Search papers, labs, and topics across Lattice.
This study addresses the limitations of traditional autonomous navigation methods in unstructured terrains, which often lead to short-sighted decision-making due to restricted depth sensing. By utilizing satellite imagery to extract long-range traversability-aware frontiers from first-person-view images, the authors create a dataset that enhances the navigation capabilities of robots without relying heavily on human demonstrations. The results show over a 10% improvement in long-range off-road navigation performance compared to existing techniques, alongside a reduction in required human interventions during real-world trials.
Extracting long-range navigational insights from satellite imagery significantly boosts off-road navigation performance, reducing reliance on human input.
Standard methods for autonomous navigation in unstructured terrain are prone to myopic behaviors in long-horizon scenarios. The use of metric maps built from LiDAR or cameras provides necessary local geometry and semantic information but is strictly limited by depth sensing range. By discarding data beyond the mapping horizon robots suffer from suboptimal, short-sighted decisions. To recover this lost information, we focus on extracting long-range traversability-aware frontiers directly from first-person-view (FPV) images. By leveraging satellite imagery, we compute the set of feasible navigation paths for a dataset of image/pose pairs and use them to supervise our network, reducing the need for extensive human demonstration data. We demonstrate that this approach improves performance in long-range off-road navigation over existing methods by more than 10% in various offline benchmarks and reduces the number of human interventions incurred in a set of real-world experiments. More details can be found at https://theairlab.org/ss_frontiers_iros .
