On Energy-Efficient NOMA Designs for Heterogeneous Low-Latency Downlink Transmissions
This paper investigates energy-efficient resource allocation for the two-user downlink with strict latency constraints at users. To cope with strict latency constraints, the capacity formula of the finite blocklength codes (FBCs) is adopted, in contrast to the classical Shannon capacity formula. The FBC formula explicitly specifies the trade-off between blocklength and reliability. We first consider the case where the transmitter uses super-position coding based non-orthogonal multiple access (NOMA). However, due to heterogeneous latency constraints and channel conditions at users, the conventional successive interference cancellation may be infeasible. We thus propose to use different interference mitigation schemes according to heterogeneous user conditions and solve the corresponding NOMA design problems. Though the target energy function is non-convex and implicit, optimal user blocklength and power allocation can still be identified for the considered NOMA schemes and checking the problem feasibility is simple. It is observed that when the latency requirements are more heterogeneous, the NOMA scheme cannot achieve the best energy-efficiency of the downlink. In view of this, a hybrid transmission scheme which includes time division multiple access (TDMA) and NOMA as special cases is considered. Although the energy minimization is even challenging than the pure NOMA design problem, we propose a concave approximation of the FBC capacity formula which allows to obtain computationally efficient and high-quality solutions. Simulation results show that the hybrid scheme can benefits from both NOMA and TDMA.
READ FULL TEXT