Listen to Others’ Failures: Cooperative ARQ Schemes for Low-Latency Communication Over Multi-Hop Networks

Optimization of ARQ Distribution for HARQ Strategies in Delay-Bounded Networks

“…Finally, to showcase the benefits of using CC-HARQ in a multi-hop network, we plot the minimum PDP against several values of q sum in Fig. 4 4: PDP comparison between (i) exhaustive search using original PDP expression in CC-HARQ strategy, (ii) exhaustive search using approximated PDP expression in CC-HARQ strategy, (iii) proposed low-complexity algorithm (iv) uniform ARQ distribution in CC-HARQ based strategy, and (v) ARQ distribution in non-HARQ strategy [6]. With uniform distribution, each link is first distributed with ⌊q sum /N ⌋ ARQs, and the remaining ARQs are equally shared by the first q sum mod N links.…”

“…Ongoing research in wireless multi-hop networks has already shown the challenges in achieving high end-to-end reliability with bounded delay-constraint on the packets. [5], [6] recently proposed a decode-and-forward (DF) based relaying strategy along with retransmission protocols at the intermediate nodes. In such a scheme, reliability within a hop was taken care of by the number of retransmissions allotted to a given node, and the end-to-end delay on the packets was dictated by the total number of ARQs allotted to all the nodes in the multihop network.…”

“…In such cases, advanced ARQ strategies, such as Type-2 ARQs [8] is beneficial, wherein on every retransmission, the receiver node combines the latest packet along with its previous copies to decode the information. Although an optimization problem on ARQ distribution has been studied in [5], [6] for Type-1 ARQs, there is no such effort on Type-2 ARQs to achieve high reliability in delay-bounded scenarios. Among many variants of Type-2 ARQs schemes that are often used in practical systems, a popular scheme is Chase Combining Hybrid ARQ (CC-HARQ), wherein each retransmission block is identical to the original code block [7].…”

“…In particular, we consider a multi-hop network dominated by slowly varying wireless channels with arbitrary line-of-sight (LOS) components. Along the similar lines of [5], [6], we impose an upper bound on the total number of ARQs required in the network following the CC-HARQ strategy based on the end-to-end delay requirements of the application. Subsequently, we formulate an optimization problem of allocating the optimal ARQ distribution to each intermediate link such that the packets reach the destination with high reliability.…”

“…After that, we obtain the necessary and sufficient conditions on the near-optimal ARQ distribution followed by a low-complexity algorithm to solve the optimization problem. We show that our CC-HARQ based framework outperforms the Type-1 ARQ based strategy in [5], [6].…”