Concatenated Network Coding for Large-Scale Multi-Hop Wireless Networks

We present "concatenated" network coding technique that incorporates multilevel coding across a network of relays in large-scale, multi-hop, noisy wireless networks. A key feature of the proposed approach is that more powerful code can be constructed, hence more errors can be corrected, as the information travels through the network. The decomposition of large-scale network into several small-scale networks and concatenation of small-scale network codes along the route path simplifies both encoding (forwarding) and decoding processes. This is particularly important, because the nodes need to make block encoding based on local information only. This framework is inherently self scalable, in that each new node and each additional hop contribute reliability enhancement. We analyze the probability of decoding error and the normalized throughput as a function of network configuration and number of decoding iterations.