A Serially Concatenated NOMA scheme for Cluster-Based Vehicular Communications

In this paper we apply non-orthogonal multiple access (NOMA) in vehicular networks and propose a serially concatenated NOMA scheme for cluster-based V2X communications. The vehicles transmit to the cluster head in a nonorthogonal manner, then the cluster head communicates also non-orthogonally with the base station. We employ the code-domain features of NOMA and investigate the mappings from the vehicles to the cluster head and from the cluster head to the base station. The overall structure results in a serially-concatenated factor graph, which is beneficial for eliminating the constellation overlap due to the use of binary superposition matrices. Optimizations are carried out to minimize the average pair-wise error probability, where the optimized superposition matrices are obtained by using the genetic algorithm. Numerical results in terms of the upper bound for symbol error rate are provided, which shows the feasibility of the proposed scheme for vehicular communications.