Multipath Distributed Data Reliability for Wireless Sensor Networks

In this work, we explore and evaluate data reliability in wireless sensor networks (WSNs) using a novel approach that promises to increase the net network throughput with greater energy efficiency than prevalent schemes. Our results suggest that, under investigated scenarios, proposed approach can achieve reliable communication without fast depleting energy resources of individual sensor nodes. In particular, we propose a framework in which nodes collaborate, in a distributed fashion, to provide robustness against channel induced errors as data traverses the multi-hop network. Our scheme builds on two core concepts: partial decoding of incoming packets at intermediate nodes & directed diffusion of interest over the network for selective event based reliable path reinforcement. In addition, we use path diversity with selective budgeting of network energy resources to achieve data reliability. We show that the proposed scheme outperforms conventional error robustness approaches by considerable margin for a given network energy budget. We further discuss energy-distortion tradeoffs that may lead to greater network lifetime along-with maintaining a desired level of throughput. We use iteratively decodable LDPC codes to illustrate efficiency of our scheme.