Despite all the technological enhancements to gain L3 automated driving, a driver should be ‎still available at all times to resume the control of an automated vehicle, in response to a ‎critical takeover request. In such vehicles, an in-cabin smart system or enabler should ‎monitor the driver and ensure the driver is ready to safely resume driving control. ‎However, there are two fundamental and challenging questions to be answered in this ‎domain:‎

How a driver monitoring system can accurately understand and interpret the driver’s level of readiness or attentiveness ‎using in-cabin sensors?‎

Can current DMS solutions (eye gaze, head pose) or steering wheel sensing ‎technology provide sufficient information about the actual state of the ‎driver/occupants?‎

Accurate understanding and measurement of driver readiness is not a trivial task. To have ‎a seamless transition between automated driving mode and human driving we need to ‎look further and develop the next generation of DMS enablers to fit the purpose. Many ‎recent studies confirm a green light to resume the control can not be issued solely based ‎on eye gaze or steering wheel sensing. In this session, we discuss a broader view ‎of ‎requirements for assessing driver readiness, using both vision sensors ‎and ‎human factors criteria. We also propose a new multi-modal feature fusion and deep ‎learning solution to ‎address one of the current technological challenges in this area.‎