Multihop Wireless Networks: Opportunistic Routing

7.4 Background on Symbol-Level Network Coding

In this section, we first describe the symbol-level network coding technique. Then, we give a motivating example to show the potential advantage of exploiting symbol-level diversity in MCD in VANETs.

7.4.1 A Brief Review of SLNC

Symbol-level Network Coding was recently introduced by Katti et al. (2008) to improve the unicast throughput in wireless mesh networks. It arises from the observation that in wireless networks, even if a packet is received erroneously, some small groups of bits (“symbols”) within that packet are likely to be received correctly. SLNC gathers these correctly received (i.e., “clean”) symbols aggressively, and performs network coding on the granularity of symbols. In contrast to PLNC, SLNC gains from both symbol-level diversity and network coding. In addition, since more bit errors are tolerated than PLNC, SLNC can also gain higher throughput by encouraging more aggressive concurrent transmissions.

In general, SLNC works as follows. A symbol is defined as a group of consecutive bits in a packet, which may correspond to multiple PHY symbols of a modulation scheme. Assume the source has K packets to send, each of them expressed as a vector with elements from a Galois field . The jth symbol in a coded packet at the source ...