08.02.2011
EMERGENCY
Mobile decision support in
emergency situations
Gyrd Brændeland
Locus brukerforum 2009-05-26
ICT
1
Motivation for EMERGENCY
 Improve decision support in emergency situations
 There is a need for
 Predict information needs and potential risks in emergency
situations
 This requires
 Knowledge about what information is needed in different
emergency situations
 Tools to make this information available to the operational leader
in an efficient manner
ICT
2
1
08.02.2011
The EMERGENCY project
 User-driven innovation project (BIP)
 Research is driven by end-user needs
 Typical end-users are the police, the fire service, the
ambulance service and volunteer organisations, such as
the Red Cross
 Runs from November 2008 to October 2012
 Budget: 26,3 MNOK of which 9.2 MNOK comes from the
Research Council of Norway
 Two scolarships
ICT
3
Partners
 End-user representatives
 Norwegian Red Cross
 The Directorate for Civil Protection and Emergency Planning
(DSB)
 The Police
 Research-performing partner
 SINTEF
 Industrial partners
 Geodata AS
 Locus AS
 Leading suppliers on technologies, tools and solutions relevant in the
setting of emergency planning, preparedness and operations
 Wish to extend and improve existing tools for emergency planning and
preparedness
ICT
4
2
08.02.2011
Decision support
 Decision support includes technologies and methods to
aid humans make difficult decisions
 multiple sources of information
 uncertain outcome
 Machines can keep track of a large amount of information
 Rule-based statistical predictions are better than human
reasoning in probabilistic settings
 Experience does not improve decision making when
 the same type of decision gives different result each time
ICT
5
Characteristics of emergency operations
 Need for fast and reliable action
 Decisions based on information from multiple sources
 Multiple agents involved in decision making
 Strategic
 Operational
 Tactical
 Ambulance personnel
 Police
 Fire dept
 Relieve organisations ...
 Outcome of decisions may be uncertain
 Several possible outcomes with different probabilities
ICT
6
3
08.02.2011
Example: explosion at a chemical plant
 Risk of further damage?
 Population registry
 Is this a densely populated
area?
 Should the inhabitants be
evacuated?
 Who lives nearby?
 Meteorological data, such as
wind direction and speed
 If toxic chemicals have been
released into the atmosphere
=> how are they dispersed?
 Public transport
 Are there railway/subway tunnels nearby?
 Local authorities
 Where are the sewers?
 Is there any chance chemicals may leak into the water supply?
ICT
7
ICT
8
Example: avalanche
 Limited time
 Chance of survival decreases
drastically after 18 minutes in an
avalanche
 Many uncertainties
 Risk analysis
 Is it safe to send rescue personnel into the
avalanche area?
 What is the chance of a new avalanche?
 Slope degree, snow quality, experience
 Multiple roles involved
 Police (overall responsibility)
 Operational leader in the field
 Local command
4
08.02.2011
Risk analysis in emergency operations
 Concrete risks depends on data obtained from on site
 Avalanche rescue operation
 Industrial accident
 Car crash
 But all risk assessment depends on
 Information gathering
 Experience
 Probabilistic reasoning
ICT
9
Needs in emergency situations
 Common understanding of situation among local
command
 Rescue personnel, police, etc
 Requires reliable communication
 Fast and reliable retrieval of relevant information
 Pre-emergency knowledge base to predict information needs
 Automated generation of information
 Visualisation
 Provided information must be simple to understand
 Adaptive and multi modal user interfaces
ICT
10
5
08.02.2011
Challenges
 Limited time
 Recognition-primed decision model
“When you are at the spot, you have to make do with what
you already know.”
 Difficult conditions
 Attention demanding
 Chaotic shifting situations
 Knowledge-based decision support systems are best suited for
narrow domains (frame problem)
 “The rescue leader must be creative and free himself from
written guidelines”
 Security critical information
ICT
11
Two focus areas
 Establish critical information and scenario-based decision
support prior to the emergency situation
 what type of information is required to make a decision?
 which risks are connected with a particular scenario?
