Novel Hybrid Long- and Short-Packet Based NOMA for Heterogeneous Data Collection in IWSNs

In this paper, we propose an uplink non-orthogonal multiple access (NOMA) based hybrid long-packet and short-packet frame structure for data collection in industrial wireless sensor networks (IWSNs) with heterogeneous quality of service (QoS) requirements of sensors. The proposed NOMA-based hybrid frame structure allows the superimposition between a number of short packets and a long packet during data collection, where the short packets can be firstly decoded to maintain the low-latency requirement, and the long packet is decoded later to achieve a high signal-noise-ratio (SNR). In addition, a joint short packet scheduling and pilot and block length optimization (JSLO) algorithm is proposed to minimize the maximum block error probabilities among short packets while maintaining a high SNR of long packet, with the assistance of optimal closed-form short packet scheduling results and pilot and block length expressions. The JSLO algorithm achieves near-optimal performance and a dramatic complexity reduction compared to exhaustive search and can lead to fast convergence. Numerical results demonstrate the proposed NOMA-based hybrid frame structure and JSLO algorithm can significantly reduce the maximum error probability among short packets and maintain a high level of fairness.