Search papers, labs, and topics across Lattice.
This paper introduces a LiDAR-only dynamic object detection framework that combines temporal occupancy grid-based motion segmentation with a learned bird's-eye-view (BEV) dynamic prior to improve robustness in cluttered indoor environments. The framework uses a fusion module to prioritize 3D detections when available and leverages the learned dynamic grid to recover detections missed due to proximity-induced false negatives. Experimental results demonstrate a 28.67% increase in recall and an 18.50% increase in F1 score compared to state-of-the-art methods in cluttered environments, while maintaining comparable precision and position error.
LiDAR-only dynamic object detection gets a major boost in cluttered environments by fusing temporal occupancy grids with a learned BEV dynamic prior, significantly reducing false negatives.
Reliable dynamic object detection in cluttered environments remains a critical challenge for autonomous navigation. Purely geometric LiDAR pipelines that rely on clustering and heuristic filtering can miss dynamic obstacles when they move in close proximity to static structure or are only partially observed. Vision-augmented approaches can provide additional semantic cues, but are often limited by closed-set detectors and camera field-of-view constraints, reducing robustness to novel obstacles and out-of-frustum events. In this work, we present a LiDAR-only framework that fuses temporal occupancy-grid-based motion segmentation with a learned bird's-eye-view (BEV) dynamic prior. A fusion module prioritizes 3D detections when available, while using the learned dynamic grid to recover detections that would otherwise be lost due to proximity-induced false negatives. Experiments with motion-capture ground truth show our method achieves 28.67% higher recall and 18.50% higher F1 score than the state-of-the-art in substantially cluttered environments while maintaining comparable precision and position error.
