Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

STAGE: Scalable and Traversability-Aware Graph based Exploration Planner for Dynamically Varying Environments

Akash Patel, Mario A V Saucedo, Christoforos Kanellakis, George Nikolakopoulos

TL;DR

The paper tackles autonomous exploration in GPS-denied, dynamically changing subterranean environments. It introduces STAGE, a two-layer graph framework that uses local sub-graphs built from direct point-cloud visibility and a globally stitched graph formed by overlapping sub-graphs to enable scalable exploration with minimal recomputation. A central contribution is the uncertainty-aware re-planning mechanism, which segments global paths and employs an oriented sampling space to verify traversability and update the global graph when unknown changes occur. Validation includes simulation and real-world experiments (e.g., a legged robot with LiDAR) showing higher exploration gain and lower CPU usage compared to a state-of-the-art graph-based planner, and demonstrating robust adaptation to dynamic changes. This work advances long-term autonomous navigation in SubT-like environments by offering a memory-efficient, traversability-aware planning framework that scales to large maps and remains robust to environmental changes.

Abstract

In this article, we propose a novel navigation framework that leverages a two layered graph representation of the environment for efficient large-scale exploration, while it integrates a novel uncertainty awareness scheme to handle dynamic scene changes in previously explored areas. The framework is structured around a novel goal oriented graph representation, that consists of, i) the local sub-graph and ii) the global graph layer respectively. The local sub-graphs encode local volumetric gain locations as frontiers, based on the direct pointcloud visibility, allowing fast graph building and path planning. Additionally, the global graph is build in an efficient way, using node-edge information exchange only on overlapping regions of sequential sub-graphs. Different from the state-of-the-art graph based exploration methods, the proposed approach efficiently re-uses sub-graphs built in previous iterations to construct the global navigation layer. Another merit of the proposed scheme is the ability to handle scene changes (e.g. blocked pathways), adaptively updating the obstructed part of the global graph from traversable to not-traversable. This operation involved oriented sample space of a path segment in the global graph layer, while removing the respective edges from connected nodes of the global graph in cases of obstructions. As such, the exploration behavior is directing the robot to follow another route in the global re-positioning phase through path-way updates in the global graph. Finally, we showcase the performance of the method both in simulation runs as well as deployed in real-world scene involving a legged robot carrying camera and lidar sensor.

STAGE: Scalable and Traversability-Aware Graph based Exploration Planner for Dynamically Varying Environments

TL;DR

The paper tackles autonomous exploration in GPS-denied, dynamically changing subterranean environments. It introduces STAGE, a two-layer graph framework that uses local sub-graphs built from direct point-cloud visibility and a globally stitched graph formed by overlapping sub-graphs to enable scalable exploration with minimal recomputation. A central contribution is the uncertainty-aware re-planning mechanism, which segments global paths and employs an oriented sampling space to verify traversability and update the global graph when unknown changes occur. Validation includes simulation and real-world experiments (e.g., a legged robot with LiDAR) showing higher exploration gain and lower CPU usage compared to a state-of-the-art graph-based planner, and demonstrating robust adaptation to dynamic changes. This work advances long-term autonomous navigation in SubT-like environments by offering a memory-efficient, traversability-aware planning framework that scales to large maps and remains robust to environmental changes.

Abstract

In this article, we propose a novel navigation framework that leverages a two layered graph representation of the environment for efficient large-scale exploration, while it integrates a novel uncertainty awareness scheme to handle dynamic scene changes in previously explored areas. The framework is structured around a novel goal oriented graph representation, that consists of, i) the local sub-graph and ii) the global graph layer respectively. The local sub-graphs encode local volumetric gain locations as frontiers, based on the direct pointcloud visibility, allowing fast graph building and path planning. Additionally, the global graph is build in an efficient way, using node-edge information exchange only on overlapping regions of sequential sub-graphs. Different from the state-of-the-art graph based exploration methods, the proposed approach efficiently re-uses sub-graphs built in previous iterations to construct the global navigation layer. Another merit of the proposed scheme is the ability to handle scene changes (e.g. blocked pathways), adaptively updating the obstructed part of the global graph from traversable to not-traversable. This operation involved oriented sample space of a path segment in the global graph layer, while removing the respective edges from connected nodes of the global graph in cases of obstructions. As such, the exploration behavior is directing the robot to follow another route in the global re-positioning phase through path-way updates in the global graph. Finally, we showcase the performance of the method both in simulation runs as well as deployed in real-world scene involving a legged robot carrying camera and lidar sensor.
Paper Structure (9 sections, 2 equations, 7 figures, 2 algorithms)

This paper contains 9 sections, 2 equations, 7 figures, 2 algorithms.

Table of Contents

  1. Introduction
  2. Related Works
  3. Contributions
  4. STAGE planner architecture
  5. Local Exploration
  6. Global Re-Positioning Planning
  7. Uncertainty Aware Re-Planning
  8. STAGE Evaluation
  9. Conclusion

Figures (7)

  • Figure 1: Overview of the STAGE Local Exploration Planning scheme.
  • Figure 2: Overview of the STAGE Global Re-positioning Planning scheme.
  • Figure 3: STAGE planner overview visualization in A-H. Steps A-E depict the sequential global graph build through the association of the local sub-graphs. Step F visualizes the concept of obstructed path in the global graph during global re-positioning and the removal of connected nodes of the global graph in the untraversable area. Step G visualizes the sub-graph on an unexplored nearby alternative area to reach the re-positioning global frontier. Step H depicts the global path-way segments on the adapted global graph.
  • Figure 4: STAGE planner's adaptive re-planning behaviour in presence of unknown change to the global map. The obstacle in [B] is added when robot was out of line of sight. In [C] re-planned path ways are highlighted in green.
  • Figure 5: Semantic Traversability $\mathbb{ST}$ image. The refined path segment (red arrow) is computed based on the most traversable zone depicted on the $\mathbb{ST}_{image}$, where traversability is represented by a set of semantic labeled classes describing the smoothness of the terrain (e.g. flat, rough, obstacle, etc.).
  • ...and 2 more figures