Learning Visual Semantic Map-Matching for Loosely Multi-Sensor Fusion Localization of Autonomous Vehicles | IEEE Journals & Magazine | IEEE Xplore