Distributed Space-Time Coding

Recently, attention has been focused on wireless networks. Methods to exploit diversity using the antennas of different users in a cooperative way have been studied, in order to look for the diversity known to be achieved in the multiple antennas point to point case. We consider a snapshot of a wireless network, where a transmitter and a receiver are distinguished, with enough computational capacity and power to encode/decode the data. The other nodes are relays, small devices with few power and computational ability, with only one antenna (as for example in sensor networks). In this scenario, it is not realistic to assume nodes can decode the information. We thus propose a strategy based on the idea of Distributed Space-Time Coding, where the nodes just do a very simple operation before forwarding the information. This operation has the effect of having the nodes cooperating to encode a virtual space-time code, thus the name of distributed space-time coding.

Designing codes for multiple antennas at both the transmitter and receiver nodes.

We are interested in designing distributed space-time codes, for the case when the transmitter and receiver nodes have several antennas. Such scenario has for goal to increase the data rate of the network. We present a scheme which relies on cyclic division algebras. It is suitable for any number of transmit/receive antennas and nodes. It is shown to perform better than random codes, with less encoding complexity.
We finally analyze the behaviour of the codes when some nodes are down, and show that the codes are resistant to node failures.
A scheme for the particular case when the transmitter and receiver only have one antenna is presented in

Designing codes when no channel information is available.

The above strategy requires the receiver node to know all the channel paths involved in the communication process. This assumption may not be realistic. In
a strategy is proposed to code over a wireless network with no channel knowledge. More precisely, the relay nodes operate without knowing the fadings that affected their received signal, and the receiver decodes knowing none of the channel paths.
This strategy has been generalized to multiple antennas in
The proof of the generalization, as well as further results are presented in

Joint work with Prof. B. Hassibi

Last update: March 8 2007