Evaluating the Effects of Different Layer Multiconnectivity on Reliable Multihop Industrial WSNs

Abstract

Increasing the reliability of multihop industrial wireless sensor networks (IWSNs) is an important challenge to enable continuous data collection in industrial processes. One approach for increasing reliability is multiconnectivity (MC), which enables simultaneous data transmission through two independent network links. In practice, MC can be implemented on all OSI layers, resulting in different parts of the network stack being duplicated. In this article, we compare the implementation of MC above the MAC layer and above the network layer, corresponding to an implementation on a per-hop basis or an end-to-end basis in multihop networks. For a comprehensive consideration, the approaches are discussed analytically, simulated and deployed in a real-world scenario to provide foundational knowledge about the effects of different layer MC on reliability. We show that the per-hop approach offers higher reliability than the end-to-end approach in homogeneous networks, while both approaches outperform the single-connectivity baseline. In comparison, the end-to-end approach outperforms the per-hop approach and the baseline depending on the network topology in heterogeneous networks. Apart from the topology and the link quality, we reveal the influence of fragmentation on both approaches. All influencing factors, in combination with external factors introduced by a real-world deployment, led to a decreased data loss rate of up to 56% compared to the single-connectivity baseline.