About the project
Future marine operations will increasingly rely on distributed networks of autonomous platforms. This project will investigate how acoustic measurements exchanged between uncrewed vehicles and deployed sensors can be exploited to estimate ocean environmental conditions over wide areas, combining acoustic tomography, signal processing, machine learning and environmental modelling.
Future marine operations are expected to employ large numbers of uncrewed underwater and surface vehicles operating alongside distributed environmental and acoustic sensing systems. The effectiveness of many maritime capabilities, including acoustic communications, navigation, surveillance and sonar systems, depends strongly on environmental conditions. Accurate knowledge of the ocean state is therefore critical for predicting system performance, reducing uncertainty and supporting operational decision-making. However, environmental observations are typically sparse and geographically limited, creating a need for techniques that can infer environmental conditions across much larger areas.
While conventional oceanographic sensors provide measurements at discrete locations, acoustic propagation is influenced by environmental parameters including temperature, salinity (leading to sound speed structure) and seabed properties. Acoustic signals transmitted and received across a distributed network therefore contain information about the intervening environment. Extracting and exploiting this information offers the potential to estimate environmental conditions over spatial scales far greater than can be achieved through direct sampling alone.
This project will investigate novel approaches for wide-area ocean environmental inference using observations collected by fleets of unmanned vehicles and distributed sensor networks. Drawing on concepts from underwater acoustic tomography, the research will focus primarily on signal processing, estimation and inverse methods, while also exploring the potential application of machine learning and data-driven approaches. Particular emphasis will be placed on developing techniques that are applicable to arbitrary fleets, meaning sensing networks with varying numbers of platforms, differing sensor capabilities, changing geometries and dynamic membership as assets enter, leave or manoeuvre within the network.
The project will examine how distributed acoustic observations can be fused with local oceanographic measurements to estimate parameters such as sound speed structure, temperature distribution, water column variability and other environmentally relevant characteristics. The resulting approaches could enable persistent wide-area environmental awareness and support improved prediction and optimisation of sonar, communication and autonomous system performance in future maritime operations.
