3D Virtual Reality Testing-Tool for GFHB-Algorithm

This paper describes the product of a project, evaluating the feasibility to develop and validate algorithms and software for the automotive feature Glare-Free-High-Beam (GFHB). Aiming to evaluate the developed algorithms, a fully virtual testing platform was created to test, debug and visualize algorithm behaviour and performance in a 3D scene. GFHB consumes data from a camera-based object-detection system in order to cut off specific high-beam segments, so that the affected traffic participant will not perceive a glaring from the ego-vehicle´s high-beam. One major challenge in developing such algorithms is the availability of testing means and feasibility of testing activities in a controlled working environment, especially because real-world tests need to be performed during darkness / night time. Furthermore, the mentioned algorithms are complex in nature and contain a high amount of process steps, making it quite difficult to break down, isolate and analyse components the of the main software model, while performing a real-world test drive. Therefore, a 3D virtual-reality testing tool has been developed, based on MATLAB/Simulink and the built-in Virtual Reality Toolbox, which enables the developers of GFHB algorithms to rapidly prototype and test several algorithmic approaches on calculating the so-called Glare-Free-Areas. The testing tool is able to use pre-recorded driving scenarios and it visualizes the outputs of the GFHB algorithms, while simulating these, based on the recorded vehicle data. The users of this testing tool are free to decide, whether to visualize only several specific process steps of the control algorithm, or the full end-to-end process chain (from object detection to LED control of the headlamps).