Multi-channel network coding in tree-based wireless sensor networks

It has been recognized that using network coding improves the reliability and throughput of lossy wireless sensor networks (WSNs). Likewise, it is known that using multi-channel networking protocols can dramatically reduce end-to-end latency and improve throughput. However, to date there has been virtually no research as to how to combine these two techniques in a WSN. This paper presents MuCode, a novel tree-based multi-channel WSN protocol that uses network coding. The MuCode protocol uses a set of graph construction algorithms that support synchronized channel switching to take advantage of wireless eavesdropping for multi-path network coded packet delivery. The performance in MuCode has been evaluated against Sensecode, a standard tree-based WSN network coding protocol, as well as several other schemes. The results demonstrate that MuCode delivers superior performance for combined measures of throughput and latency.