Iterative Decoding for Wireless Networks

The invention of turbo codes and low density parity check (LDPC) codes has made it possible for us for design error correcting codes with low decoding complexity and rates close to channel capacity. However, such codes have been studied in detail only for the most basic communication system, in which a single transmitter sends data to a single receiver over a channel whose statistics are known to both the transmitter and the receiver. Such a simplistic model is not valid in the case of a wireless network, where multiple transmitters might want to communicate with multiple receivers at the same time over a channel which can vary rapidly. While the design of efficient error correction codes for a general wireless network is an extremely hard problem, it should be possible to design such codes for several important special cases. This thesis takes a few steps in that direction. We analyze the performance of low density parity check codes under iterative decoding in certain simple networks and prove Shannon-theoretic results for more complex networks. More specifically, we analyze the iterative decoding algorithm in two very important special cases: (a) when the transmitter and receiver have no prior knowledge of the channel and (b) when the channel is a multiple access channel. We also apply iterative decoding to some non-LDPC codes on the binary symmetric channel and the additive white Gaussian noise channel. Finally, we derive capacity results for a class of wireless multicast networks and a class of fading channels.