Regular Raptor Codes Based on LDGM with Optimal Degree Distributions for Internet of Things

We consider a coding scheme suitable for Internet-of-Things (IoT), which is extendable for future IoT rateless transmissions. In this paper, we propose new Regular Raptor codes based on low density generator matrix (LDGM) for IoT uplink wireless communications providing high reliability, which are also simple to support long battery lifetime for high-density wireless networks. We prefer a simple Regular LDGM code as a precode prior to the Luby Transform (LT) codes, called LDGM-Raptor codes, since LDGM codes have low computational complexity. To achieve high performance, indicated by minimum rate-loss, we design optimal degree distributions of LDGM-Raptor codes using extrinsic mutual information transfer (EXIT) chart. We evaluate the proposed codes under additive white Gaussian noise (AWGN) and frequency-flat Rayleigh fading channels using a series of computer simulations. We found that with the proposed Regular degree distributions the performances of LDGM-Raptor codes are good and close enough to the Shannon limit, especially when the block is long. The design is expected to be useful to support future IoT devices requiring low complexity for longer battery lifetime.