Conditional multimodal anchors improve 3d mapping in unknown environments
All Roads Lead to Rome: Flow-driven Multi-Anchor Exploration for Open-Environment Active 3D Mapping
Robotics
Summary
Mapping unknown places in 3D is hard because robots only see parts of the space at a time and there are many possible ways to explore next. Existing methods try to guess a single next goal but that can lead to poor decisions in new environments. The authors propose a method that suggests multiple possible goals, called anchors, based on the current map state. These multiple options help the robot pick better exploration paths, improving mapping quality and efficiency.
What this means in practice
- •For autonomous robot developers: Enable robots to explore and map unfamiliar indoor and outdoor areas more efficiently by providing multiple candidate navigation goals.
- •For drone navigation engineers: Improve long-range drone path planning for 3D reconstruction in complex environments by using conditional multimodal anchors for goal selection.
Authors
Yang Li, Aming Wu, Zihao Zhang, Ziju Han, Sijia Zhang, Yahong Han
Abstract
To advance the development of embodied intelligence, Open-Environment Active 3D Mapping has attracted increasing attention, aiming to perform a long-horizon and shortest trajectory exploration for reconstructing unseen scenarios. Since only limited information about unseen environments is available, methods built on the closed-set assumption, i.e., assuming that the test environments are similar to those seen during training, cannot generalize satisfactorily. In existing active mapping methods, long-horizon exploration is often guided by predicting a coarse long-range goal and then converting it into an executable path. However, this stage is usually formulated as single-point prediction. Under partial observability, the same local observation may correspond to multiple plausible exploration directions, making such deterministic prediction prone to brittle decisions and degraded performance in unseen scenarios. Our experiments further verify that this is a key factor underlying their weak generalization. To address this issue, we reformulate long-horizon target prediction as conditional multimodal anchor generation using Conditional Flow Matching.Instead of predicting a single goal, our method learns a conditional distribution over coarse exploration anchors from the current mapping state. These anchors are first converted into executable candidate paths through obstacle-aware planning. We then apply exploration-mode clustering to compress geometrically similar trajectories and reduce candidate redundancy. Finally, a hierarchical selection module selects the most promising mode and reranks paths within it to produce the final executable trajectory. Experiments show that our method improves generalization and reconstruction efficiency in open environments.