Fig. 6: Limitations of the model.
From: Human-like driving behaviour emerges from a risk-based driver model
a Tactical costs: The DRF model can only perceive physical risk from objects such as cars, trees, etc. However, it cannot perceive the risk from oncoming traffic which is currently not in its field of view. Hence, at an intersection, rather than slowing down, it will speed up, since there is larger road-area available, which is contrary to what a human would do. This can be solved by introducing additional ‘tactical costs’ that artificially increase the risk of an intersection (red square). This approach can be extended to other elements such as traffic lights or zebra crossings. b Predicted path: For simplicity, the DRF model currently uses a circular arc for predicting the path (for preview time tla seconds). This circular path arises due to the assumption that the current steering angle (δ) and speed (v) will be held constant over the preview time. However, we can optimise for a vector of steering angles and speed (as is done in a Model Predictive Control). This allows for a flexible DRF and better prediction of microscopic trajectories. c Surround DRF: In this paper, the DRF only extends in front of the vehicle (top). However, the risk field extends on all four sides. The bottom image is merely a suggestion, and the shape has not been investigated. This ‘surround DRF’ will help predict human-driving behaviour in additional scenarios such as: being followed by another car, being overtaken, lane change manoeuvres, etc. d Uncertainty in dynamic obstacles: The DRF represents the driver’s (self) perception-action uncertainty. However, the motion of dynamic obstacles is less predictable. This uncertainty was ignored in this paper, but will have to be accounted for in future iterations of this model.
