The progressive automation of driving is fundamentally changing the role of the driver, from continuous vehicle control to supervision and occasional intervention. Yet, rather than making the human less relevant, increasing autonomy makes understanding the driver’s state a critical requirement for safe and effective human–vehicle interaction. Why does autonomous driving still need bioengineering? Because there is still a human to measure.
Recent advances in neurophysiological and physiological monitoring have demonstrated the potential of biosignals to continuously characterize relevant dimensions of the driver’s state, including cognitive workload, vigilance, stress, fatigue, and engagement. This talk will review selected evidence from experimental studies, including research conducted using EEG and peripheral physiological measures, highlighting what can already be reliably measured and what remains beyond the reach of current technologies.
However, demonstrating that a human state can be detected under controlled experimental conditions is only the first step towards its deployment in real-world vehicles. The transition from laboratory prototypes to automotive applications raises a set of fundamental bioengineering challenges. These include the development of unobtrusive and robust sensing technologies, reliable signal processing in the presence of motion and environmental artifacts, calibration and personalization procedures, cross-subject and cross-session generalization, and the definition of meaningful ground truths for validating physiological estimates of human state.
Ultimately, the challenge is not simply to measure the driver, but to transform noisy and highly individual biosignals into robust, interpretable, and actionable information for intelligent vehicles. Addressing these challenges will require a convergence of bioinstrumentation, biosignal processing, AI, and human factors, moving from driver monitoring towards truly human-aware autonomous vehicles.