Robotics and Sensor-Based Localization

Simultaneous Localization and Mapping, or SLAM, is the problem of building a consistent map of an unknown environment while simultaneously figuring out where within that map a robot or vehicle is located — a chicken-and-egg challenge that lies at the heart of autonomous navigation. Researchers approach it using sensors ranging from monocular cameras and RGB-D depth cameras to lidar, extracting geometric and visual features that are fused into representations like point clouds or graph-based pose estimates, often in real time. In aerospace contexts, where GPS is unreliable or absent entirely — think planetary rovers, inspection drones in confined structures, or spacecraft approaching an uncharted body — robust localization becomes safety-critical rather than merely convenient. Open questions center on how to maintain accuracy as maps grow large and as lighting or surface conditions change, and how to make these systems resilient enough to handle the sensor noise and compute constraints of actual flight hardware.

Works

94,232

Total citations

1,404,400

Keywords

SLAM3D MappingVisual OdometryRoboticsLocalizationPoint Cloud