| KEYWORD |
Real-Time Pilot Activity Recognition for Civil Aviation Scenarios
keywords FLIGHT SIMULATION, PILOT DIGITAL ASSISTANT, PILOT REAL TIME MONITORING, SINGLE PILOT OPERATIONS
Reference persons GIORGIO GUGLIERI
External reference persons Dr. Nicola Puca (nicola.puca@polito.it)
Research Groups 08- Dinamica, controllo e simulazione del volo
Thesis type NUMERICAL AND EXPERIMENTAL
Description Context of Application: With an anticipated shortage of professional pilots, the aviation industry is exploring Single-Pilot Operations (SPOs) to reduce operational costs. To maintain high safety standards in a reduced-crew scenario, integrating automated assistance systems is crucial. A promising approach to virtual pilot assistance involves dynamically adjusting the level of automation based on real-time assessments of pilot behaviour, comparing detected actions against established situational procedures.
Aim of the Thesis: This thesis aims to develop a real-time activity detection algorithm to identify and classify deviations in pilot behaviour in the specific context of a civil flight mission. The algorithm must be able to process time sequences of observed events from cockpit recordings to infer whether the operator is performing the expected type of activity. A crucial aspect is leveraging mission knowledge through an established reference workflow.
One of the key outcomes of the task detection algorithm is to provide timely, context-related alerts when non-nominal events (e.g., missed or incomplete tasks) occur. An intermediate step towards the final algorithm will involve implementing pseudo real-time analyses. The tool will need to be validated against a set of pre-recorded data streams obtained during previous test campaigns.
See also real-time pilot activity recognition for civil aviation scenarios.docx
Required skills Minimum programming skills. Background in aeronautics.
Deadline 30/07/2025
PROPONI LA TUA CANDIDATURA