Search papers, labs, and topics across Lattice.
This paper introduces the Ego-Centric Octree (ECO), a novel spatial data structure designed for mobile robots that process continuous point streams in real-time. By dynamically bounding the mapping space to the robot's immediate surroundings and employing an efficient incremental update algorithm, ECO significantly reduces computational overhead and latency. Evaluations show that ECO achieves up to 67.52% faster updates compared to traditional methods, enhancing real-time spatial perception in dynamic environments.
ECO slashes update times by over 67% while maintaining a balanced tree structure, revolutionizing real-time spatial perception for mobile robots.
Constructing octrees for mobile robots that process continuous point streams in real time poses significant computational and memory challenges. Standard global structures often suffer from high latency and unbalanced tree growth. We introduce the Ego-Centric Octree (ECO), a spatial data structure that acts as a 3D sliding window, dynamically bounding the mapping space to the robot's immediate surroundings. ECO uses an efficient incremental update algorithm that categorizes the environment into shift-out, shift-in, and overlap regions, eliminating redundant global coordinate transformations. Evaluations on the KITTI benchmark demonstrate that ECO reduces update times by up to 25.60% (24.87% on average) compared to full static reconstruction and by up to 67.52% (54.60% on average) compared to a bounded incremental baseline. Furthermore, ECO substantially lowers the total system latency of downstream tasks, running up to 34.17% faster than full reconstruction in voxel-map generation. In dynamic scenes, ECO naturally retains a short-term temporal memory of moving objects, providing useful temporal context while keeping update cost bounded and the tree balanced for real-time spatial perception.
