Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper addresses the challenge of localizing robots in untraversed environments by leveraging rich semantic data from OpenStreetMap. The authors introduce a method that utilizes Vision-Language Models (VLMs) to extract relevant landmarks from egocentric observations and establish correspondences with a large semantic map, overcoming the limitations of existing approaches. Key results show that their lightweight matcher, distilled from the VLM, efficiently computes correspondences across extensive mapped landmarks, enabling robust pose estimation even under varying environmental conditions.
Robots can now localize in unfamiliar environments using rich semantic data, achieving robust performance across diverse conditions without prior traversal.
We want robots to localize in previously untraversed environments against commonly available prior data. Rich semantic data available from OpenStreetMap can be useful in this task. However, existing methods either ignore this semantic information, directly matching panoramas and overhead imagery, or dramatically compress the semantic information, working with a small set of fixed classes. To leverage this rich semantic information, two challenges need to be overcome. First, useful semantic information needs to be extracted from the robot's egocentric observations. Second, the observed information must be quickly associated with the large prior semantic map (e.g., up to 628 km^2). We show that VLMs are effective at both extracting relevant landmarks from panoramas, and identifying feasible correspondences between these landmarks and prior overhead landmarks. However, using VLMs to propose all correspondences scales poorly as the number of mapped landmarks increases. Instead, we propose distilling a lightweight matcher from a VLM which computes correspondences for all entities in a map. We use this output to form an observation likelihood which is fused over time with a Bayes filter to create a time series of pose estimates. To support further investigation into generalizable cross-view methods that leverage semantic information, we release a dataset of extracted semantics and evaluation trajectories spanning eleven environments, including panoramas we collected in a snowstorm and at night in Boston. We demonstrate our method, trained on a single city's fair-weather data, generalizes across location, lighting, weather, and other challenges. Code and datasets are available at https://efahnestock.github.io/loci/.