Answers are grounded in this paper and cite their sources.
Underwater communication presents significant challenges due to factors such as signal attenuation, absorption, multipath propagation, limited bandwidth, high latency, marine noise and fluctuating channel conditions that make data collection difficult in underwater environments. This work reviews the data collection methods via (i) multi‐hop routing and (ii) autonomous underwater vehicles (AUVs) in underwater acoustic sensor networks (UASNs), which are the subset of internet of underwater things (IoUT). We present an in‐depth analysis of routing protocols developed over the past two decades (2006–2026) with their advantages and limitations. In addition, we discuss the role of reinforcement learning (RL), a core subfield of machine learning, in enhancing routing efficiency by optimising key metrics such as next‐hop selection, energy consumption, packet delivery ratio and network lifetime. Additionally, our work investigates the evolution of acoustic modems from traditional to software‐defined and the capabilities of emerging multimodal acoustic modems. Furthermore, we evaluate the capabilities of simulation tools and underwater model libraries available to replicate real‐world scenarios. Our analysis presents the need for improved model libraries that include hardware‐accurate profiles. This work bridges the gap between theoretical modelling and practical underwater communication requirements by identifying the key research gaps in channel, energy, noise modelling, medium access control (MAC) layer support and RL learning capabilities, along with the integration of next‐generation communication technologies, and provides future research directions for data collection in the IoUT environment.
No takes yet. Share an insight, caveat, or question.
R et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: