Simulation of weather conditions perturbation for an automotive LIDAR sensor
Résumé
Perception algorithms based on LiDAR sensors are more and more widely used in intelligent transport systems, such as ADAS and autonomous vehicles. As new technological advancements evolve, industrial LiDAR sensors are becoming much more affordable and equipped on modern vehicles. Despite their accuracy for mapping the environment and estimating precise distances, the operation of LiDAR is known to be heavily affected by adverse weather conditions, thus reducing the operational design domain of vehicles equipped with this type of sensor. Tests in simulation of intelligent transport systems have been widely used in the industry throughout the whole cycle of development, to accelerate R&D and reduce time to market costs. It is thus important to improve the virtual models of LiDAR sensors to ensure the reproduction of realistic and adverse weather conditions in simulation test. This is the aim of project CVH (Tool-based chain for the validation and certification of the connected automated vehicle). This paper introduces a methodology of acquiring real data and through processing be able to create and improve the current model of perfect sensors by including the effects of perturbations, whether it be hardware or weather caused.
Origine | Fichiers produits par l'(les) auteur(s) |
---|