GRID-FAST: A Grid-based Intersection Detection for Fast Semantic Topometric Mapping

TL;DR

GRID-FAST addresses scalable navigation in large environments by deriving semantic topometric maps from 2D grids through a novel intersection-detection framework. It integrates gap-based gap segmentation, wall/object filtering, opening detection, overlap resolution, and intersection optimization to produce a compact skeleton that supports fast planning and decision-making. Validation across indoor, cave-like, and outdoor maps, with comparisons to Voronoi-based baselines, shows GRID-FAST achieves up to 92% fewer nodes and competitive runtimes, while enhancing frontier detection and path safety. The approach offers a practical, online-capable tool for robust robotic navigation and multi-robot coordination, with potential to integrate further semantic labels and higher-level planning strategies.

Abstract

This article introduces a novel approach to constructing a topometric map that allows for efficient navigation and decision-making in mobile robotics applications. The method generates the topometric map from a 2D grid-based map. The topometric map segments areas of the input map into different structural-semantic classes: intersections, pathways, dead ends, and pathways leading to unexplored areas. This method is grounded in a new technique for intersection detection that identifies the area and the openings of intersections in a semantically meaningful way. The framework introduces two levels of pre-filtering with minimal computational cost to eliminate small openings and objects from the map which are unimportant in the context of high-level map segmentation and decision making. The topological map generated by GRID-FAST enables fast navigation in large-scale environments, and the structural semantics can aid in mission planning, autonomous exploration, and human-to-robot cooperation. The efficacy of the proposed method is demonstrated through validation on real maps gathered from robotic experiments: 1) a structured indoor environment, 2) an unstructured cave-like subterranean environment, and 3) a large-scale outdoor environment, which comprises pathways, buildings, and scattered objects. Additionally, the proposed framework has been compared with state-of-the-art topological mapping solutions and is able to produce a topometric and topological map with up to \blue92% fewer nodes than the next best solution.

Figures (15)

• Figure 1: An illustration of the steps taken by the proposed method to create a topometric map. The process takes an occupancy map as input, identifies all intersections, and creates a topometric and topological map. The steps are shown on a section of the map for illustration purposes, but the method is applied to the complete map.
• Figure 2: An example of how the gaps $g_{(i-1)k_1}$, $g_{ik_2}$, $g_{(i+1)k_3}$, and $g_{(i+1)k_4}$ could be connected. Here, $G^+_{ik_2}$ has two elements and is treated as a gap detection.
• Figure 3: A perfect cross intersection aligned with the first scanning angle. The figures show the result of using different numbers of scan directions. The gap detection groups are indicated by blue lines and the arrows indicate the directions the image was scanned in.
• Figure 4: (a) Unfiltered occupancy map where some of the gaps belonging to $\mathcal{G}_{nt}$ are marked. (b) Filtered occupancy map where small openings have been removed. (c) Map showing a gap part of the set $\mathcal{G}_t$ that is connected to an object. The points on the wall $W_n$ around the object have fewer cells with a neighboring occupied point than the threshold $f_{obj}$, and therefore the object is removed. (d) Final filtered map where small objects have been removed.
• Figure 5: A part of a map with three gap detections is shown. The direction of the gap detection and the inside points of the gaps are marked by the arrows in the figure. In this situation, there is an opening between the inside point of $g_3$ and $g_4$. There will not be an opening between the inside points of $g_1$ and $g_2$, as they do not fulfill the second condition P2 in sub-section \ref{['sec:opDetect']} and thus will be converted directly to opening detections. $g_5$ will be converted to an opening detection as it does not overlap any other gap detections.

Theorems & Definitions (5)

• Definition 1
• Definition 2
• Definition 3
• Remark 1
• Definition 4