Upload
kayla-griffith
View
221
Download
4
Tags:
Embed Size (px)
Citation preview
Personalised Support for Reflective Learning
in Fire Risk Assessment
Wichai Eamsinvattana
Supervisors: Vania Dimitrova, School of Computing David Allen, Business School
Presentation Plan
• Motivation: Reflective Learning for Fire Risk Assessment
• Goal and Research Questions
• PORML Framework
• Contribution
• Conclusions
• Training needs– Lack of effective training models in Fire and Rescue
Services (FRS)
• Importance of risk assessment skills– Crew commanders’ risk assessment skills have major
impact on the efficiency and effectiveness of dealing with fire accidents
• Need for personalisation– Crew commanders come from diverse backgrounds,
have different experience, frequently changing jobs, and often do not have enough practice
Motivation
Disconnected from the real context
–Simulated environment–Off-site–Not very effective
Training at Fire and Rescue Services
Conditions rapidly change and various aspects (e.g. road traffic, lethal chemicals)
Previous experience is important
Crew commander’s risk assessment is crucial
Many crew commanders are inexperienced
Review of the risk assessment normally takes a lot of time (a month, 3 months or more)
(interviews with UK FRS representative)
What Happens in Reality
Problem: Can we find intelligent ways to capture the real risk assessment activities and create appropriate learning scenarios?
Our Approach: Personalised mobile reflective learning
The Problem & Our Approach
Mobile technology to support decision making at FRS
Aimed at decision makingbut learning is not considered
Virtual reality to support reflective learning
Learning by reflecting on FRS activities but disconnected from the real environment
No research has been conducted to use mobile technologies for reflective on-the-job learning.
Existing Research
(From a thesis in School of Education, University of Leeds)
Goal:• Examine how the Activity Theory can be utilised to develop
personalised mobile learning environments to support reflective on-the-job training at Fire and Rescue Services.
Goal and Research Questions
Research Questions:• Can we use Activity Theory to inform the
design of an intelligent agent that captures a user’s risk assessment experience?
• How can a holistic model of context be developed by exploiting an ontological model of generic risk assessment and semantic enhanced location information?
• Can we design context-adapted dialogue to capture a user’s risk assessment experience and promote reflection?
Dynamic Risk Assessment
Ontology
Intelligent Dialogue AgentUser Login
Service
Semantic Location Service
Game Analyzer
Dialogue Game
useuse
use
maintain
Map Properties Database
uses
convert
UK Map from Ordnance Survey
A Sample Map Area(GML file)
link p
lace
related conce
pts
select extr
act l
ocat
ion
an
d pl
acesselect record
PORML System
Interact via Internet
Position x,y(GPS)
Reflective Question
User Dialogue InteractionLog File
Web-Based User Interface
User Current Activity
User Database
Personalised On-the-job Reflective Mobile Learning (PORML) Framework
Fire Incident Site
Incident Commander
Mobile Internet Browser
Incident Commander
Reflection-on-Action
After Fire Activity Completed
Review Fire Risk Assessment Activity
Fire Incident Site
Incident Commander
Mobile Internet Browser
Incident Commander
Reflection-on-Action
After Fire Activity Completed
Review Fire Risk Assessment ActivityCapture
Semantic Location
Capturing Semantic Location
A Sample Map Area(Ordnance Survey, UK)
Incident Location
Interest Area
+ GML File
Semantic Map Data
Incident Location
Interest Area
+ GML File
Semantic Map Data
Map Properties
ID Place or Building’ s
Name
Place or Building’ s
Typeor Area (Approximately)
Place Group
Easting Northing
001 Six Bells Pub Public House 551,997.92 256,249.54 Building
002 Six Bells Car Park
Car Park 552,015 256,247 Non-Building
003 Garage1 Garage 552,028.22 256,243.33 Building
’ s ’Centre Coordinate of Object
001 Six Bells Pub Public House 551,997.92 256,249.54 Building
002 Six Bells Car Park
Car Park 552,015 256,247 Non-Building
003 Garage1 Garage 552,028.22 256,243.33 Building
Some Property Entry
Dynamic Risk Assessment (DRA Ontology)
Garage
Public House
Car Park
Link Property to Concept
Dynamic Risk Assessment
Ontology
Intelligent Dialogue AgentUser Login
Service
Semantic Location Service
Game Analyzer
Dialogue Game
useuse
use
maintain
Map Properties Database
uses
convert
UK Map from Ordnance Survey
A Sample Map Area(GML file)
link p
lace
related conce
pts
select extr
act l
ocat
ion
an
d pl
acesselect record
PORML System
Interact via Internet
Position x,y(GPS)
Reflective Question
User Dialogue InteractionLog File
Web-Based User Interface
User Current Activity
User Database
Personalised On-the-job Reflective Mobile Learning (PORML) Framework
Fire Incident Site
Incident Commander
Mobile Internet Browser
Incident Commander
Reflection-on-Action
After Fire Activity Completed
Review Fire Risk Assessment Activity
Domain Ontology (DRA Ontology)
for Fire Risk Assessment
Creating Domain Ontology for Fire Risk Assessment
Chimney Fire
Assess Hazardous Substances
Working in Roof Space Activity Level
ActionLevel
Obtain Information on Risks
Operation Level
DRA Ontology
Chimney Fire
Working in Roof Space
Crew Member
Assess Hazardous Substance
Obtain Information on Risks
haveRole
haveRole
Assess HazardousSubstance
Crew Member
