RoA-Planner: Rotatable Area-Based Path Planner in Dense Spaces
Hyouk Ryeol Choi, Yeongwoo Son, Hyunyong Lee, Hansol Kang, Ji Man Park, SeongWon Nam, JaeYoung Oh, Bumsu Yi, Junha Song, SooYeon Choi, Bogeun Kim, daegeun song
AI summary
Problem
Existing path planning methods struggle with conservative approximations, high computational costs, and safety guarantees when navigating asymmetric rectangular robots through dense, obstacle-cluttered spaces.
Approach
The authors define Rotatable Areas as safe heading-angle ranges for collision-free rotation, then use geometric area and edge conditions to evaluate local motions, transforming the SE(2) planning problem into an efficient 2D search via a quadtree and modified A* algorithm.
Key results
- Novel Rotatable Area concept for precise rotational collision checking
- RoA-Planner framework using quadtrees and modified A* for efficient dense-space navigation
- Geometric area and edge conditions that reduce SE(2) planning to a 2D problem
- Real-time, collision-free path generation validated in simulations and physical quadruped experiments
Why it matters
Enables reliable, real-time navigation for rectangular mobile robots in critical applications like warehouse automation and urban patrol where space is tightly constrained.
Abstract
Path planning in obstacle-dense environments is a challenging problem, particularly for robots with asymmetric rectangular footprints. To address this problem, we propose a novel collision-checking approach, called a Rotatable Area, which represents a range of heading angles where the robot can rotate without colliding with obstacles. Based on the relationship between two rotatable areas, we define safe local motion and extend this concept to the RoA-Planner, a path planning frame- work in SE(2) dense space. We validate our planner through extensive simulations and real-world experiments in complex and narrow environments. The results demonstrate that our method achieves fast planning speed while ensuring safety and robustness, making it suitable for practical applications. Note to Practitioners—Navigating complex environments is critical for applications, such as warehouse automation, facility management, or urban patrol. Existing path planning methods either suffer from low collision-checking accuracy in complex environments or require a long computation time to ensure fidelity. To tackle this problem, this paper introduces the Rotat- able Area, a collision-checking method tailored for asymmetric rectangular robots. By checking local motion between rotatable areas, a collision-free path can be planned in a fast and precise manner. This approach is applicable to any shape of rectangular robots with holonomic movement and particularly effective in obstacle dense and narrow environments. The method was validated in scenarios derived from industrial fields and service domains, demonstrating efficiency and safety in real-time.