VRU Integration in Simulations

Abstract

Various simulation methods, such as human-in-theloop or model-in-the-loop, have developed and are being used in more and more applications, as they offer several advantages such as lower costs or less risk. In the automotive sector, these methods of simulation are used to simulate traffic, including Vulnerable Road Users (VRU). Modelling VRU behavior in simulation comes with several challenges, e.g. due to the high degrees of freedom and the spontaneous decision-making process of humans compared to other traffic participants. A general goal of simulation is to model reality as realistic as possible. To achieve a realistic simulation, the components and actors must be calibrated. The calibration process requires real world data to match and align the simulation data with real world data. The collection of this data is feasible with different methods. In this project, two methods were used to capture macroscopic behaviors, such as movement decisions of humans. This work focuses on the VRU actions and decisions rather than the exact microscopic 3d-modelling of the VRU, like dynamics or kinematics.