An Integrity Assessment Framework for Multi-modal Vehicle Localization

Integrity monitoring of sources and process associated with localization systems used in autonomous navigation applications is of paramount importance. In this work, we present a novel source-level integrity monitoring method with a focus on generalization, scalability and versatility. A cross-consistency assessment based framework is developed to monitor the integrity of different spatial data sources. The proposed method is applied to several datasets, which demonstrates promising behavior of our integrity markers in identifying and quantifying unreliable data from sources. Several scenarios, which are normal occurrences in challenging driving conditions, are considered in the experiments and the advantages of the proposed integrity monitoring process are highlighted.