Search papers, labs, and topics across Lattice.
The paper introduces Serialized Red-Green-Gray (SerRGG), an improved method for efficiently updating roadmap graphs in dynamic motion planning scenarios. SerRGG extends the Red-Green-Gray (RGG) framework by using batch serialization and vectorization to accelerate the classification of roadmap edges as valid, invalid, or uncertain using GPU acceleration. Experiments show that SerRGG achieves a 2-9x speedup compared to the sequential RGG implementation, making it suitable for real-time robotic applications.
Updating motion planning roadmaps in dynamic environments just got an order of magnitude faster with a GPU-accelerated edge validation scheme.
Motion planning in dynamic environments, such as robotic warehouses, requires fast adaptation to frequent changes in obstacle poses. Traditional roadmap-based methods struggle in such settings, relying on inefficient reconstruction of a roadmap or expensive collision detection to update the existing roadmap. To address these challenges we introduce the Red-Green-Gray (RGG) framework, a method that builds on SPITE to quickly classify roadmap edges as invalid (red), valid (green), or uncertain (gray) using conservative geometric approximations. Serial RGG provides a high-performance variant leveraging batch serialization and vectorization to enable efficient GPU acceleration. Empirical results demonstrate that while RGG effectively reduces the number of unknown edges requiring full validation, SerRGG achieves a 2-9x speedup compared to the sequential implementation. This combination of geometric precision and computational speed makes SerRGG highly effective for time-critical robotic applications.
