Case Studies
Selected projects in applied AI and technology development
Maritime Domain Awareness
Satellite-derived vessel risk transparency and GPS jamming detection for maritime security
The problem
Maritime security depends on knowing where vessels are, but ships engaged in illicit activity routinely disable or spoof their GPS transponders, creating blind spots for regulators and insurers.
Our approach
Funded by ESA, we are developing an Earth Observation Vessel Transparency Index (EO-VTI) that fuses satellite imagery with AIS tracking data to identify ‘shadow fleet’ vessels and characterise their behaviour. A parallel capability detects GPS jamming across monitored regions.
Outcome
In active development (2026), with an agentic AI extension – a multimodal copilot integrating satellite imagery, vessel tracks, and sanctions data – funded through a separate AWS programme. Targets maritime fleet risk assessment at scale.
Tech: Earth observation (SAR & optical), AIS data fusion, anomaly detection, agentic AI, AWS cloud
Immersive Tourism through Earth Observation
AI-driven voice-interactive Earth observation for planetariums and Virtual Reality
The problem
Planetariums and museums want to show audiences the living, changing Earth, but lack tools to transform satellite data into compelling, interactive visitor experiences.
Our approach
Funded by ESA through its Kick-Start programme, we are co-developing SpaceDome with Armagh Observatory & Planetarium: an AI-driven experience that lets audiences explore Earth observation data through voice interaction on a planetarium dome and in VR. We provide EO data pipeline design (Copernicus/Sentinel), AI integration, and show design, with participatory workshops to shape the visitor experience.
Outcome
This project is in development during April–September 2026 as a proof-of-concept. Cross-border collaboration (Northern Ireland + Republic of Ireland) demonstrates a model for joint innovation across jurisdictions.
Tech: EO data pipelines , AI/NLP voice interaction, immersive visualisation (dome, VR), participatory design with agentic orchestration
Construction Digitisation
AI-driven test orchestration for earthworks quality assurance
The problem
Earthworks testing – verifying ground compaction on every major construction project – remains almost entirely paper-based. Technicians record results by hand, training takes a decade, and quality gaps drive costly rework.
Our approach
Working with major UK construction firms, we are building a spatiotemporal test orchestration platform. It ingests 3D site models, generates testing schedules across thousands of grid-cell and layer combinations, and re-prioritises dynamically as construction progresses. An AI quality assurance module encodes senior technician expertise and flags anomalies before they escalate.
Outcome
Paid scoping exercise with industry partners complete. Innovate UK-supported development programme in progress, targeting a full digital thread from field testing through lab analysis to compliance certification.
Tech: Spatial-temporal optimisation, AI/ML anomaly detection, offline-first field capture, real-time equipment integration, 3D site modelling
Satellite Flood & Water Mapping
Open-source AI for flood mapping from space
The problem
When floods strike, response and recovery need accurate maps of inundation across larger areas, but manual satellite image analysis is slow and cannot scale to national-level events.
Our approach
Project Maji is a clean-room implementation of a satellite flood mapping pipeline, developed during the Road to SKA workshop in South Africa (February 2026). It uses a U-Net deep learning architecture to detect flood water and permanent water bodies from Copernicus Data Space imagery (Sentinel-1/2). The approach builds on the WorldFloods model and published open research by Portalés-Julià, Mateo-García, Purcell et al. (2023, Scientific Reports) as part of the ESA Frontier Development Lab programme.
Outcome
Operational and publicly available on GitHub. Now expanding to Irish catchments — the Suir and Shannon floodplains — for domestic flood risk management.
Tech: Earth observation (optical), machine learning, AWS