In the pursuit of achieving Level 3 automated driving, the necessity for a driver’s constant availability to resume control remains crucial. Addressing this, an in-cabin smart system must effectively monitor and interpret the driver’s readiness. Current challenges include the accuracy of driver monitoring systems (DMS) in gauging attentiveness solely through eye gaze or steering wheel sensing. This may not be sufficient to assess the driver’s level of situational awareness. With a focus on multi-modal data fusion and deep learning models for simultaneous evaluation of in-cabin data and traffic scenes, these challenges can be tackled. By integrating in-cabin sensors and considering human factors, the model aims to revolutionize DMS enablers for a seamless transition between automated and manual driving. This presentation will also showcase research that emphasizes that assessing driver readiness requires a comprehensive approach beyond traditional methods, offering a promising solution to enhance the safety and efficiency of automated vehicle operation.

Discover more about:

• The critical need for accurate driver monitoring
• The challenges in current driver monitoring systems
• Multi-modal fusion and deep learning
• The role of human factor research in improving AI-based DMS
• The emergence of vision language models in DMS