Cognitive Radio Networks: Imagination or Reality?

1
Cognitive Radio NetworksImagination or Reality?

• Joseph B. Evans
• Deane E. Ackers Distinguished ProfessorofElectri
  cal Engineering Computer Science

2
Definitions and More Questions

• Cognitive
• Is it sense act learn?
• Or learn sense act?
• Radio
• Radios with the right capabilities exist
• But with focus on what layer(s) of the system?
• Is the focus on local subsystem and/or single
  parameter?
• Networks
• Information shared across network?
• Coordinated actions?

3

• Radios

4
KU Agile Radio (KUAR)

• Compact portable agile radio
• Consists of RF, digital, and power supply boards
• Flexible hardware
• Current KUAR supports 5 GHz band operations
• Future implementation to support 2.4 GHz band
• 30 MHz independent Tx/Rx complex baseband
• Waveform A/D 14-bit _at_ 80 Msps and D/A 16-bit _at_
  160 Msps
• PowerPC 405 (266 MHz), 32 MB SDRAM / 32 MB Flash
• Ethernet
• Xilinx Vertex II Pro 2 PPC cores 21K logic
  cells 300 MHz
• Flexible software
• Linux OS (Kernel 2.4)
• C SCA implementation from Virginia Tech (OSSIE)
  used for software framework
• KUAR control processor fully participates in
  wired network with standard network services
• Capability for dynamic service and spectrum
  access, and rapid service creation

5
Cognition and Radio
Source Preston Marshall, DARPA
6
Technology Trends
Agile Radios
Cognitive Radio Networks?
7

• Networks

8
Smarter Nodes to Smarter Networks
Across Networks
Within Devices Protocol Stacks
Within Networks
Within Devices Protocol Stacks
9
Cognitive Radio Architectures

• Cognitive radio networking capabilities
• Radio adaptation and collaboration
• Spectrum coordination for flexible wireless
  access
• Autoconfiguration
• Networking service discovery, naming, addressing
  and routing
• Cross layer network management overlay to provide
  aggregated representations of the cognitive
  subnetwork state to the future Internet

10
Cognitive Wireless Networks

• Cognitive Wireless Network with Multiple
  Network-Layer Overlays

11
The Bottom Line

• Adaptation learning are already in the network
• But
• Cognition in the network is immature
• Much work needs to be done on proving the concept
  and then exploring deployment

12

• Cognition

13
Cognition For Example, Spectrum
Source Preston Marshall, DARPA (Modified)
14
How Cognition Might Help

• Each technology can throw tough situations to
  other more suitable technologies

Source Preston Marshall, DARPA (Modified)
15
Distributed Sensing, Control, Learning

• Sharing observations from sensing
• Sharing information for reasoning and learning
• Protocols and languages?

16
Opportunity in Chaos

• It may be best to investigate from different
  parts of the protocol architecture
• Configuration management of radios in a node
  because of recent trends towards multi-radio
  systems
• Network management because it is less firmly
  embedded in the architecture

17
Final Thoughts

• Radio platforms
• They exist, with varying degrees of
  sophistication
• Adaptation in response to sensing
• Most radio systems already do this to some extent
• Learning
• It is clearly incorporated on some time scale
• But how does it evolve over time?
• How it is shared or networked?

18
Cognitive Radio NetworksImagination or Reality?