Filtering Distributed Information to Build a Plausible Scene for Autonomous and Connected Vehicles.

To make their decisions, autonomous vehicles need to build a reliable representation\r\nof their environment. In the presence of sensors that are redundant, but not necessarily\r\nequivalent, that may get unreliable, unavailable or faulty, or that may get attacked, it is of\r\nfundamental importance to assess the plausibility of each information at hand. To this end,\r\nwe propose a model that combines four criteria (relevance, trust, freshness and consistency)\r\nin order to assess the confidence in the value of a feature, and to select the values that are\r\nmost plausible.We show that it enables to handle various difficult situations (attacks, failures,\r\netc.), by maintaining a coherent scene at any time despite possibly major defects.