A Mobile-Agent Approach for
Location Tracking in a
Wireless Sensor Network
Yu-Chee Tseng, Sheng-Po Kuo,
Hung-Wei Lee, and Chi-Fu Huang
HSCC Lab, NCTU
Agenda
Introduce Sensor Network
Tracking issue in the sensor network
Tracking protocol using mobile agents
Conclusion and future work
Sensor network introduction
A wireless ad-hoc network with thousands of
inexpensive nodes/sensors
Sensor can sense, compute and
communication
Emerge MEMS-based technology
Connect the physical world and the internet
Internet
Sensor network introduction
Research issues
Deployment problem
Coverage problem
Power saving
Tracking problem
Applications
Disaster detection
Environment control
Location-based service
Agenda
Introduce Sensor Network
Tracking issue in the sensor network
Tracking protocol using mobile agents
Conclusion and future work
Tracking issue - Positioning
Fine-grained localization
Time based methods (ToA / TDoA)
Angle based methods (AoA)
Received Signal Strength Indicator (RSSI)
Signal pattern matching
Coarse-grained localization
IR-based
RF-based
Tracking issue - Positioning
Technology
Triangulation
Trilateration
Multilateration
Tracking problem
Dynamic positioning problem
Rely on history data
Need continued collaboration between a
group of sensors
Sensor deployment of the
Tracking protocol
Regular deployment
Triangular deployment
Square deployment
Irregular deployment
Tracking protocols
Centralized
All data sent back to the centralized server
Zone-based
Data collected in zone and sent back to server
later
Agent-based
Data collected by agent and agent moves with the
object
Mobile-agent based
tracking protocol
Properties
Agent-based
Regular deployment (Triangular)
Advantage
Save power consumption
Localized
Distributed
Robust
Agenda
Introduce Sensor Network
Tracking issue in the sensor network
Tracking protocol using mobile agents
Conclusion and future work
Agents
Two kinds of agents
The master agent
One master agent will keep following the moving
object and continue assigning slaves to help it track
The slave agents
Two slave agents help the master do triangulation
Be assigned and revoked by the master agent
Environment
Protocol overview
S S
M
Protocol overview
S S
M
Protocol overview
S
S
M
Protocol overview
S
S
M
Protocol overview
M
S S
Protocol details
Election and positioning
S1 S2
S1 S2 A0
A1
r1 r2 F0
S6 S0
F1
E0
S0 r0
S5
Protocol details
Inhibit messages
Protocol details
State transition diagram
Conclusion
Our tracking protocol is localized and
distributed
Save transmission power
Future work
Specify the irregular deployment
Simulations