TopoNav: Topological Graphs as a Key Enabler for Advanced Object Navigation

2025-09-01

AI Summary

ObjectNav faces challenges in dynamic, long-horizon tasks due to long-term memory decay and insufficient spatial reasoning. To address this, we propose a spatial memory framework built upon an updatable topological graph: the graph serves as a unified representation integrating real-time visual observations with persistent semantic-structural knowledge, enabling joint modeling of adjacency, connectivity, and scene semantics; a dedicated graph update mechanism supports incremental map construction and goal-directed path re-planning under environmental dynamics. Evaluated on standard ObjectNav benchmarks—including AI2THOR and RoboTHOR—our method achieves state-of-the-art performance, improving task success rate by +8.2% and path efficiency (SPL) by 23.5%, with particularly strong generalization in multi-room and highly dynamic environments. The core innovation lies in elevating the topological graph from a static map to an active spatial cognitive hub capable of memory integration and relational reasoning.

Technology Category

• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning
• Planning, Routing, and Scheduling: Model-Based Reasoning
• Intelligent Robots: Localization, Mapping, and Navigation

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding

Abstract

Object Navigation (ObjectNav) has made great progress with large language models (LLMs), but still faces challenges in memory management, especially in long-horizon tasks and dynamic scenes. To address this, we propose TopoNav, a new framework that leverages topological structures as spatial memory. By building and updating a topological graph that captures scene connections, adjacency, and semantic meaning, TopoNav helps agents accumulate spatial knowledge over time, retrieve key information, and reason effectively toward distant goals. Our experiments show that TopoNav achieves state-of-the-art performance on benchmark ObjectNav datasets, with higher success rates and more efficient paths. It particularly excels in diverse and complex environments, as it connects temporary visual inputs with lasting spatial understanding.

Problem

• Addresses memory challenges in long-horizon object navigation tasks
• Leverages topological graphs for spatial knowledge accumulation and reasoning
• Improves navigation performance in dynamic and complex environments

Innovation

• Topological graphs as spatial memory
• Captures scene connections and semantics
• Enables effective reasoning for navigation