The front-end acts as the interface between raw sensor data and the system's abstract model of the world. Its primary goal is to process high-frequency sensor information in real-time. SLAM-Former: Putting SLAM into One Transformer - arXiv

Simultaneous Localization and Mapping () is a foundational technology in robotics that allows a device to build a map of an unknown environment while simultaneously tracking its own position within it. This process is often described as a "chicken-and-egg" problem: to build an accurate map, the robot needs to know its location, but to determine its location, it needs a map.

Modern SLAM systems are generally organized into a modular and back-end structure. 1. The Front-End (Data Interpretation)