Research Analyzer
← Back ICRA 2026

FineNav: A Versatile Framework Enhancing Ground Robot Navigation in Unstructured Environment

Jinghui Wang, Chenyang Wang, Yuxuan Cao, Zelong Sun, Wang Xi, Jianping He

PDF

AI summary

Key figure (auto-extracted from paper)
FINENAV enables reliable ground robot navigation in complex 3D environments by decoupling real-time local perception from scalable global mapping through a novel hierarchical system.
ground robot navigation unstructured environments hierarchical mapping voxel grid octree open-source framework

Problem

Existing frameworks fail to simultaneously handle dynamic obstacles, non-planar terrain, and multi-story structures due to the trade-off between real-time computational efficiency and large-scale memory demands.

Approach

The authors introduce FINENAV, a full-stack framework that couples a high-rate local voxel grid for rapid obstacle and terrain analysis with a memory-efficient global octree for persistent, large-scale mapping, all wrapped in a modular, Nav2-compatible architecture.

Key results

  • Novel hierarchical mapping system balancing real-time latency and memory efficiency
  • Validated full-stack navigation across dynamic, uneven, and multi-story real-world scenarios
  • Modular architecture with seamless Navigation2 (Nav2) ecosystem compatibility
  • Open-source release of the complete navigation stack

Why it matters

It provides a practical, deployable solution for autonomous ground robots in complex real-world settings, lowering the barrier for next-generation mobile robotics development.

Abstract

Autonomous navigation of ground robots in un- structured 3D environments remains a fundamental challenge, as it requires accommodating dynamic obstacles, non-planar ground, and multi-story structures within a unified framework. In this paper, we propose a versatile navigation framework named FINENAV. It features a novel hierarchical mapping system that couples a high-rate local voxel grid for real- time perception with a scalable global octree for persistent storage. This design balances low-latency performance with large-scale mapping capabilities, enabling reliable navigation in unstructured environments. Moreover, the entire navigation pipeline is refactored into modular and reusable components, while maintaining compatibility with existing 2D navigation ecosystems. We validate FINENAV on a wheeled robot, demon- strating its versatility across diverse scenarios. FINENAV is released as open-source software for the community.

Index terms

Software Architecture for Robotic and Automation Software Middleware and Programming Environments

Related papers