Driver Monitoring System
Our Driver Monitoring System reliably detects drowsiness and provides alerts. International legislation make this a mandatory safety feature starting 2024. Additionally our scalable solution also supports facial recognition, emotion detection and advanced gesture controls.
- Compliance to General Safety Regulations (GSR)
- Increased Safety e.g. by drowsiness and attentiveness detection
- Increased Comfort e.g. by driver identification using facial detection
- Modular and scalable solution also suitable for small and niche application
- EuroNCAP 4-5 star rating relevance
- Supports autonomous driving L2+ and higher, e.g. take-over request management in L3
Our modular system consists of three elements. The core builds the DMS software algorithm, which is designed to operate not only on our own ECU, but may be integrated also in 3rd party ECU or domain controller hardware.
- 1MPix Sensor
- 56° x 37° FoV Lens
- NIR LED Illumination
DMS Electronic Control Unit (ECU)
- Renesas R-Car V3M
- Linux and AUTOSAR OS
- CAN FD vehicle interface
- Algorithm data on CAN
DMS Software Algorithm
- Attentiveness and Drowsiness (GSR compliance)
- Facial recognition
- Emotion and Gesture
Driver Monitoring System in action
The system monitors the driver’s most recent behavior including head position, eye gaze (area and duration) and further signals from the car such as vehicle speed and traffic signs. Based on this information the attentiveness of the driver can be classified into a scale ranging from 1 (attentive) to 4 (distracted).
This information can then be used to dynamically adapt driver assistance functions such as blind-spot detection warnings or emergency brake assist.
The driver’s drowsiness is diagnosed based on eye blink duration and frequency as well as eye opening and closing velocity. For this, the absolute opening of left and right eye is constantly being monitored and evaluated.
Based on the Karolinska Sleepiness Scale (KSS) an easy to use four level drowsiness scale is calculated ranging from 1 (alert) to 4 (sleepy).
If drowsiness of the driver is detected, the driver can be alerted in time. Additionally this information can be used to dynamically adapt driver assistance functions such as adaptice cruise control.
Each driver can be uniquely identified using facial features such as 3d face geometry and skin texture.
New users can added using a guided enrollment procedure and idenficiation data is for privacy reasons only stored locally.
The system is robust agains glasses and tolerant to aging effects by continuous feature learning.
Example use cases
- Favorite radio channel
- Spotify playlist
- Color theme
- Seat adjustment
Business Development – Interior Electronic Functions