Predicting real-world effects of highly automated driving
In this work package, Bo, Ignacio and Silvia will develop microscopic traffic flow models that aim to predict safety and efficiency of traffic with different penetration rates of highly automated vehicles.
These models will predict effects of highly automated driving on the traffic-flow characteristics (capacity, fundamental diagram, flow stability, shockwave propagation). The models will incorporate cognitive side effects of automation and human error probabilities identified in the literature and WP1.
Bo will model variability within drivers during highly automated and manual driving. Mental workload and driver distraction have substantial influences on perceptual thresholds and are important contributors to accident risk, and will have a central role in these newly developed microscopic models.
Ignacio will quantify individual differences in driving skill and driving style. The focus herein is on effects of personality, gender, age, and other individual differences that are likely to affect human behaviour during highly automated driving. Ignacio will also study national differences in behaviour including cross-cultural comparisons and differences in road infrastructure.
Silvia will integrate driver models and the quantified variations within and between drivers as delivered by Emin and Ignacio into microscopic traffic flow models. These will include effects of transient manoeuvres such as merging, splitting, platoon entry, platoon exit, and authority transitions between manual and highly automated driving.
The traffic-flow model parameters will be calibrated with empirical data collected by means of driving simulators (WP1–WP3), and with naturalistic driving and FOT data as available by the partners. The effects of our proposed HMI concepts (WP2) on safety and congestion will be predicted for several relevant road networks and future road topologies, with varying penetration levels of automation.