OPERA: An Optimal Progressive Error Recovery Algorithm for Wireless Sensor Networks

Wireless Sensor Networks (WSNs) require robustness against channel induced errors while retransmission based schemes prove too costly for energy constrained sensor nodes. Channel coding with low rates can ensure increased reliability, and it can eliminate the need for costly retransmissions. However, low channel coding rates on an end-to-end basis (between a source sensor and a destination base-station) requires a large number of transmissions of redundant data. In this paper, we propose an Optimal Progressive Error Recovery Algorithm (OPERA) over WSNs. Under OPERA, individual intermediate sensors, which are relaying data toward the base station, partially and optimally channel-decode the incoming packets while employing a progressive decrease in parity bits as data reaches the final destination. OPERA requires significantly lesser processing than would be required for complete decoding or full decoding/encoding at the sensor nodes; and OPERA significantly reduces the total number of transmissions when compared to optimal end-to-end channel coding schemes. We use iteratively decodeable LDPC codes for this purpose. OPERA not only provides a partial processing framework, but also an algorithm to optimally map the decoding iterations over the multi hop network. We provide a comparison between our iteration assignment scheme and random iteration assignment, and show that our scheme performs considerably better. Finally, while being naturally motivated by the sensor reachback problem, we further develop a fairness-based OPERA scheme for the allocation of channel-decoding LDPC iterations to sensor nodes taking into consideration the life-expectancy of each sensor.