Fig. 4: Validating the model in road scenarios using literature.

Each row represents one scenario and the columns compare two different metrics in that scenario. The DRF model results are compared to the results from literature (Supplementary Notes 1–8) in the adjacent subfigures (Supplementary Figures 3–6). In the DRF model subfigures, the black and grey markers represent the sport and normal parameter settings, respectively. 1 Curve radius: 1a and 1b show that the DRF model predicts the decrease in ‘curve-cutting’ (quantified using TTR) as curve radius increases. 1c and 1d show the speed at the curve centres. The sport setting of DRF cuts the curves more (1a) and drives at higher speeds (1c) compared to the normal setting. 2 Lane width: 2b shows that the (mean ± SE) standard deviation of lateral position (SDLP) of the vehicle increases as the lane width increases. The DRF model (2a) can predict this trend. 2c and 2d (mean ± SD) show that the speed at which drivers negotiate a road increases as the lane width increases. 3 On-road obstacles: In 3b, the ‘wide’ obstacle encroaches more onto the road compared to the ‘narrow’ obstacle. The minimum lateral deviation (3b) is calculated from the trajectories in 3a. Drivers moved away from the parked cars (3c: lane centre = 0, bars indicate 95% CIs). 3b shows that the DRF model showed a similar trend of moving away from the obstacle. Drivers drove slower when there were parked cars, as compared to when there were no parked cars encroaching the road (3f: bars indicate 95% CIs). 3e shows that the DRF model slows down for obstacles covering the road partially. 4 Roadside furniture: In the asymmetric case, the mean lateral deviation from the lane centre is away from the parked cars (4b) and away from water (more dangerous than grass) in 4c. Subfigure 4c shows the distribution of lateral position of the participants. 4e and 4f show that in the symmetric condition with ‘danger’ on both sides of the lane, the DRF model correctly predicted that the drivers drove slower than in the asymmetric case. The mean lateral deviation (4b) and mean speed (4e) are calculated from the trajectories in 4a and 4d, respectively.