Abstract
In this paper, we analyze the performance of sensors to find optimal navigation solutions for fully autonomous vehicles (AVs). We consider sensory and environmental uncertainties that may prevent AVs from achieving optimal navigation accuracy, which could in turn threaten the safety of road users. As the fusion of all sensors might not be able to resolve these problems, we propose a new approach called PROBE (preparing for roads in advance of barriers and errors) that consists of five algorithms: error analyzer, path finder, challenge detector, challenge analyzer, and path marker. The objective of these algorithms is to provide the highest level of navigation accuracy obtainable in navigation-challenging conditions. The final outcome of these algorithms is visualization of challenging conditions on a selected route. We experimented with PROBE using routes in different countries and the results show that PROBE performs well in detecting, analyzing, and marking navigation challenges in advance. PROBE's outcome can be used to decide on the appropriate sensors in advance of challenging conditions.