Mobility is undergoing a profound transformation towards Cooperative, Connected, and Automated Mobility (CCAM). The European BERTHA project, led by the Institute of Biomechanics (IBV) with 14 partners from 6 countries, aims to develop a Driver Behavior Model (DBM) to make autonomous vehicles safer, more human, and predictable, crucial aspects for their social acceptance.
In a CCAM ecosystem, autonomous vehicles interact with other road users like pedestrians, cyclists, and drivers. A vehicle might be technically correct, but if its behavior is perceived as strange, abrupt, or aggressive, human trust will diminish. Predictability is essential for trust and acceptance.
To address this need, BERTHA has developed a Driver Behavior Model (DBM) comprising four modules: Perception, Cognition, Affective-Emotional, and Motor Control. These modules have been validated using data from simulations and real-world tests.
The Perception Module uses machine learning to replicate 360° human attention, interpreting and anticipating driver gaze in complex scenarios, thereby enhancing safety and trust.
The Cognition Module, based on the COSMODRIVE model, simulates the driver's risk assessment and decision-making, determining the safest driving behavior according to perceived risk.
The Affective Module, built on Bayesian Networks, predicts key mental states like fatigue and stress, enabling vehicles to adapt to human variability.
The Motor Control Module generates probability distributions for a range of driver actions, capturing the inherent variability in human driving styles and translating it into nuanced vehicle responses.
Helios De Rosario-Martínez, a researcher at IBV, notes that the launch of these software components is a starting point for CCAM research in human driver behavior.
Rigorous validation of the modules using Field Operational Tests (FOT) ensures their robustness and reliability for integration into automated mobility systems, fostering social and regulatory trust.
BERTHA positions itself as a global leader in human-like automated mobility, addressing the modeling of holistic, personal, and cultural aspects of driving. Its scalable approach sets a new benchmark for digital validation in the CCAM industry.
Andrés Soler Valero, BERTHA project coordinator, states that the DBM module launch marks a turning point, allowing manufacturers to iterate faster and more accurately in developing next-generation Advanced Driver-Assistance Systems (ADAS).
