A crisis can occur with little or no warning, anywhere, anytime and it can happen either locally or globally. The SmartRescue project will explore how communication technology can be used in acute crisis situations.
What is the SmartRescue Project?
Mobile wireless devices such as smartphones have become a widespread and typical asset. At the same time, such devices are equipped with ever more advanced sensor technology, including accelerometer, digital compass, gyroscope, GPS, microphone, and camera. This has enabled an entirely new class of mobile device based applications that connects low-level sensor input with high-level events.
The SmartRescue project at University of Agder, Norway, have explored how this kind of communication technology can be used in acute crisis situations, where individuals need to be alerted about immediate threats, and be supported with plans for evacuating the affected area in the safest possible way. Our focus has been on the first phase of an acute and severe emergency situation in which human life and health are endangered, and where individuals for the moment are partially left to themselves, for example, because the traditional response apparatus is suppressed, delayed or paralyzed by the crisis.
Scenario: Fire in building
Our focus has been on fire in building, forecasting the extent, impact and development of the fire.
SmartRescue App Technologies
- Android app based
- Smartphone based sensing
- Data sharing through Publish-Subscribe Notification System
- Artificial intelligence techniques (Bayesian Network, K-Nearest neighbor algorithm) for threat assessment and forecasting
- Ant-based algorithms for evacuation planning
SmartRescue App Features
The app user can share sensor information through a publish-subscribe system, meaning that in a crisis situation, users can filter the sensor information they wish to know, such as peers’ locations or temperature surrounding them. It has simple activity recognition that detects someone’s movement, and senses the environment surrounding the users, for instance, if they are inside in office light, in a dark area, or outdoor.
By exploiting the sensor readings from the devices (temperature, humidity, pressure) in a fire hazard scenario, the user app can assess and predict the fire development based on a color-coding scheme. Besides, the SmartRescue app has an indoor localization feature that can track people within a building during a fire, such as victims trapped in a room. Although the app is prototyped for building fire, its applicability can be extended into other fire scenarios such as a ship or tunnel fire.
Cooperation with Local Authorities
Serious Game for Testing SmartRescue App
1. Testing the fire development feature in ISCRAM Summer School, University of Tilburg, the Netherlands, 18 August 2014
2. Testing indoor localization, fire assessment and prediction features of SmartRescue App in University of Agder, Campus Grimstad, 20 November 2014
Workshops and Dissemination
The SmartRescue Project has been involved in various workshops and dissemination events, such as:
- SmartRescue Workshop, Grimstad, June 2012
- Forskningsdagene, Grimstad, 22 September 2012
- Evaluation Workshop for Mandal Terrorist Attack Drill, Kristiansand, 23 October 2013
- CIEM Workshop, Kristiansand, 27 November 2013
- ISCRAM Summer School, University of Tilburg, the Netherlands, 13-22 August 2014
- CIEM Workshop, Kristiansand, 4 November 2014
- Digitalkonferansen Change IT, Kristiansand, 18 March 2015
- Kristiansandskonferansen, Kristiansand, 21 April 2015
Publications and Teaching
In the SmartRescue project we have published 16 peer-reviewed international conference papers and 5 scientific journal papers and 1 book chapter.
In teaching, the SmartRescue project has been used as a case for student projects, Master’s theses and PhD projects. The topics covered span a wide range of challenges, including simulating fire based on Bayesian networks, publish-subscribe systems for real-time emergency management, hazard visualization, evacuation planning, as well as fire detection and prediction using machine learning techniques such as Naïve Bayes Classifier and Decision Trees.
JOURNALS
CONFERENCE PAPERS
Simin Rasouli, Ole-Christoffer Granmo, and Jaziar Radianti (2015). “A Methodology for Fire Data Analysis based on Pattern Recognition towards the Disaster Management”. the 2nd International Conference on Information and Communication Technologies for Disaster Management (ICT-DM’2015) to be held in Rennes (France), 30 November - 2 December 2015.
Vimala Nunavath, Jaziar Radianti, Martina Comes, Andreas Prinz (2015) . The impact of ICT support on information distribution, task assignment and teams' situational awareness in search and rescue operations. The 2015 International Conference on Computing and Network Communications (CoCoNet'15). December 16-19, 2015, Trivandrum, India.
Mehdi Ben Lazreg, Jaziar Radianti, Ole-Christoffer Granmo (2015). A Bayesian Network Model for Fire Assessment and Prediction. The First International Workshop on Machine learning, Optimization and big Data - MOD 2015 Volume Editor(s): Giuseppe Nicosia, Panos Pardalos, Mario Pavone, Giovanni Maria Farinella and V. Cutello.
Radianti, J., Granmo, O., Bouhmala, N., Sarshar, P., Yazidi, A., Gonzalez, J. (2013).
Crowd Models for Emergency Evacuation: A Review Targeting Human-Centered Sensing, 2013 46th Hawaii International Conference on System Sciences, 2013 Pages: 156-165 , DOI: 10.1109/HICSS.2013.155 (Nominated for the best paper award)
Ole-Christoffer Granmo, Jaziar Radianti, Morten Goodwin, Julie Dugdale, Parvaneh Sarshar, Sondre Glimsdal, Jose J. Gonzalez (2013).
A Spatio-temporal Probabilistic Model of Hazard and Crowd Dynamics in Disasters for Evacuation Planning, Recent Trends in Applied Artificial Intelligence, Lecture Notes in Computer Science Volume 7906, 2013, pp 63-72.
BOOK CHAPTERS