Unlocking Data for Smart Firefighting
It has been more than 15 years since Computer Science Professor Nalini Venkatasubramanian first realized the relevance of her research to firefighting. “We started looking at the area of information flow for emergency response soon after 9/11,” she says, referring to her work with Computer Science Professor Sharad Mehrotra on a large National Science Foundation (NSF) Information Technology Research project called Rescue. “In the process of working through the project… it came to our attention that firefighting was a domain that could use a lot of the technologies we were developing.”
Since then, they have worked on a variety of related projects, including SAFIRE, which focused on Situational Awareness for Firefighters and was funded by the Department of Homeland Security (DHS). Partnering with firefighting agencies across Southern California, they have closely aligned their research efforts in the Donald Bren School of Information and Computer Sciences (ICS) with the practical needs of various fire departments. Venkatasubramanian and her team also contributed to a National Fire Protection Association (NFPA) report, “Developing a Research Roadmap for the Smart Fire Fighter of the Future.”
Venkatasubramanian and Mehrotra’s current project, SciFire, is a collaboration with Casey Grant and Kathleen Almand of the NFPA. Through the National Institute of Standards and Technology (NIST), they recently received a second year of funding for their work to better enable infrastructure and information resilience for smart firefighting.
Delivering Data for Situational Awareness
The SciFire project continues the earlier work on situational awareness, using emerging Internet of Things (IoT) technologies to make improvements. During the project’s first year, the researchers explored how best to gather and prioritize heterogeneous sensory data from smart spaces for various stakeholders during structural fires — though, as Grant notes in a piece written for the NFPA Journal, the data could be used during other emergency situations as well. The team is now working to improve the reliability and resilience of the information gathered and distributed.
“It’s a very exciting project, but we’re taking it step by step because there are a lot of technology challenges at the device layer, network layer, information layer and human layer,” says Venkatasubramanian. “We’re taking a structured approach that addresses both infrastructure and information reliability challenges to develop meaningful technologies for firefighters in failure-prone scenarios.”
Understanding Real-World Challenges
Enthusiasm for new technology must be balanced with the realities of the job. Thanks to previous experiences participating in drills and “live burn” training exercises, Venkatasubramanian knows all too well the challenges that come with real-world deployment. “During live burns, camera visibility was reduced because of smoke,” she says. “We even experienced failure when components melted!” Solutions must consider the extreme conditions as well as the building materials and how they interfere with data transmission and communication.
Furthermore, firefighters can’t be distracted from their main objective, so any engagement with technology must be seamless. Also, because firefighters already carry a significant amount of equipment to perform their job, researchers can’t keep piling on more and more devices. As one firefighter indicated, “I’m not a Christmas tree!”
Leveraging Smart Buildings
This is where IoT can come into play, and Donald Bren Hall is serving as a testbed for the project. For over a year now, Venkatasubramanian and Mehrotra have been working with other ICS researchers, including Informatics Professor Alfred Kobsa and a team of Information Systems Group (ISG) students, to transform DBH into a smart building through the Testbed for IoT-based Privacy-Preserving PERvasive Spaces (TIPPERS) project. The TIPPERS team has built a system that manages more than 7,000 sensors in DBH and dynamically translates raw observations from those sensors into semantically meaningful interpretations (such as occupancy levels in different parts of the building or the location of people).
“Buildings are smart these days,” explains Venkatasubramanian. “An infrastructure might have data that can help during the firefighting process, but firefighters need access to that information.” The plan is to create a digital version of a Knox Box. A typical Knox Box is a small safe mounted on an outside wall that holds keys to the building, which firefighters can access during an emergency. “We’re trying to create an electronic Knox Box, an Ebox, so when there is electronic information associated with the building, firefighters can access it during a fire.” The Ebox aims to provide information such as floor plans, Wi-Fi passwords, sensor data and surveillance camera footage from the building.
In addition to creating an Ebox to gather and store this information, the team has also developed FireDex, a fire data exchange system that defines access levels and prioritizes information. “Given network constraints, it’s less important to send camera feeds of a lecture or an empty hallway than to send a visual of regions impacted by the fire,” says Venkatasubramanian. FireDex uses software-defined mechanisms to decide how to prioritize data that’s useful for firefighters and deliver that data to the appropriate people in real time.
Addressing Privacy Concerns
Another challenge with smart spaces is knowing when and how to release otherwise private information. Fortunately, the TIPPERS project is particularly well suited for this work, given its emphasis on privacy-preserving spaces.
“Privacy is a very interesting issue,” says Mehrotra. “You want to be found when a fire happens, but you don’t want to be tracked all the time. Can we design mechanisms that determine when and what to capture during an event?” One option is to tell system users during initial set up that if smoke in the building reaches a certain level, or if an alarm goes off, it will trigger the system to capture otherwise blocked location data to ensure you can be found during an emergency.
Scaling the System
The project involves taking what is learned from DBH and scaling it for high-rise buildings. “We’re building the data management technology to scale, so as the sensor data comes in for six floors, we can scale it up to 100 floors using simulation and modeling technologies,” says Venkatasubramanian. Eventually, this could be expanded to gather data from multiple buildings to address fires in larger smart spaces and in the wildland urban interface (WUI).
“The WUI, which is a little more like the Paradise fire, is different from structural files,” she explains, “but FireDex could work here as well.” The system could collect information from different sources at various locations, including sensors that capture air quality. “Getting external information about the weather and which way the wind is blowing and assimilating all that information could help firefighters more strategically battle fires and better keep themselves safe.”
So hopefully, by the time another 15 years rolls around, all of this research will have been transformed into a well-integrated tool for firefighters that offers seamless access to critical information.
— Shani Murray