GOFusion: Real-Time Gaussian-Occupancy Submap Fusion for Fisher-Informed Autonomous Robot Navigation
In this letter, we present GOFusion, a real-time visual mapping system for autonomous robot navigation that anchors Gaussian primitives to local occupancy maps, providing both spatial awareness for path planning and high-fidelity appearance that enables information-driven robot navigation.
ProofPaper ↗
Key points
- GOFusion builds 3D Gaussian maps from pose estimates and local occupancy maps obtained via visual-inertial SLAM and enables pose corrections to maintain both geometric and visual consistency.
- Building on the appearance-rich Gaussian map, we introduce a rasterizer-integrated Fisher approximation that quantifies per-primitive sensitivity online, serving as an information-driven heuristic for Gaussian pruning and next-best-view planning.
- We evaluate GOFusion on public datasets and in real-world drone flights.
- Finally, we demonstrate fully autonomous drone exploration for Gaussian map refinement using the proposed Fisher-informed heuristic.
Sources (1)
- [1]GOFusion: Real-Time Gaussian-Occupancy Submap Fusion for Fisher-Informed Autonomous Robot NavigationarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 18, 01:31 PM
In this letter, we present GOFusion, a real-time visual mapping system for autonomous robot navigation that anchors Gaussian primitives to local occupancy maps, providing both spatial awareness for path planning and high-fidelity appearance that enables information-driven robot navigation.
GOFusion builds 3D Gaussian maps from pose estimates and local occupancy maps obtained via visual-inertial SLAM and enables pose corrections to maintain both geometric and visual consistency.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 18, 2026Google Open-Sources Harness2, a Recursive Agent Harnessing Framework
- Oct 18, 2026Localization and Neural Splatting on the Edge: An FPGA Scene Memory for Robots
- Oct 18, 2026Elevatormusic/DeepSeek-V4.1-Flash-1017-EXL3-TensorFold-2x-DGX-Spark: DeepSeek-V4.1-Flash-1017 EXL3 2.9bpw on 2x NVIDIA DGX Spark (GB10) with TensorFold: 2.1-2.3x the vLLM kit's decode speed.
- Oct 16, 2026[Research] 2,788 tokens/s on a single RTX 5090 with a diffusion LLM — Introducing ORL (fully open source)
- Oct 16, 2026Learning Robot Metareasoning: Adaptive Computation Budgets for Real-Time Control
- Oct 15, 2026Granite Switch 4.1: Chat, Decisions, and Speech from One Endpoint