Simulation testing framework for Automated Driving Localization

In nominal condition, Highly Automated Driving System needs to perform four tasks in parallel: localization, perception, planning and control. As a consequence, both Urban and Highway architecture for Highly Automated Driving Functions agree that ‘accurate’ localisation is a key function to enable Automated Driving Systems (ADS) for path planning and decision making [1]. On the one hand, current technological (SW & HW) solution datasheet provide nominal performances that are not always met. On the other hand, no consensus
exists between all automotive actors (OEM, Tier-1 and NMP) on quantitative requirements to be put on localisation function embedded on Automated Driving Systems. Main challenges that simulation is facing are the completeness of the virtual environment and the representativeness of the virtual signals. Through this article, we propose concrete solutions to address both; A co-simulation approach to construct a comprehensive tool chain and a sensor realistic behaviour modelling to increase the level of representativeness of the system features.