross 8 cells. The algorithm shows that it can only choose the direction to the next cell that neither contains a ‘w’ nor holds the value greater than its current value (to avoid going backward). Therefore, the only direction it can go is east. It will continue towards east until it reaches cell 6 that holds the wall as next position to the east; and then, it will turn north. By following the descendent distance number marked on the map, the avatar can finally reach the goal G1.

Figure10 Proof-of-Demonstration

6. Conclusion and Future Work
The major objective of the project is to design a collaborative training simulator for the purposes of emergency preparedness and response. The collaboration of diverse personnel and applications will provide improved opportunities for team training. The capability to train responders and commanders together on a wide range of scenarios will enable the development of effective incident management teams. These teams can build on their shared sets of experiences developed through incident management training. The shared experiences will develop the understanding of capabilities and command decisions in the team resulting in increased cohesiveness and effectiveness. The simulation industry is always evolving, which continually pushes the applications to a wider, better, and faster level. Future additions to SimSITE include, but not limited to, the ability to record the simulation in a database, better collision detection algorithm, involving more types of simulators, and cooperation with other federates.

7. Acknowledgements
The authors acknowledge the financial support of CANARIE. Figure 9 Path Selection

8. References
[1] X. Shen, J. Zhou, A. El Saddik, and N. D. Georganas , “Architecture and Evaluation of Tele-Haptic Environments”, Proc. 8th IEEE International Symposium on Distributed Simulation and Real Time Applications (IEEE DS-RT 2004), October 2004, Budapest, Hungary [2] Modeling and simulation for emergency response workshop report: http://www.mel.nist.gov/msidlibrary/doc/nistir7071.pdf [3] HLA, Defense Modeling and Simulation Office (DMSO),“High Level Architecture for Simulations Interface Specification”, Version 1.3.

5. Proof-of-Concept Demonstration
The evacuation simulation of emergency is implemented as the proof-of-concept demonstration to perform emergency training among students or employees who work in a multi-storey building with complex floor plans. The demo is tested among three labs at the University of Ottawa: DISCOVER, MCRLab and PARADISE lab.


推荐相关:

SimSITE The HLARTI Based Emergency Preparedness and Response ....pdf

SimSITE The HLARTI Based Emergency Preparedness and Response Training Simulation_专业资料。Many wake-up calls have been received for emergency response, due ...


Distributed Simulation for Emergency Management based on the ....pdf

Emergency Operations Centers (EOCs) and on-site,...Proceedings of the 11th Simulation Symposium ASIM ... A heterogeneous simula... 暂无评价 8页 免费 ...


...Simulation In Emergency Management Training Systems.pdf

Knowledge Centred Simulation In Emergency Management Training Systems_专业资料。www-version: 1.0 Knowledge Centred Simulation In Emergency Management Training ...

网站首页 | 网站地图
All rights reserved Powered by 酷我资料网 koorio.com
copyright ©right 2014-2019。
文档资料库内容来自网络,如有侵犯请联系客服。zhit325@126.com