Artificial intelligence (AI) is rapidly advancing across multiple industries, including aviation and aerospace. At the Institute of Aerospace Technologies at the University of Malta, we are conducting research on the safe application of AI in areas such as human-machine interaction, flight data analysis and air traffic management (ATM).
In the ARTIAP project, we are developing an AI system using intelligent agents and large language models (LLMs) to help airline pilots synthesise information, interact with automation and solve in-flight issues through a natural language interface. The system is being fine-tuned with aviation datasets - such as operation manuals - to ensure it can handle diverse queries and support pilots effectively.
For flight data analysis, we make use of the flight data monitoring (FDM) system on large commercial aircraft. This system records hundreds of parameters - such as speed, temperature and fuel flow - several times a second during flight.
Traditionally, post-flight analysis of this data - to detect abnormalities and deviations from operating procedures - is laborious and time-consuming. In the SmartFDM project, we developed machine learning (ML) tools to automatically detect anomalies in individual flights and patterns across fleets. Such tools enable airlines to address issues faster and improve efficiency.
The WAGE project took this concept a step further by performing the analysis in real-time. Live flight data is compared with historical trends and any discrepancies alert pilots to unusual aircraft behaviour. The system also performs short-term predictions to pre-empt potential failures. This predictive capability functions as an early warning system for pilots.
In ATM, we are using AI to manage the growing complexity of global air travel, ensuring safety and efficiency from takeoff to landing. In the ASTRA project, we developed a solution to predict and resolve traffic hotspots before they occur, while, in the TADA project, we are developing a solution to assist air traffic controllers to sequence aircraft around congested airports.
Across all these initiatives, our goal is not to replace humans but to provide them with decision support, reduce workload and promote human-AI teaming. In fact, in all our projects, we collaborate with end users and involve them throughout the project life cycle - from concept definition to validation. This is fundamental to build trust in AI systems.
To further promote trust, we employ explainable and transparent AI techniques and safeguards against hallucinations and other AI-related risks.
ARTIAP, SmartFDM and WAGE are financed by Xjenza Malta through the FUSION R&I Programme, while ASTRA and TADA are funded by SESAR 3 Joint Undertaking under the European Union's Horizon Europe R&I programme.
Jason Gauci is a senior lecturer and Robert Camilleri is an associate professor at the Institute of Aerospace Technologies.
A prototype space debris sensor, designed for detecting and quantifying the size and velocity of space debris. The prototype will be tested within a simulated space environment, where a 3mm ball bearing will be shot at it at 5km/s and the sensor's response will be measured. The sensor was developed through project SmartShield and financed by Xjenza Malta through the FUSION Space Upstream Programme.
* A 2023 report by the Federal Aviation Administration (FAA) estimated that there is a one-trillion-to-one probability of a piece of space debris fatally injuring a passenger during flight. In November 2022, a re-entering object caused the closure of European airspace, delaying 645 flights. While there were no incidents, it is estimated that this event had an economic impact of several millions of euros. The FAA report anticipated that the number of hazardous fragments surviving re-entries each year will reach 28,000 by 2035. Hence, despite the small probability of an aircraft being directly hit, the rapid growth in rocket launches of satellite constellations has scientists urging for actionable solutions.
For more science news, listen to Radio Mocha on www.fb.com/RadioMochaMalta/.
* On average, there were more than 35,000 flights per day in European airspace in July 2025.
* In 2024, the US military conducted the first real-life dogfight between a human-piloted F-16 and an AI-controlled fighter jet but it did not reveal who won.
* The longest-range aircraft is the A350-900ULR, with a range of 9,700 nautical miles (11,162 miles) and an endurance of more than 20 hours.
* Regional aircraft (70-100 seats) represent 36% of the total flown hours. The EU Clean Aviation programme views their short-haul operation as an ideal candidate for introducing a hybrid-electric solution.