how to ride a sci

Download Report

Transcript how to ride a sci

Tin-Chih Toly Chen
Department of Industrial Engineering and Systems
Management
Feng Chia University
Objectives
 To share personal research interests in AmI
 Not a comprehensive survey
 Including some topics being worked on, or to be
investigated in the near future
 Hope to work with you on some topics
Introduction: Ambient Intelligence (AmI)
(1)
 coined by European Commission in 2001 (Ducatel et al.,
2001)
 sensible autonomy
 Ambient Intelligence is the vision of a future in which
environments support the people inhabiting them. This
envisaged environment is unobtrusive/transparent,
interconnected, adaptable, dynamic, embedded, and
intelligent (Cook et al., 2007; Sadri, 2011).
Introduction: Ambient Intelligence (AmI)
(2)
 interfacing with human senses rather than focusing on
computer-based input and output devices
 Sensors/detectors are embedded in everyday objects that
can communicate with each other
 Environment is sensitive to the user’s need, and even can
anticipate the user’s need or behavior.
 related topics: context-aware computing, ubiquitous
computing, pervasive computing, everywhere computing,
human/artificial intelligence, machine learning, agentbased software, robotics, etc.
Introduction: Ambient Intelligence (AmI)
(3)
 already used in everyday lives:
 thermostats
 movement sensors that control lighting
 movement sensors linked to a security alarm for detecting
intruders
Toshiba Smart TV can be controlled by hand gestures
Pattern Recognition
Human System Interaction
AmI Categories and IE Categories
 Ubiquitous Computing: Network Analysis, e-Commerce,
Mobile Commerce, RFID Applications
 Context Awareness: General Psychology, Safety and Health
Management, Human Information Processing
 Intelligence: Artificial Intelligence, Soft Computing, Data
Mining
 Natural User-system Interaction: Human Factors, Human
Computer Interaction, Occupational Psychology
Procedure of Developing
and Applying AmI (1)
Motion
decomposition
Motion analysis
Scenario generation
Law, privilege, data
security
consideration
(Huang and Chen, 2013)
Human-system
interface design
Data/message
transmission
Data analysis
algorithm
Performance
evaluation
Procedure of Developing
and Applying AmI (2)
(Cook et al., 2007)
 sense features of the users and their environment
 reason about the accumulated data
 select actions to take that will benefit the users in the
environment
(Garzotto and Valoriani, 2012)
 requirements  specifications  mockups  functional
prototypes  beta-systems
AmI System Architecture
 three layers: interface layer, flows handler layer, application
layer (Coronato and De Pietro, 2008)
 four layers: users, communication service provider, system
server, service locations (Chen and Wang, 2013)
Performance Measures of AmI
Applications (1)
 the intended goal
 cost efficiency – the costs of the establishment and
applications of the system (Sadri, 2011)
 learning efficiency – the time to learn a new rule (Sadri,
2011)
 learning completeness – the number of new rules that still
need to be learned (Sadri, 2011)
 usability (Lambrecht et al., 2011)
 process quality (Lambrecht et al., 2011)
Performance Measures of AmI
Applications (2)
 comfort – the degree of comfort that can be improved by
the system (Sadri, 2011)
Problems of the Existing Methods
 the lack of a systematic procedure of developing AmI
applications
 Cost-effect analysis of AmI applications has seldom been
done.
 Most successful AmI applications are because of massive
government support.
 The usefulness of some AmI technologies are being
questioned, e.g. e-bike.
 No one can be always successful, e.g. Siri and map
navigation of i-phone.
An Example (1)
It’s a location-aware
service (LAS) problem!
Definitions
 an instance: a user
 a state (stateful) machine, which evolves as user
interactions are caught (*)
 flows handler, which handles the interface’s state machine
and makes it evolve as interactions come from the lower
layer (*)
 context-aware system (CAS): A context-aware system uses
context to provide relevant information and/or services to
the user (Dey, 2001).
 location-aware service (LAS), which is a special CAS that
utilizes the location of the user to adapt the service
accordingly (*)
LAS in the Literature
 Chen et al. (2013) developed location-based parking
finding services for park-and-ride (PnR) facilities
provided by Australian transport authorities.
 The number of available parking spaces decreases with
time.
 Simple fuzzy rules were proposed to evaluate the
parking availability.
destination
current
position
Problem Scope
Ambient Intelligence
Contextaware
services
Location
-aware
services
scope
Ubiquitous
Mobile
commerce
computing
(Chen and Wang, 2013)
Scenario
00:00
00:01
Order lunch through mobile phone
•System detects the user’s location and speed
•Determine the JIT service location
00:05
Receive a message “go to Mac shop E”
Receive the order
Start making the lunch
02:10
Arrive at Mac shop E
Lunch is ready
4-tier
Architecture
System Server
Abstracted Road Map
1
D
1
2
0.7
B
0.8
A
I
2
2
2
4
3
3
C
F
3
3
H
0.8
1
3
E
4
2
C
G
2
D
5
J
Determining the JIT Service Location
 Procedure:
 Treat each service location as a destination, and
calculate its JIT path. Assume the JIT path length of s(m)
is j(m).

