Driver monitoring system
A driver monitoring system that can detect if a driver is fit to drive in the case of semi-autonomous driving
The driver monitoring systems offers a unique solution to facilitate communication between a driver and an autonomous vehicle. As a result, the system enables the creation of an easily readable, fluid and safe environment to ease the transition of driving control between driver and autonomous vehicle.
3 key advantages:
High quality database: the use of numerous sensors and more than 150 simulator drivers ensured the creation of a high quality database to develop this system
Variety of models: the Fit2drive index is built upon a variety of sub-models that analyze everything from driver respiration to eye movement or steering wheel grip.
What's new?
2 important advances
First, the system benefits from a huge, qualitative database that was built using a variety of sensors: Electro-Cardio-Graphy (ECG), Electro-Dermal- Activity (EDA), Respiration Rate (RR), Eye Tracker, Thermographic camera, and a steering wheel grip sensor (developed by CEA). More than 150 drivers participated in the data acquisition phase. Second, this qualitative database enabled the development of a variety of submodels to detect the state of the driver and deliver an index representative of this state. These sub-indexes are be merged into an innovative Fit2Drive index that can be used by the Human-Machine-Interface in order to offer the driver information, warnings and recommendations for the use of autonomous driving capabilities.
|
What’s next?
3 phases of development:
- Creation of a database thanks to more than 150 test drivers and numerous sensors in our driving simulator.
Development of models to detect various driver states (visual concentration, fatigue, manual activities, etc.) in order to provide a holistic Fit2Drive index.
Underway: design and implementation of live tests in real demonstration vehicles.
Transfer of Technology expected to be feasible by mid-2023. Once models are used for car and truck markets, the system can be expanded to support other markets such as production lines that require interactions between operators and robotic machinery.
|