Spectrum Management for Wireless Networks with Hybrid of Controls and Game Theory Approaches

1
Spectrum Management for Wireless Networks with
Hybrid of Controls and Game Theory Approaches

• Maciej Zawodniok
• Electrical and Computer Engineering
• Missouri ST
• March 2nd, 2009

2
Motivation

• Interference among the neighboring wireless
  devices
• Spectrum management becomes challenging due to
• Large number of deployed wireless network
• Uncontrolled deployment e.g. mobile, ad hoc
  networks
• Wide range of wireless systems and technologies

3
Goals and Objectives

• Context-aware spectrum management
• Fair sharing of radio resources
• Heterogeneous networks
• Different wireless technologies
• Accommodation of inaccurate or incomplete
  information
• Sensing, information from other networks, etc.
• Detection of intruders that intend to disrupt
  network operations
• Strategy selection to counter such attacks and
  ensure a high QoS

4
Proposed Approach
Game-based Decision Making and Context Discovery
Radio Spectrum
Sensing
5
Proposed Approach

• Game theory for spectrum management
• Cooperative and non-cooperative games are
  suitable for
• Behavior analysis other nodes and networks
• Strategy selection and decision making
• When to share resources? when to avoid specific
  networks? Which radio channels to use and when?
• Challenges
• Solutions for complex and nonlinear type game
  scenarios are difficult to find and to guarantee
  performance
• Inaccurate and incomplete information
• For example, interference from ordinary devices
  and malicious nodes difficult to distinguish via
  simple scanning

6
Adaptation Using Learning Methodology

• Control-based methodologies provide learning
  capabilities
• Neural networks, adaptive control, etc.
• Often used to linearize parameters of nonlinear
  systems
• Online update of control schemes parameters
  (weights) to minimize error in a dynamic
  environment
• Example Adaptive control scheme to select
  backoff interval for CSMA/CA system that meets
  desired performance 8 see the white paper
• Established analysis tools
• Lyapunov approach
• Controller and observer design

7
Verification and Validation

• Lyapunov approach
• Stability and convergence analysis
• Simulations
• Network simulators (Ns2/Ns3/OPNET) for
  performance evaluation in various scenarios
• Missouri ST Hardware platform
• Missouri ST Motes - proved to be a valuable test
  environment

Mote 2
USB Dongle
Mote 1
8
Missouri ST Motes

• Series of Missouri ST Motes used in several
  testbeds
• Evaluation of network protocols (OEDSR, AODV,
  ADFS, MMCR)
• Monitoring, diagnostics, and prognostics (M/D/P)
  applications
• RFID and asset management
• Robotics
• More on websites ims.mst.eduweb.mst.edu/mjzx9c
• Demos, publications list

M2 Mote
Mote F1
9
Questions?

• Thank you!