Hybrid ARQ Based Data Reliability Framework for WSNs using LDPC Codes

Abstract

Reliability is the most important feature of event-driven wireless sensor net- works. Due to the resource constrained nature of sensor networks, achieving reliability is challenging with e cient utilization of communication resources. Designing energy-e cient data communication mechanisms plays a vital role in maximizing the life-time of wireless sensor networks. Since communica- tion is accomplished through lossy error prone wireless link where errors are introduced in a packet while transmission from the source to destination. Motivated by above mentioned challenges, we propose a distributed HARQ based data reliability framework for reliable data delivery adaptable to vary- ing network conditions. Proposed error-correction framework is capable of transmitting data reliably with least amount of retransmissions. Our work is based on the fact that rate of energy depletion is higher in transmission and reception as compared to processing at sensor CPU. We integrated Low Density Parity Check (LDPC) codes in order to achieve high level of relia- bility in distributed fashion for e cient data communications. Intermediate relay nodes perform certain degree of processing in order to partially recover corrupted data. Such approach has signi cant impact on the lifetime of net- work and helps in reducing the bit error rate at destination. Framework is implemented using Crossbow Imote2. Actual on-sensor implementation of the proposed and traditional techniques enabled us to show the e cacy of our framework with fair comparison.