Opportunistic Communication: A New System Design

1
Opportunistic CommunicationA New System Design

• Main reference
• Opportunistic Beamforming Using Dumb Antennas
• P. Vishwanath, D. Tse, R. Laroia, IT 2002

2
Opportunistic Communication

• One line summary
• Transmit when and where the channel is good

3
Downlink of a wireless communication
4
Traditional Approach

• Two steps
• Orthogonalize the users with respect to time,
  frequency or codes.
• Compensate for fading of individual links via
  power control, interleaving, frequency hoping
  etc.
• Goal is to create a set of reliable point to
  point links
• All 2G wireless systems (GSM,CDMA) are designed
  on this philosophy

5
Information Theoretic Result

• Total throughput achieving strategy
• At any time transmit full power only to the best
  user.
• Assumption
• Channel tracked by receiver and SIR fed back to
  BS.
• Peak transmit power constraint.

6
Fading channel is Better Than AWGN
7
Multi-user Diversity

• Diversity in wireless comm. arises form
  independent path
• Traditional forms of diversity include time,
  freq, space
• Multi-user diversity arises from independent
  fading channels across different users
• Fundamental difference
• Multi-user diversity takes advantage of rather
  than
• Compensate fading
• But it has problem Fairness and Delay

8
Proportional Fair Scheduling

• At time slot t, given
• 1) users average throughputs T1(t), T2(t),
  T3(t) in past window
• 2) current request rate R1(t), R2(t), R3(t)
• transmit to the user k with the best
• Average throughputs Tk(t) updated by an
  exponential filter

9
Limitation of Multi-user Diversity Gain

• Amount of diversity gain depends mainly on
• Range of SIR fluctuation
• Rate of SIR fluctuation in window
• Number of active users

10
A Natural Question and the key idea

• Question
• Is it possible to dictate nature to have more
  fluctuation ?
• Idea
• Artificially induce channel SIR fluctuations that
  have
• Larger range and Faster time schedule
  fluctuations

11
Random Beamforming
12
Slow Fading Environment Before
13
Slow Fading Environment After
14
Opportunistic Beamforming slow fading
Dumb antennas can approach the performance of
true Beamforming When there are many users in the
systems with less feedback.
15
Opportunistic Beamforming fast fading
Improve performance in fast fading Ricean channel
16
Overall performance Improvement
17
Comparison with space time code

• Space time code Intelligent use of transmit
  diversity to improve reliability of
  point-to-point link
• In contrast, opportunistic beamforming requires
  no special multi-antenna encoder or decoder nor
  MIMO channel estimation
• The mobiles are completely oblivious to the
  existence of multiple transmit antenna
• Antennas are truly dumb, but yet can surpass
  performance of space time code (with proportional
  scheduling)

18
Cellular Environment Opportunistic Nulling

• In a cellular systems, users are scheduled when
  their channel is strong and interference from
  adjacent base station is weak
• Multi-user diversity allows interference
  avoidance
• Transmit to user with good channel and very low
  interference
• Opportunistic beamforming combined with
  opportunistic nulling

19
Summary

• Traditional wireless communication design
  principles
• Make individual point-to-point links close to
  AWGN
• Average the interference (CDMA)
• A new design paradigms
• Modern design principle
• Large and Rapid channel fluctuations are
  preferable
• Proactive Stance
• Induce Larger and Faster Channel
  fluctuations
• Requirement
• 1) Sufficient number of users in the system
• 2) Scheduling problem