Pilot and Aircraft State Estimation for Adaptive Cockpit AI

The post-doctoral research will focus on:

• Multimodal state estimation combining physiological, behavioural and flight-system data to assess pilot workload, stress, trust, and startle responses in relation to the operational context.
  Postdoctoral Research Position: Pilot and Aircraft State Estimation for Adaptive Cockpit AI
• Machine learning and data fusion: developing and comparing machine learning approaches, including classical models, recurrent neural networks and hybrid neuro-symbolic methods, to achieve robust, interpretable and safety-aware state estimation.
• Adaptive AI assistance: evaluating how integrated team-state estimates can inform the level and type of assistance provided by the intelligent cockpit assistant, including cognitive workload mitigation and startle recovery.
• Experimental validation: assessing model performance and assistance strategies in simulation environments of increasing realism, from synthetic piloting task (MATB) to X-Plane and PEGASE flight simulators, under both nominal and non-nominal flight scenarios. Different levels of assistance will be assessed.