STEGNav: Spatio-Temporal Event Graph Reasoning for Multimodal Lifelong Object Navigation
Robotics
Summary
The authors address the problem of navigating unknown places to find objects or follow instructions, which needs remembering what’s been seen and deciding where to go next. They note that previous methods had trouble telling similar objects apart and balancing searching for new parts of the environment with aiming for the target. To fix this, they propose STEGNav, a system that keeps track of spatial details and time-based navigation memories in a combined graph, helping the agent choose better next steps without extra training. Their tests show improved success rates and efficiency, supporting that combining space and time information helps in complex navigation tasks.
Authors
Yang Chen, Zhenyu Huang, Wenbo Fu, Danyang Peng, Shi-Yu Tian, Kun-Yang Yu, Lan-Zhe Guo
Abstract
Multimodal lifelong navigation requires an agent to autonomously explore unseen environments while sequentially completing navigation tasks specified by object categories, language descriptions, or reference images. Existing methods primarily accomplish these tasks by constructing state-centric semantic scene graphs. By treating scene graphs as persistent repositories of semantic observations, these methods struggle to distinguish similar instances, jointly represent semantic targets and exploration frontiers, and effectively exploit navigation memory and trajectory experience. To address these limitations, we propose Spatio-Temporal Event Graph Navigation (STEGNav), a training-free framework that extends conventional scene graphs into spatio-temporal event graphs along complementary spatial and temporal axes. The spatial axis performs query-conditioned instance grounding and jointly represents semantic targets and occupancy-aware exploration frontiers characterized by reachability, path cost, and exploration utility. The temporal axis employs trajectory-aware dual-window memory to retain recent decision--trajectory events and verified cross-subtask navigation outcomes. A VLM-based navigation agent reasons over the resulting spatio-temporal event graph and selects either a target instance or an exploration frontier as its next navigation goal. STEGNav achieves 66.3% SR and 39.7 SPL on GOAT-Bench, as well as SR scores of 64.0% and 69.4% on HM3Dv1 and HM3Dv2, respectively. Ablation studies and error analyses validate the complementary effects of the two axes, demonstrating that event-driven spatio-temporal representations improve navigation reliability and cross-subtask experience reuse.