Age-of-Information Dependent Random Access in NOMA-Aided Multiple-Relay Slotted ALOHA

TL;DR

The paper addresses information freshness in a two-hop relaying network with random access by proposing an age-dependent NOMA scheme for the second hop. It develops AD-NOMA-MR, derives analytical AoI expressions, and shows how phase-2 NOMA with multiple power levels can greatly reduce AoI, approaching an ideal, collision-free second hop as power levels grow. The results indicate up to 76.6% AoI improvement over the prior ADRA-MRU policy and provide design guidelines for selecting the access probability, age threshold, number of relays, and power levels. This work demonstrates the practical value of NOMA-assisted relaying in maintaining freshness for IoT applications with stringent timeliness requirements.

Abstract

We propose and evaluate the performance of a Non-Orthogonal Multiple Access (NOMA) dual-hop multiple relay (MR) network from an information freshness perspective using the Age of Information (AoI) metric. More specifically, we consider an age dependent (AD) policy, named as AD-NOMA- MR, in which users only transmit, with a given probability, after they reach a certain age threshold. The packets sent by the users are potentially received by the relays, and then forwarded to a common sink in a NOMA fashion by randomly selecting one of the available power levels, and multiple packets are received if all selected levels are unique. We derive analytical expressions for the average AoI of AD-NOMA-MR. Through numerical and simulation results, we show that the proposed policy can improve the average AoI up to 76.6% when compared to a previously proposed AD Orthogonal Multiple Access MR policy.

Figures (5)

• Figure 1: Illustration of the setup adopted in this work.
• Figure 2: Example of the evolution of the of an user. This user transmits at $t_1$, $t_2$ and $t_3$. At $t_1$ and $t_3$, the packets are successfully received at the sink, resetting the to one. On the other hand, at $t_2$, the packet is lost, and its is increased by one in the subsequent time slot.
• Figure 3: Illustration of collisions in the considered setup. At time slot $t_1$, all packets are successfully retrieved. At time slot $t_2$, there is a power collision which causes all packets to be lost.
• Figure 4: versus $K$, for the AD-NOMA-MR scheme with different power levels $L$, $N=30$ users, phase-1 erasure rate $\varepsilon_U=0.3$ and numerically optimized values of $p$ and $\delta$ (see Table \ref{['tab:opt_delta']}).
• Figure 5: Optimal $K$ (top) and the ratio of the obtained in the ADRA-MRU (with no erasure on the phase-2) policy and AD-NOMA-MR policy (bottom) with $L\in[2,4]$, as the quality of the channels deteriorate.