By Paula Smyczek, Faculty of Fire Safety Engineering; Piotr Tofilo, Section of Construction Safety; Adam Krasuski, Section of Computer Science, Main School of Fire Service, Warsaw, Poland
This article describes a research proposal that has been awarded the 2017 Chief Donald J. Burns Memorial Research Grant by the SFPE Foundation through a partnership with Bentley Systems. The initiative will further develop the Main School of Fire Service’s existing research to develop a system that uses simple radio beacons attached to existing smoke detectors, which uses BIM technology to allow calculation of firefighter locations in a building.
This article contains a description of the proposed research project, which focuses on the visual localization system for fire brigade using radio beacons, acceleration measurements and Building Information Modeling (BIM) technology. Reliable information about the location of firefighters in the building is vital for safe firefighting operations since it provides increased level of awareness for commanding officers who then can use available staff safely and more efficiently.
The proposed research area has been actively studied in the Main School of Fire Service for some time now (ICRA project) and a number of specific problems already have been addressed related to data handling (acquisition, monitoring, correction, adding, integration). So far, significant progress has been made and the university has developed some technical solutions that now can be further tested, improved and integrated with available MicroStation V8i program and all Bentley BIM applications solutions.
Current State of Research
The proposed system uses localization devices— radio beacons and receivers—and acceleration measurements to provide the exact location of each firefighter in a building. Radio beacons are simple electronic devices that can be attached to existing smoke detectors. The signal from beacons and the acceleration measurement is used to calculate the position of the firefighter equipped with measurements. Current real-time positions of firefighters can be presented to the commanding officer and remote command centers using a web interface and wireless connection. The key part of the visualization system is BIM technology, which provides visual context with floor layouts and important relevant information about the building.
System ICRA (more information about this system is in articles 1–4) is a wider concept that tests a number of new ideas. It has been developed with sensors allowing for data acquisition independent of the commander and firefighters.
The system includes:
- stationary and mobile sensors to increase commander’s situational awareness;
- navigation and localization technology of firefighters inside the building;
- models for estimating the location of civilians in buildings based on an analysis of signals from mobile phones;
- technology for reporting firefighter activities;
- models for estimating the size and effects of fire; and
- a buffer for sensory data.
The visual localization system for a fire brigade is used for determining and reporting the position of firefighters in real time in the building during a rescue and firefighting action. For this purpose, the system included radio beacons installed in the building (with a density of one or two beacons per room, coinciding with the density distribution of fire detectors). Changes in the propagation of the radio signal strength (RSSI) are measured and then adapted to the current readings to determine the most probable position. The navigation technology inside the building is achieved through the use of dead reckoning, which consists of estimating a traveled path from a known position (in this case, the GPS data collected by the building). The position is corrected by readings and radio fingerprinting techniques.
Figure 1: Mounted radio beacon
Figure 2: Radio beacons Figure 3: Radio receiver
The system consists of two elements:
- Radio beacon (Figure 2) – equipment installed in the building, potentially integrated with fire devices. The radio operating frequency is 867 MHz, usable at a random time unit (1.2–2 s.) as a marker.
- Radio receiver (Figure 3) – an integrated system of protective clothing, listening for radio beacons located nearby and calculating the ratio of noise to signal for them. This information is a preprocessed signal that is smoothed in time and uses an artificial neural network to determine position.
To adapt a building to a positioning system, it is necessary to provide for radio beacons in the building and collect radio fingerprinting. This consists of recording the signal strength in known positions and creating the artificial neural network based on the training exercise (such as the process illustrated in Figure 4).
Figure 4: The result of collecting radio fingerprinting in an example of a selected radio.
The data is processed in real time by a data acquisition unit integrated with clothes and transmitted to the module ICRA by radio. From there, the firefighters are getting into the system and are visualized by their commander. Their position is determined by geographical coordinates. The system operates only in the building, in the open area using the GPS receiver.
Figure 5: The position of firefighter.
Thus far, the localization system ICRA works in principle, but it has weaknesses in terms of precision of localization for both types of localization devices—radio beacons/receivers and accelerometers. The problem appears as fluctuation of the firefighter position, which is serious, because it sometimes leads to situations where a firefighter may be temporarily in a different room or sometimes on a different floor from what is shown. For practical purposes, such lack of precision may lead to bad decisions. Thus, there is a need to combine readings from both sources to increase the precision of localization. This requires developing dedicated algorithms that will use various approaches to stabilize the location of the firefighter, including probabilistic assessment of various events such as a sudden change of position or using BIM data to reduce the probability of unlikely situations such as passing through the wall or through the floor.
The Main School of Fire Service has already has had good experience with the proposed system and possesses most of the equipment needed to conduct further research. The university is a convenient place to conduct the proposed research because of its significant number of laboratories and staff members with specialties covering all areas of fire brigade activities, including firefighting, search and rescue, building fire protection, and computer science. This provides good access to required fire equipment and a modern didactic base that will be necessary during the project. Moreover, the team has the ability to perform tests on fire service officers in buildings that are suitably prepared for this process.
The plan for further research includes:
- Reduce the number of radio beacons to one ad-hoc mounted in the building by firefighters.
- Transition from one 867 MHz frequency to UWB technology using ultra wideband radar and decaWave devices
- Use magnets and magnetometer to correct Zero-Updated Velocity errors.
- Use algorithms to smooth the way firefighters move.
- Implement typical rescue procedures to correct errors.
- Change rescue tactics and procedures to better cooperate with the system.
The ability to localize firefighters precisely in a building during their activities would be a huge achievement. The results of this research will contribute to:
- significantly improving the management process during firefighting activities;
- improving search and rescue operations by ensuring they will be carried out more efficiently and effectively;
- finding firefighters if they cannot get out of a building by themselves;
- reducing the number of casualties among firefighters during their activities;
- improving infrastructure safety and fire service preparedness; and
- improving fire brigade safety during training for, responding to and operating in building emergencies.
The proposed research will be carried out within the stated timeframe and a presentation about the project will be submitted to the 2018 SFPE North America Conference and to the official SFPE journal, Fire Technology, for possible publication.
- A. Krasuski, A. Jankowski, A. Skowron, D. Ślęzak: From Sensory Data to Decision Making: A Perspective on Supporting a Fire Commander. In: Proceedings of the 2013 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT 2013). Atlanta, Georgia USA 17-20 November 2013. IEEE Proceedings, Vol. 188. Xplore (2013), pp. 237–244. DOI 10.1109/WI-IAT.2013.188.
- A. Krasuski: A Framework for Dynamic Analytical Risk Management at the Emergency Scene. From Tribal to Top Down in Risk Management Maturity Model. In Proceedings of FedCSIS'14 Conference. IEEE Xplore 2014.
- M. Meina, et al.: Tagging Firefighter Activities at the Emergency Scene. Summary of AAIA’15 Data Mining Competition at Knowledge Pit. Proceedings of the Federated Conference on Computer Science and Information Systems. ACSIS, Vol. 5, 2015, pp. 367–373. DOI: 10.15439/2015F426.
- M. Meina, A. Krasuski, K. Rykaczewski: Model Fusion for Inertial-based Personal Dead Reckoning Systems. In Proceedings 2015 IEEE Sensors Applications Symposium (SAS 2015). Zadar April 2015.