Advanced Driver-Assistance Systems (ADAS) enhance safety through Driver Monitoring Systems (DMS) and effective in-vehicle communication. Vision-based DMS, relying on gaze and facial cues, can detect distraction or drowsiness, while vehicle sound interaction delivers non-intrusive auditory warnings without distracting the driver. This work presents an integrated system combining head-pose-based visual distraction monitoring with spatial auditory interaction. In its current implementation, the DMS estimates overt visual distraction from calibrated head orientation in a real in-vehicle setting. The sound interaction module delivers directional auditory warnings, employing spatialization techniques, including binaural rendering and crosstalk cancellation. Experimental evaluations support the feasibility of the presented system.
Driver Attention Monitoring and Vehicle Sound Interaction: Experimental Evaluation / Generosi, A., Bruschi, V., Agostinelli, T., Terenzi, A., Cecchi, S., Mengoni, M.. - (2026), pp. 96-101. (6th IEEE International Workshop on Metrology for Automotive, IEEE MetroAutomotive 2026 San Faustino complex, ita 2026) [10.1109/MetroAutomotive69354.2026.11644633].
Driver Attention Monitoring and Vehicle Sound Interaction: Experimental Evaluation
Generosi A.;Bruschi V.;Agostinelli T.;Terenzi A.;Cecchi S.;Mengoni M.
2026-01-01
Abstract
Advanced Driver-Assistance Systems (ADAS) enhance safety through Driver Monitoring Systems (DMS) and effective in-vehicle communication. Vision-based DMS, relying on gaze and facial cues, can detect distraction or drowsiness, while vehicle sound interaction delivers non-intrusive auditory warnings without distracting the driver. This work presents an integrated system combining head-pose-based visual distraction monitoring with spatial auditory interaction. In its current implementation, the DMS estimates overt visual distraction from calibrated head orientation in a real in-vehicle setting. The sound interaction module delivers directional auditory warnings, employing spatialization techniques, including binaural rendering and crosstalk cancellation. Experimental evaluations support the feasibility of the presented system.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.


