Optimal Progressive Error Recovery for Wireless Sensor Networks using Irregular LDPC Codes

We study the problem of providing reliable data transmission in energy constrained wireless sensor networks (WSNs). Low rate channel coding can increase reliability and eliminate the need of costly retransmissions for sensor data. However, low rate channel coding on end to end basis puts a considerable burden in terms of transmit energy on resource constrained sensor nodes. We propose a scheme that progressively provides error resilience as information reaches the final destination. Precisely, we present a novel framework for processing within the network (in-network) using irregular low density parity check (LDPC) codes for channel coding, in which nodes progressively decode the information at intermediate nodes. We not only present the in-network processing setup, but also an optimal progressive error recovery algorithm (OPERA) that optimally maps the decoding iterations to attain maximum throughput at the destination node. We use density evolution algorithm for belief propagation decoding of LDPC codes to cast the optimization problem and use dynamic programming to reach the solution. We compare the performance of our scheme with end to end channel coding and establish the efficiency of proposed solution for a given energy budget. Finally, we give a comparison between our scheme and random iteration assignment for decoding at intermediate nodes, and show that our scheme performs considerably better.