Dynamic Risk Assessment
Ontology
Intelligent Dialogue AgentUser Login
Service
Semantic Location Service
Game Analyzer
Dialogue Game
useuse
use
maintain
Map Properties Database
uses
convert
UK Map from Ordnance Survey
A Sample Map Area(GML file)
link p
lace
related conce
pts
select extr
act l
ocat
ion
an
d pl
acesselect record
PORML System
Interact via Internet
Position x,y(GPS)
Reflective Question
User Dialogue InteractionLog File
Web-Based User Interface
User Current Activity
User Database
Personalised On-the-job Reflective Mobile Learning (PORML) Framework
Fire Incident Site
Incident Commander
Mobile Internet Browser
Incident Commander
Reflection-on-Action
After Fire Activity Completed
Review Fire Risk Assessment Activity
A Main Part of PORML
(Dialogue Interaction)
Using Dialogue Interaction to Capture User Experience (Activity)
Collect User and Location Information
User Current Experience
Express Utterance (Agent)
User Utterance
Update
PlanCommunicativeAct
NoNext Utterance
Check End Dialogue
Analyse
Yes
Record Dialogue Log File
User Dialogue
Interaction Log File
DRA Ontology
Reflective
Question
Template
Start to get date and time of incident, position (x,y), and incident place
Collect Context Information DG
Initial Actions DG
Explanation DG
Initial Control Measures DG
(for first situation)
Identify Risk Assessment DG(identify hazards)
Mode and System DG
Additional Control Measure DG
Explanation DG
Reflection DG
Reflection DG
Explanation DG
Explanation DG
Explanation DG
Feedback DG
Start Dialog Game (DG)
Situation Assessment DG(who was harm, risk rating)
Reflection DG
Reflection DG
Suggest Actions
Explanation DG
Explanation DG
Feedback DG
Feedback DG
Reflection DG
Nex
t Situ
atio
n
Dynamic Risk Assessment
No, I didn’t perform
Yes, I perform
Reflection1
Reflection2
No, I didn’t provide
Yes, I privide
Reflection1
Reflection2
Episode1
Episode2
Episode3
Episode4
Episode5
Episode6
Sub-Episode4
Sub
Sub
Sub
Sub
Sub-Episode
Sub
Sub
Sub
Sub
Sub
Sub
Sub
Sub
Sub
Sub
Sub
Dialogue Structure
Explain concept (term meaning) in a statement
Explain concept
Explain concept
Explain concept
No, I didn’t identify
Explain concept
Explain concept
Feedback Selection/Result
Feedback Selection/Result
PerformActions DG
End Dialog DG
Reflect Low RiskReflect Medium RiskReflect High Risk
As Fire Developed
Start to get date and time of incident, position (x,y), and incident place
Collect Context Information DG
Initial Actions DG
Explanation DG
Initial Control Measures DG
(for first situation)
Explanation DG
Reflection DG
Reflection DG
Explanation DG
Start Dialog Game (DG)
No, I didn’t perform
Yes, I perform
Reflection1
Reflection2
No, I didn’t provide
Yes, I privide
Reflection1
Reflection2
Episode1
Episode2
Episode3
Sub
Sub
Sub
Sub
Sub-Episode
Sub
Dialogue Structure
Explain concept (term meaning) in a statement
Explain concept
Explain concept
Identify Risk Assessment DG(identify hazards)
Mode and System DG
Additional Control Measure DG
Explanation DG
Explanation DG
Feedback DGSituation Assessment DG
(who was harm, risk rating)
Reflection DG
Reflection DG
Suggest Actions
Explanation DG
Explanation DG
Feedback DG
Feedback DG
Reflection DG
Nex
t Situ
atio
n
Dynamic Risk Assessment
Episode4
Episode5
Episode6
Sub-Episode4
Sub
Sub
Sub
Sub
Sub
Sub
Sub
Sub
Sub
Sub
Explain concept
No, I didn’t identify
Explain concept
Explain concept
Feedback Selection/Result
Feedback Selection/Result
PerformActions DG
End Dialog DG
Reflect Low RiskReflect Medium Risk
Reflect High Risk
As Fire Developed
PORML Dialogue Example
Collect Basic Information
Dialog: AGENT What was the incident place name you deal with?
Dialog: User It was a SixBells Pub
SixBells Pub
Confirm/Send SixBells Car Park
Garage1
Building11
Building11a
BDa
Collect Initial User and Location InformationCollect Basic Information
Dialog: AGENT How were the weather conditions during fighting the Chimney Fire?
Dialog: User Weather Rain , Wind Low , Visibility Good
Freeze High Bad
Confirm/Send Rain Low Good
Snow
Sunny
Collect Basic Information
Dialog: AGENT Which fire type do you want to assess?
Dialog: User I want to assess Chimney Fire
Building Fire
Confirm/Send Chimney Fire
Farm Fire
High Rise Building Fire
Public Entertainment Venue Fire
Rural Area Fire
Secure Accommodation Fire
Collect Basic Information
Dialog: AGENT What was the incident place name you deal with?
Dialog: User It was a SixBells Pub
SixBells Pub
Confirm/Send SixBells Car Park
Garage1
Building11
Building11a
BDa
Example on iPhone 3G Used in Summative Evaluation
• A new context modelling algorithms that capture a user’s risk assessment experience based on semantic-enhanced geographic data and an ontological model of general risk assessment activity
• Demonstration how these algorithms can be utilised in a novel pedagogical environment to promote reflective learning
• Examination of whether the new technological solutions could be deployed in the Fire and Rescue Services training practice
Contribution
• This work presents a new opportunity for personalised reflective learning using dialogue to reflect the fire risk assessment activity in emergency services based on the user’s real job experiences
• The potential benefits of modelling this assessment context are quick risk assessment linked to the real situation
• The user experiences and location environments are involved and taken into account the risk assessment activities
Conclusions
Thank you!