Fig. 5: Validating the model in traffic scenarios using literature.

Similar to Fig. 4, each row represents one scenario and the two metrics in the two columns compare the DRF model results to trends shown in the literature (Supplementary Notes 9–14). For the DRF model figures, the black and the grey markers represent the sport and normal parameter settings, respectively (Supplementary Figs. 7–9). 1 Car following: 1b and 1c indicate that the preferred time headway is independent of the speed. In 1c, the circular markers indicate the median and the whiskers indicate 25th and 75th percentile. 1e and 1f show that the braking intensity (represented by the acceleration at brake initiation) increases as the approach speed to the obstacle increases. 2 Overtaking: 2b and 2c show that the DRF model could correctly predict that the overtake-distance increases as the speed of the overtaken car increases. In the sport setting, the model covers larger distance than in normal setting, indicating ‘smoother’ trajectories in the sport setting. However, the DRF model does not come back to its own lane sufficiently (2a). Subfigures 2e and 2f show that the predictions of the DRF model agree with the results in literature that show the time to collision (TTC) at the start of the overtake manoeuvre increases, as the speed of the overtaken car increases. 3 Oncoming traffic: In 3b and 3c, the minimum lateral deviation is shown on the y-axis. The condition where no oncoming cars were present is indicated by ‘absent’. The DRF model simulated one car that drove on the oncoming lane’s centre (‘centre’ in 3b) and another car that was offset towards the ego lane (‘offset’ in 3b). In normal and sport setting the DRF model moved away from the oncoming traffic, which is in agreement with the driver’s behaviour. 3e and 3f show that the DRF model slowed down, like humans (3f), when it encountered oncoming traffic. In 3c and 3f, the black markers indicate mean, and whiskers indicate the  ±SD.