Based on an extensive field study of current firefighting practices, we have developed a system called Siren to support tacit communication between firefighters with multiple levels of redundancy in both communication and user alerts. Siren provides a foundation for gathering, integrating, and distributing contextual data, such as location and temperature. It also simplifies the development of firefighting applications using a peer-to-peer network of embedded devices through a uniform programming interface based on the information space abstraction. As a proof of concept, we have developed a prototype context-aware messaging application in the firefighting domain. We have evaluated this application with firefighters and they have found it to be useful for improving many aspects of their current work practices.

1. Siren: Context-aware
Computing for Firefighting
Xiaodong Jiang, Nicholas Chen, Jason Hong, Kevin Wang
University of California, Berkeley
Leila Takayama James Landay
Stanford University University of Washington

2. March 14, 2015 2Siren: Context-aware Computing for Firefighting
Motivation
 Firescause great harm
 1,755,000 firesin the United Statesin 1998
 4000 deaths, 25000 injuries, 100 firefighter
deaths/ year
 Information intensive task
 “Firefighting ismaking a lot of decisionson
very little information”
 Convergence of sensing and
communication

3. March 14, 2015 3Siren: Context-aware Computing for Firefighting
Research Overview
 Goal
 Support interaction between firefighters
working inside urban structures
 Approach: User-centered Design
Design Principles:
4 monthsof field study
Evaluation:
Quick feedback
Prototype Solution:
Off-the-shelf technology

4. March 14, 2015 4Siren: Context-aware Computing for Firefighting
Outline
 Field studiesresults
 From field to design
 Siren architecture
 Siren messaging application
 Evaluation
 Conclusion & Future work

5. March 14, 2015 5Siren: Context-aware Computing for Firefighting
Field Studies
 Method
 4 months, 3 depts
 Interviewsin
work environment
 1 field exercise
 2 emergency calls
 Participants
 1 assistant chief
 4 battalion chiefs
 2 captains, 2 engineers
 5 firefighters

6. March 14, 2015 6Siren: Context-aware Computing for Firefighting
Fieldwork Results
 Firefightershave an incomplete picture
 Situational assessment isdone by sending advanced
team into fires
 Individual firefighter confined to a limited space
 Current communication systemsare inadequate
 Require explicit control
 Broadcast and ephemeral
 Detached from information gathering
 Firefightersoperate in harsh environments
 Inherently unreliable nature of any infrastructure

7. March 14, 2015 7Siren: Context-aware Computing for Firefighting
Communication
“There isa lot of noise on the fire
ground. You’re inside; the fire is
burning; it makesnoise; there’s
breaking glass; there’schain saws
above your head where they’re
cutting a hole in the roof; there’s
other rigscoming in with sirens
blaring; lotsof radio traffic;
everybody trying to radio at the
same time.”

8. March 14, 2015 8Siren: Context-aware Computing for Firefighting
Need for Tacit Communication
 Context-dependent sharing of mission-
critical information
 “See through the eyesof fellow firefighters”
 Opportunistic, spontaneous, persistent
 When communication happensiscontext
dependent
 Often triggered by external events, without
human initiation
 Can be replayed and shared

9. March 14, 2015 9Siren: Context-aware Computing for Firefighting
Tacit Communication: An Example
C
A
B Abandon

10. March 14, 2015 10Siren: Context-aware Computing for Firefighting
From Field to Design
 Tacit communication
 Ad-hoc P2P
 Delayed multi-hop
 Context dependence
 Robustness: multiple levelsof redundancy
 Path redundancy
 Storage redundancy
 Version redundancy
 Feedback redundancy

11. March 14, 2015 11Siren: Context-aware Computing for Firefighting
Siren Architecture
Adapted Context Fabric (Hong et. al.) for a P2P
network of Wi-Fi-enabled Pocket PCs
Siren Modular Applications
Siren
Context Rule Engine
InfoSpace
Storage
Manager
Multi-hop
Communication
Manager

12. March 14, 2015 12Siren: Context-aware Computing for Firefighting
Storage Manager
 An InfoSpace consistsof tuples
 Each firefighter ownsan InfoSpace containing data
about himself, other peersand the environment
 Tagsto specify expiration
 InfoSpace snapshot
 A view maintained by each device of entitieson the
Siren network
 Snapshot taken by aggregating sensor readingsover
a given time window
 InfoSpace operators(insert, retrieve) unify
storage and communication

13. March 14, 2015 13Siren: Context-aware Computing for Firefighting
Multi-hop Communication Manager
 Message format
 XML over HTTP
 Semantic slots
 Time-based message queuing & routing
 One queue for each source
 Queue issorted by time
 Keep forwarding to all neighborsuntil all semantic
slotsexpire
 Version redundancy
 Not overwrite older messagesasnewer versions
arrive – why?
Message id: ….
Source: ….
Time_Created:
Semantic Slot 1:
Semantic Slot 2:
Semantic Slot 3:
Type: location
Value: 525 Soda Hall
Confidence: 70%
Expiration:

14. March 14, 2015 14Siren: Context-aware Computing for Firefighting
Context Rule Engine
 Take InfoSpace snapshotsasinput, generate
user alerts
 Production system consisting of “if-then” rules
 Currently supportsfive typesof alerts
 “dangerousplace”
 “danger to oneself”
 “next to dangerousplace”
 “othersin danger”
 “instructions”
IF (firefighter F IN room A) AND
(surrounding temperature > 1500F)
THEN (generate_alert (firefighter F1 in danger)) AND
(generate_alert (room A is a dangerous place))

15. March 14, 2015 15Siren: Context-aware Computing for Firefighting
Siren-based Messaging Application
for Firefighters
 Running on Wi-Fi-enabled Pocket PCs
 A set of context rules
 A map-based user interface
 Use of Berkeley motes
 Location beacon
 Mobile sensing
 Environmental sensing

16. March 14, 2015 16Siren: Context-aware Computing for Firefighting
Video

17. March 14, 2015 17Siren: Context-aware Computing for Firefighting
Evaluation
 Used by a group of four firefightersin their fire
house
 Positives
 Communication redundancy for large complex urban
structures(hospitals, high rises, warehouses)
 Location tracking enhance firefighter safety
 Like UI with no manual input
 Concerns
 Ruggednessof technology
 Risksof new technology
 Deployment issues

18. March 14, 2015 18Siren: Context-aware Computing for Firefighting
Ruggedness of Technology
“It’sgot to be tough; you’ve got
to be able to drop it; it’sgot to
be able to take extreme heat; it’s
got to be able to get wet.”

19. March 14, 2015 19Siren: Context-aware Computing for Firefighting
 Thermal Imaging Cameras
 PASSSystem
Risks of New Technologies

20. March 14, 2015 20Siren: Context-aware Computing for Firefighting
Deployment Issues
 Tighter integration with existing tools(use
in conjunction with thermal imager to
narrow search areas)
 Sensor deployment
 “Most effective toolsfor usare always
those we can directly accessand maintain”

21. March 14, 2015 21Siren: Context-aware Computing for Firefighting
Conclusion & Future Work
 Conclusion
 User-centered design of Siren
 Support tacit communication needs
 Siren messaging application & evaluation
 Future Work
 Dealing with uncertainty in sensor data
 Multimodal interaction techniques
 Deployment strategy

22. Thanks to:
Berkeley Fire Dept
El Cerrito Fire Dept
Alameda Fire Dept
NSFITR, CITRIS
Lawrence Leung Xiaodong Jiang
NicholasChen
Jason I. Hong
Kevin Wang
Leila A. Takayama
JamesA. Landay
http://guir.berkeley.edu/siren