Improved Modelling of the Position-Independent, Fine-Scale Lateral Movement of Vehicles Within Their Lane

The lateral movement of vehicles within their lane has a significant impact on the range of vision of sensors that automated driving functions depend on. The range of vision, in turn, is crucial for the objects visible to the sensors, and thus detectable by downstream algorithms. All together these factors decide on the situational awareness of an automated driving function. With the growing importance of simulations in the validation of those, it is necessary to simulate the lateral movement of vehicles within their lane. For this, submicroscopic behavior models have been developed to be used in simulations such as microscopic traffic simulations and when applying the maneuver-based approach in scenario-based testing. In earlier work, it has been shown that the use of a two-level stochastic model that divides the lateral movement in a fine and a coarse part is a suitable approach to model the lateral movement of vehicles within their lane. Since the introduction of the model, improvements of the Markov model being responsible for the coarse movement have already been suggested. This paper tackles the current limitations of the noise model for the fine movement by introducing enhancements that eliminate manual steps and significantly improve the model's results.