HPHS: Hierarchical Planning based on Hybrid Frontier Sampling for Unknown Environments Exploration

TL;DR

HPHS tackles unknown-environment exploration by fusing fast LiDAR-based frontier sampling with a hierarchical planning framework. It reduces planning complexity through subregion segmentation and uses a DTW-informed revenue model to sequence subregions, while a multi-criterion frontier selector balances travel cost, orientation, and information gain. Empirical results in simulation and real-world tests show substantial improvements in exploration speed, path length, and coverage compared with state-of-the-art baselines. The combined approach enables real-time, scalable exploration with robust performance in varied environments, and code release promises reproducibility and community uptake.

Abstract

Rapid sampling from the environment to acquire available frontier points and timely incorporating them into subsequent planning to reduce fragmented regions are critical to improve the efficiency of autonomous exploration. We propose HPHS, a fast and effective method for the autonomous exploration of unknown environments. In this work, we efficiently sample frontier points directly from the LiDAR data and the local map around the robot, while exploiting a hierarchical planning strategy to provide the robot with a global perspective. The hierarchical planning framework divides the updated environment into multiple subregions and arranges the order of access to them by considering the overall revenue of the global path. The combination of the hybrid frontier sampling method and hierarchical planning strategy reduces the complexity of the planning problem and mitigates the issue of region remnants during the exploration process. Detailed simulation and real-world experiments demonstrate the effectiveness and efficiency of our approach in various aspects. The source code will be released to benefit the further research.

Figures (10)

• Figure 1: A robot performs autonomous exploration in an unknown environment using the proposed method. The yellow points represent the frontier points, the purple point is the next target point, and the black curve is the trajectory of the robot. The green grids are divided subregions.
• Figure 2: The overall framework diagram of the exploration system.
• Figure 3: Frontier points sampling directly from the LiDAR data. The red dots represent the point cloud. The purple and green dots represent the inserted frontier points when condition 1 or 2 is satisfied respectively. The radius and polar angle of the point in the polar coordinate system are expressed by $r$ and $\theta$.
• Figure 4: Subregion segmentation and selection during the exploration. The green grids represent the remaining subregions after being filtered, and the dark blue grid in each map refers to the subregion that should be visited at the current moment. The other unfilled grids represent subregions that are filtered out, as there are no frontiers in the interior. As the size of the map changes, subregions should grow dynamically.
• Figure 5: The illustration of calculating the heuristic gain. The total gain of a frontier point is composed of the traveling gain, the orientation gain, and the information gain. The information gain uses a $k\times k$ information kernel for the comprehensive evaluation.