Develop novel algorithms for multi-agent reinforcement learning (MARL), enabling multiple autonomous assets to learn how to interact and collaborate towards a specified task.
Use of multiple autonomous vehicles, whose fused measurements can be used to provide high-quality survey data within the marine environment.
Objectives
Controlling a collection of unmanned autonomous vehicles poses the two key questions, listed below. Addressing these would be the key objectives to address during this research program.
Monitor radiated acoustic and non-acoustic emissions from maritime platforms using a combination of in-situ sensors fixed to the platform and remote autonomous vehicles operating in the marine environment around the platform.
Develop analysis strategies for deriving understanding from data collected from multiple platforms, with different sensors, sensing different properties of the environment, over multiple spatial and temporal scales.
Co-design AI learning algorithms and learning theoretic guarantees so to provide deep learning systems with strong empirical performance as well as theoretical guarantees.
This project will develop new machine learning inspired algorithms to extract, communicate and fuse information from networks of sensors, converting sensing into actionable information.
This project seeks to implement a dImplementation of differentiable physics engine for distributed motion planning in multi-agent loco-manipulation tasks.
The aim of this project is to enhance the coordination capabilities of autonomous agents during the deployment, either with other autonomous platforms or with human operators.
The central aim of this research is to investigate how foundation models can be effectively leveraged to enhance decision-making capabilities in single autonomous agents operating within complex, real-world environments.
Humans with experience are very able to know when something is going wrong with the system they are using and to compensate for it. In autonomous system this intuition needs to be built into the algorithmic design.
The objective is to mix filtering methods with optimisation and deep learning to design advanced efficient tracking methodologies, with strong theoretical guarantees.
The project aims to develop new computational algorithms to optimize the acquisition and processing of event data (e.g., single-photon data) and enable fast imaging in extreme environmental conditions using statistical and learning based approaches.
This project will develop novel Artificial Intelligence (AI) techniques for position location, considering both the algorithm’s performance and its ability to be implemented efficiently in hardware.
We will apply state-of-the-art signal processing techniques (e.g. online sequential Bayesian inference, reinforcement learning) to make spin-based quantum sensors faster, more robust against noise and changing environments, and more user-friendly, with the goal of detecting weak nanoscale magnetic resonance signals from small ensembles of molecules.
In this work we will deliver highly original solutions to two important problems in our society, i.e., the physical limitations of data-storage platforms, and their booming energy demand.
Co-designing advanced CI algorithms and computing hardware to reconstruct images with a new regime of precision, robustness, efficiency, and scalablity, enabling in particular addressing low-resource setting applications, with a specific focus on low-SWAP platforms.