Title: Beamforming in Wireless Sensor Networks
1Beamforming in Wireless Sensor Networks
- Omar Bakr
- ombakr_at_cs.berkeley.edu
R. Mudumbai, J. Hespanha, U. Madhow, Scalable
Feedback Control for Distributed Beamforming in
Sensor Networks, Proc. 2005 IEEE International
Symposium on Information Theory (ISIT 2005),
Adelaide, Australia, September 2005.
2Advantages of Beamforming
- More coverage for less power
- Less interference gt improved capacity
- Basic idea use a cluster of nodes to emulate
elements of a phased array - Challenges relative locations of the different
nodes are unknown and clocks are not synchronized - Beamforming has to be done in a distributed and
scalable fashion
3Beamforming Scenario
Sender
Beamformers
Receiver
4Randomized Beamforming Algorithm
- Each node needs to apply the proper phase shift
to the signal in order to maximize the signal in
strength in a particular direction - How to choose the right phase delay? Use feedback
- Each node independently adjusts its phase delay
by ? or -? with equal probability at each time
slot - If an SNR improvement is detected, then the new
delays are used otherwise revert to the previous
delays
5Performance Evaluation (100 nodes)
6Weaknesses of this approach
- Works well only when the channel is time
invariant - May not converge if the channel varies at high
rates - Requires a noiseless feedback channel
- Not clear what beam pattern gets produced with
the random weight assignment