INLP formulation
 Find the m* that minimizes p – j(m*). s(m*) is the JIT
location.
 the remaining route: shortest path problem
• n nodes. The start point and
destination are nodes 1 and n,
respectively.
• lij: length of the path connecting
nodes i and j; i, j = 1 ~ n; i  j
• lij =  if there is no connection
between the two nodes
• no back path is allowed, namely, lij
=  if i > j
• di: length of the route from the
start point to node i. d1 = 0
• p: service preparation time
Modified Dijkstra’s Algorithm (1)
1.
Set d 0  0, and di   for i  0. Set the current node to the start
point.
2. Evaluate the suitability of node i as
 di
if
di  p

si   p
 0 otherwise

3.
If there is no unvisited service location in the remaining path
of the current node, go to step (4); otherwise, consider all
successors of the current node. For each successor, calculate
the distance and evaluate the suitability. Update the distance
and suitability if the suitability increases.
Modified Dijkstra’s Algorithm (2)
4. Mark the current node as visited.
5. If the highest suitability is 1, or if all service locations have been
visited, go to step (5); otherwise, set the current node to
“unvisited” and assign the highest suitability to the current
node and return to Step (3).
6. The service location with the highest suitability determines the
JIT service location. Stop.
Performance Measures (1)
 Average waiting time
 Comfort – The maximum comfort of a user is the comfort
that results from avoiding any waiting time. In this regard,
the proposed methodology is indeed effective.
 Cost efficiency – The proposed system uses a readily
available cell phone as the interface; the user does not need
to purchase additional supporting devices.
 Learning efficiency – Aside from the fuzzy Dijkstra’s
algorithm, no new rule needs to be to learned.
Performance Measures (2)
 Learning completeness – No new rules need to be learned
and the learning completeness of the proposed system is
100%.
Cost-effect Analysis (1)
(All are assumed values)
 Costs:
 Server side:

Server: 50000 NTD/2 years or 2083 NTD/month

Network connection: 1000 NTD/month

Administration & maintenance: 30000 NTD/month

Total: 2083 + 1000 + 30000 = 33083 NTD/month
 Client Side: free
 Service location: not considered
Cost-effect Analysis (2)
 Effects:
 Reduced waiting  improved customer satisfaction  increased
purchases and sales by 10%
 Original sales: 120 customers/day * 100 NTD/customer * 30.5
days/month * 10 stores = 3660000 NTD/month
 Increased sales: 3660000 * 10% = 366000 NTD/month
 Profitability: 30%
 Increased profits: 366000 * 30% = 109800
 Return on investment (ROI) = (109800 – 33083) / 33083 = 232%
IE Concepts or Techniques Applied
 Network Analysis
 the concept of Just in Time
 Mathematical Programming, Operations Research
 Fuzzy Logic, Artificial Intelligence, Soft Computing
 Mobile Commerce, e-Commerce
 System Analysis and Development
 Cost-and-effect Analysis, ROI
Topics Unsolved (1)
 A better way is needed to define the suitability/timeliness
of a path.
 Uncertainty – the positioning inaccuracy, changes in the
user’s speed, unstable network connections, humanassisted service preparation, etc.
 A user not only requires timely service, but also has to
reach his/her destination as soon as possible, which leads
to a bi-objective decision-making problem.
 To improve the efficiency of problem solving – parallel
processing
Topics Unsolved (2)
 modification of other algorithms
 Dijkstra’s algorithm
 Bellman-Ford algorithm
 A* search algorithm
 Floyd-Warshall algorithm
 Johnson’s algorithm
 How to deal with a system with multiple service locations
and multiple users at the same time must be explored.
AmI-related Journals
 IEEE Intelligent Systems Magazine, IEEE (SCI)
 Journal of Ambient Intelligence and Smart Environments
(JAISE), IOS (SCI)
 Journal of Ambient Intelligence and Humanized Computing,
Springer (EI)
 International Journal of Ambient Computing and Intelligence
(IJACI), IGI-Global (EI)
AmI-related Conferences
 European Conference on Ambient Intelligence
 International Conference on Ubiquitous Robots and Ambient
Intelligence
 Ambient Intelligence Forum
 International Scientific Conferences in the Ambience
 International Conference on Ambient Intelligence and
Ergonomics (AmI&E)
References (1)
 A. Coronato, G. De Pietro (2008) Middleware mechanisms for
supporting multimodal interactions in smart environments.
Commputer Communications, Vol. 31, pp. 4242-4247.
 M. Jeon, S. W. Lee, and Z. Bien, (2011) Hand gesture recognition using
multivariate fuzzy decision tree and user adaptation. International
Journal of Fuzzy System Applications, Vol. 1, Issue 3, pp. 15-31.
 F. Sadri (2011) Ambient intelligence: A survey. ACM Computing Surveys,
Vol. 43, No. 4, Article 36.
 T. Chen, and Y. C. Wang (2013) Establishing the fuzzy just-in-time
ubiquitous service networked system. Submitted.
 M. Huang, and T. Chen (2013) Establishing a just-in-time and
ubiquitous output system. Submitted.
References (2)
 A. K. Dey (2001) Understanding and using context. Personal and
Ubiquitous Computing, Vol. 5, pp. 20-24.
 D. J. Cook, J. C. Augusto, and V. R. Jakkula (2007) Ambient intelligence:
technologies, applications, and opportunities.
 K. Ducatel, M. Bogdanowicz, F. Scapolo, J. Leijten, and J.-C. Burgelman
(2001) Scenarios for ambient intelligence in 2010. IST Advisory Group
Final Report, European Commission, Brussels.
 J. Lambrecht, M. Kleinsorge, J. Kruger (2011) Markerless gesture-based
motion control and programming of industrial robots. 16th IEEE
international Conference on Emerging Technologies and Factory
Automation.
References (3)
 F. Garzotto, M. Valoriani (2012) “Don’t touch the oven”: motion-based
touchless interaction with household appliances. International
Working Conference on Advanced Visual Interfaces, pp. 721-724.
Thanks for your listening
Have a nice day~