Performance of Low Rank Parity Check Codes for Multi-source Wireless Sensor Networks
Oussama Habachi, Vahid Meghdadi, Jean‐Pierre Cances, Imad El Quachchach · HAL (Le Centre pour la Communication Scientifique Directe) · 2020
Random linear network coding (RLNC) is one of the most promising high-performance techniques for wireless sensor networks (WSNs). One of its most relevant characteristics is that it allows intermediate nodes to combine incoming packets and send only the combined packets. Nevertheless, few works have focused on the use of multi-source networks using error correcting codes. In fact, when an intermediate node fails, errors may occur and since RLNC combines packets from different sources, several packets may be affected. Furthermore, we consider a realistic model for the wireless links in the network taking into account the ambient thermal noise. In this paper, we consider the problem of multi-source networks and we propose a novel error correction mechanism using modified low rank parity check (M-LRPC) decoding algorithm based on LRPC code, as an outer code and a convolutional code as an inner code to correct sparse errors. Moreover, we investigate the performance of the proposed coding technique in terms of success decoding rate. Then, we derive a theoretical expression for the decoding probability of the proposed M-LRPC. Simulation results show the accuracy of the proposed bound. These results prove that the proposed scheme significantly improves the decoding probability compared to Gabidulin codes.