Shortest Uplink Scheduling for NOMA-Based Industrial Wireless Networks

The power-domain nonorthogonal multiple access (NOMA) based on successive interference cancellation (SIC) provides opportunities for fast media access in industrial wireless networks. In this article, given the traffic loads of wireless sensors, we study the shortest uplink scheduling (SUS) problem by joint power allocation and wireless sensor (WS) scheduling. A key term named maximum decoded level (MDL), which models the transmitting characteristics of WSs under SIC, and thus, lays the foundations for revealing a sufficient and necessary condition for successive transmissions under SIC, is presented in the first step. Then, guided by the theoretical condition that decouples WS scheduling from power allocation, we present a two-step greedy algorithm for the SUS problem in the case of continuous transmit powers. We also prove that the proposed algorithm is optimal for two regular cases. One is for any traffic loads under 2-SIC, the other is for unit traffic load under k-SIC. Furthermore, in the case of discrete transmit powers, we further propose an optimal algorithm under 2-SIC and an approximation algorithm under k-SIC by adapting the above-mentioned greedy algorithm for the case of discrete transmit powers. Experimental evaluations reveal the effectiveness of the three algorithms.