Training
ADAS Systems: sensing, computing, actuation
From sensor to actuator: how an ADAS function perceives, decides and acts on the vehicle.
- Duration
- 2–3 days
- Format
- On-site · modular · hands-on workshops on real cases
This training follows the complete ADAS chain: sensors, fusion, planning and control, then actuation through steering, propulsion and braking. Content is adapted to the participants’ programs and to the functions they develop or integrate. Workshops use realistic cases: function specification, actuator interfaces and validation scenarios.
Program
01
ADAS architecture & functions
- Function landscape: ACC, AEB, LKA, lane change assist, park assist
- Levels of driving automation (SAE J3016)
- Centralized and distributed E/E architectures for ADAS
- Function chain: sensing, computing, actuation, HMI
02
Sensing
- Radar: FMCW principle, range, velocity and angle resolution
- Mono and stereo cameras, lane and object detection, ML-based perception
- Ultrasonic sensors for parking and low-speed functions
- Sensor calibration, blockage, misalignment and weather effects
03
Computing: fusion & environment model
- Multi-object tracking with Kalman filters
- Object-level fusion, data association, track management
- Environment model: objects, lanes, free space
- Situation analysis: target selection, time-to-collision
04
Planning & control
- Path planning and trajectory generation under comfort constraints
- ACC speed and distance control; ISO 15622 performance requirements
- AEB warning and braking cascade
- Lateral control for lane keeping and lane centering
05
Actuation & arbitration
- Steering interfaces: angle and torque overlay requests
- Braking and propulsion interfaces: deceleration and torque requests
- Driver override, arbitration and degradation modes
- Workshop: specify the ACC-to-brake interface and its fallback
06
Safety, validation & regulations
- SOTIF (ISO 21448) and ISO 26262 applied to ADAS
- Scenario-based testing across MIL, SIL, HIL and VIL
- Euro NCAP protocols, UN R152, UN R79 (ACSF), UN R171
- ISO/PAS 8800 for machine-learning components