 Develop solutions for efficient on-site access to
information and decision support
 how can relevant information from different sources be made
optimally accessible on mobile equipment
 how to collect and utilise experiences from emergency operations
with mobile equipment
 develop work-domain specific guidelines for the interaction
between personnel on the accident location and a mobile unit
ICT
12
6
08.02.2011
Contribute to new knowledge
 How information from risk analyses may support
personnel in taking correct decisions at the accident
location
 How relevant information from different sources can be
made optimally accessible on mobile equipment
 How to collect and utilise experiences from emergency
operations with mobile equipment
 Work-domain specific guidelines for interaction between
personnel on the accident location and a mobile unit, to
achieve optimal work support.
ICT
13
Field studies
 Developed technology in the project must be tested in the
course of the project
 Technology: methodology, procedures, or software and
user interfaces for mobile devices
 Testing is performed in case studies in collaboration with
end user groups
ICT
14
7
08.02.2011
Field studies in 2009
 Avalanche search and rescue course in Sunnmørsalpene,
18-22 March
 Outcome
 Group-based evaluation of DISKO SAR pilot – distributed command
control for search and rescue
 Established baseline
 Requirements gathering



What type of decisions do they need support for?
What type of information do they need to make decisions?
What type of functionality/hand held devices are feasible to use in the
field?
 Fire exercise in the autumn 2009
 forest wild fire exercise
 building in the city
ICT
15
Expected results of EMERGENCY
 Framework for specifying pre-emergency knowledge needed as onsite decision support in emergency situations
 Methods for risk analysis that
 predict information needs;
 develop scenarios;
 collect necessary knowledge to be available for future emergency
operations.
 Adaptation of existing software tools for risk analysis management to
the application area of emergency situations
 Tool for managing emergency work processes from risk analyses and
real-time contextual information
 Prototypes of adaptive and possibly multi modal support tools for
emergency responders on mobile devices
ICT
16
8
08.02.2011
Exploitation of results
 End user representatives will exploit the results as endusers in rescue operations in situations of emergency and
as end-users in emergency planning
 Industrial partners will exploit results to develop novel
business concepts to improve and extend existing tools
 Improve emergency centre applications for emergency planning
and preparedness
 Develop novel user interfaces that are adapted for the different
needs of different end-users
ICT
17
Progress plan
Month
Year
Activity
Documentation
1
2008
Establish advisory board.
Further detailing of plans.
1-6
2008 -2009
R&D work. Trial based on straw-man
process.
Establishment of straw-man process
based on state-of-the-art methods/tools.
Assessment report
7-12
2009
R&D work. Trial based on toolset v.1
Toolset & handbook v.1
Assessment report
Start-up of PhD-fellows
13-18
2009 - 2010
R&D work. Trial based on toolset v.2
Toolset & handbook v.2
Assessment report
19-24
2010
R&D work. Trial based on toolset v.3
Toolset & handbook v.3
Assessment report
25-30
2010 - 2011
R&D work. Trial based on toolset v.4
Toolset & handbook v.4
Assessment report
31-36
2011
R&D work. Trial based on toolset v.5
Toolset & handbook v.5
Assessment report
37-48
2011 - 2012
Final assessment
Final assessment report
ICT
18
9
08.02.2011
Kontrakter og avtaler
 Locus er kontraktspartner mot Forskningsrådet
 De øvrige partnerne inngår en samarbeidsavtale med
Locus
 Regulerer plikter og rettigheter, inkl. styringsgruppe
 Basert på Forskningsrådets ”modell” for samarbeidsavtaler
 Opp til konsortiet hvordan avtalen er utformet, men
Forskningsrådet vil ha kopi
 SINTEF har en egen kontrakt med Locus knyttet til
SINTEFs arbeid i prosjektet
ICT
19
Decision support in emergency
situations
ICT
20
10
08.02.2011
Categories of decision support systems
 Based on...
Communication
Documents
Knowledge
Human-machine
cooperation
ICT
21
Geographical information support in
emergency situations
 Geopol – police operation
central
 TransFirePC – fire and rescue
services
 SARA – the joint rescue
coordination centres
 NARRE – facilitate decision
support in crisis management
ICT
22
11
08.02.2011
End-user initiatives
 DISKO-SAR (Distributed
command control for search and
rescue)
 Live tracking of rescue personnel
in Google earth or equivalent
map tool utilising GPS
 Use of satellite pictures in forest
wild fire management
ICT
23
12