Online Mapping

Online mapping focuses on creating dynamic, up-to-date maps from sensor data, primarily addressing the limitations and high costs of manually creating and maintaining traditional high-definition maps. Current research emphasizes improving the speed and accuracy of online map generation using techniques like bird's-eye-view representations and Gaussian process regression, often integrating these maps directly into downstream tasks such as autonomous vehicle navigation and trajectory prediction. This work is crucial for advancing applications in robotics, autonomous driving, and remote sensing, particularly by enabling operation in unmapped environments and improving the efficiency of data processing.