Robust Adaptive Transmission

1
Robust Adaptive Transmission
Greg Pottie, UCLA Electrical Engineering
Department Tel [phone] Fax (310)
206-8495
2
Research Focus

• Dealing with high channel dynamics Doppler,
  fading modify techniques proposed in other
  contexts for Orthogonal Frequency Division
  Multiplexing (OFDM)
• Objective robust communication links, at maximum
  possible rate for wide range of scenarios, using
  systolic radios
• Issues
• Jamming training sequence usually the weak link
  apply spread spectrum methods to training signal
• Complexity must implement in radios not merely
  a study of optimal forms.

3
Robust and Adaptive Transmission

• Prior work on adaptive OFDM has considered both
  time and frequency domain estimation
• Has not considered rates of channel change
  present in air/ground communications, nor jamming
• Peak to average power ratio (PAR) reduction
  techniques considered in isolation from other
  training
• Will consider multiple uses of coded training
  sequences
• Prior work on Multi-Input Multi-Output (MIMO)
  largely concerned with slowly changing links
• Will investigate how to have same adaptive
  framework for single or many antennas, and fast
  changing channel

4
Student Researcher Progress

• Equalization (S. Kim)
• Channel characterization, study of time and
  frequency domain adaptive equalization techniques
• Spread spectrum acquisition and synchronism
  issues (R. Thrasher)
• Survey of fast synchronism techniques
• Control and sensor traffic modeling (A. Pandya)
• Gathering data from formation flight experiments
• Peak power reduction techniques (H. Chen)
• Reduced complexity method

5
Channel Conditions
Air to Air Fast fading (almost frequency
non-selective, very time selective)
6
Channel Conditions
Ground to Ground Slow fading (highly frequency
selective, not very time selective)
7
Channel Conditions
Air to Ground Worst case (frequency and time
selective)
8
Channel Estimation

• Estimation in Time vs. Frequency
• Choice depends on the channel conditions
• Fast fading with short delay spread (air to air,
  air to ground)
• Time domain estimation
• Slow fading with long delay spread (ground to
  ground)
• Frequency domain estimation
• Can share training sequence among many
  transceiver functions

9
Channel Estimation

• Snap-shot estimation and interpolation

Training symbols for channel estimation
Estimated channel
Interpolated channel
10
Conventional RLS equalization
11
BER performance ( Air to Ground )
12
Two Algorithmic Levels

• Channel identification fading rate, link loss,
  interference levels, packet durationdetermines
  OFDM/spread spectrum mix and basic tracking
  approach
• Channel tracking
• Synchronization frequency and time domain
  methods
• Equalization frequency and time domain methods
• PAR reduction in-band techniques making use of
  coded training sequence
• MIMO coding and adaptive array techniques

13
Deliverables

• Channel, traffic, and physical link models for
  simulator and MAC development
• Several generations of algorithms for
  equalization, synchronization, PAR reduction, and
  use of multiple antennas, taking into account
  systolic radio architecture and MAC
• Adaptive algorithms running on the systolic radio