Real-Time Overwater Wireless Network Design
Ref: CISTER-TR-230201       Publication Date: 6, Feb, 2023

Real-Time Overwater Wireless Network Design

Ref: CISTER-TR-230201       Publication Date: 6, Feb, 2023

This dissertation addresses various communication and networking challenges that arise when dealing with the design of real-time wireless networked systems operating over water environments. The end goal is to build up a comprehensive and tailored cross-layer framework that considers physical, data link, and network layer elements for improved real-time and overwater wireless communication performance by design. In this direction, we first focus on mitigating distinctive physical factors in overwater RF propagation (e.g. tides, intertidal zones) to then address upper-layers design considerations related to the improvement of traffic schedulability, i.e. the ability of the network to satisfy explicit timing constraints (e.g. packet deadlines). In particular, due to the layered structure of these research challenges, we organize our studies into two separate action domains: i) overwater, involving physical layer factors of communication, and ii) real-time, which entails both data link and network layers aspects.
In terms of overwater communication, special emphasis has been placed on investigating the impact of tides on link quality, and how tailored design strategies can help mitigate such an issue. by design. To this aim, we first investigate overwater RF propagation from the perspective of channel modeling and characterization to then capitalize this understanding into novel design methods that effectively improve link quality (e.g. received power). Specifically, we propose an approach that uses a tidal-informed two-ray propagation model to provide both static (e.g. sensor nodes) and mobile nodes (e.g. autonomous vessels) with better design options for antenna height and positioning. We also present a novel methodology for path loss prediction based on the non-trivial integration of the two-ray propagation model with precise location-dependent hydrodynamics (i.e. tidal model). A key aspect of this methodology is the ability to account for a reflective surface of varying altitude and permittivity as a function of the tide. This feature is crucial for an accurate path loss estimation over water environments with characteristic intertidal zones.
In terms of real-time communication, we aim to improve the real-time performance of globally time-synchronized wireless mesh networks (e.g. TSCH-based). Within this context, we endorse the idea of judicious \textit{gateway designation} for enhanced traffic schedulability at system's design time. Concretely, we introduce the concept of \textit{network centrality} from social network analysis (SNA) as an effective heuristic for gateway designation that facilitates real-time communication. We also generalize this idea to multiple gateways with the aid of the unsupervised learning method of spectral clustering. More importantly, we propose a novel metric termed minimal overlap centrality which conveniently exploits the relationship between path node-overlaps and gateway designation for improved network schedulability. In similar fashion, we propose a novel scheme termed minimal-overlap routing based on a new greedy heuristic for path-overlap minimization. This strategy increases traffic schedulability regardless of the gateway designation criterion.
In a nutshell, this dissertation offers various, novel, and effective cross-layer insights and methods aiming to improve the wireless communication performance of real-time overwater wireless networked systems by design. We summarize our major achievements in this direction as follows:
1) On the overwater action domain, we pioneered research studies on the large-scale fading impact of tides and intertidal zones in short and medium range shore-to-shore and shore-to-vessel RF communication links. This understanding led us to propose novel methods for link design and link-quality prediction able to mitigate/anticipate changes due to tides of up to 20 dB. Notably, our experimental measurements using commodity technologies (e.g. LoRa, WiFi) in different overwater settings (e.g. estuaries, marinas) showed major trends of the received power in agreement with the proposed methods.
2) On the real-time side, we brought to the forefront a new dimension in the design of real-time wireless sensor networks (RT-WSNs), complementing the more common approaches based on real-time scheduling or routing. Specifically, we introduced the idea of centrality-driven gateway designation and proposed a new metric resorting to the minimization of path overlaps. We showed by simulations that our novel minimal-overlap centrality achieves up to 50% better traffic schedulability than traditional metrics in SNA. Likewise, a minimal-overlap routing was proposed to enhance schedulability regardless of gateway designation.
We claim these original research findings constitute initial but promising results towards our end goal of building up a comprehensive cross-layer design framework specifically proposed to improve wireless communication performance in real-time and overwater networked systems.
Keywords: Wireless Networks, Real-Time Communication, Internet of Things.

Miguel Gutiérrez Gaitán

PhD Thesis, University of Porto.

Notes: Ph.D. Thesis. February 6, 2023. Porto. President of the Jury, Dr. Jaime dos Santos Cardoso, Full Professor at the Faculty of Engineering of the University of Porto. Members of the Jury: Dr. Thomas Watteyne, Research Director at the National Institute for Research in Digital Science and Technology (INRIA) in Paris, France; Dr. Svetlana Girs, Senior Lecturer at the Mälardalen University in Västeras, Sweden; Dr. Rui Campos, Assistant Professor at the Faculty of Engineering of the University of Porto; and Dr. Luís Miguel Pinho de Almeida, Full Professor at the Faculty of Engineering of the University of Porto (Advisor).

Record Date: 9, Feb, 2023