Evaluating Input Modalities for Pilot-Centered Taxiway Navigation: Insights from a Wizard-of-Oz Simulation

Abstract

Runway and taxiway incursions continue to challenge aviation safety, as pilots often experience disorientation from poor visibility in adverse conditions and cognitive workload in complex airport layouts. Current tools, such as airport moving maps on portable tablets, allow manual route planning but do not dynamically adapt to air traffic controllers' (ATCOs) clearances, limiting their effectiveness in high-stress scenarios. This study investigates the impact of different input modalities - paper-based, keyboard touch, map touch, and speech-to-text - on taxiway navigation performance, using a medium-fidelity flight simulator and a Wizard-of-Oz methodology to simulate ideal automation conditions. Contrary to common assumptions, recent studies indicate that paper-based methods outperform digital counterparts in accuracy and efficiency under certain conditions, highlighting critical limitations in current automation strategies. In response, our study investigates why manual methods may excel and how future automation can be optimized for pilot-centered operations. Employing a Wizard-of-Oz approach, we replicated the full taxiing process - from receiving ATCO clearances to executing maneuvers - and differentiated between readback and execution accuracy. Findings reveal that speech-based systems suffer from low pilot trust, necessitating hybrid solutions that integrate error correction and confidence indicators. These insights contribute to the development of future pilot-centered taxiway assistance that enhance situational awareness, minimize workload, and improve overall operational safety.