aalto1 untyped-item.component.html
Robust tracking of heterogeneous objects in sparse point clouds
Loading...
URL
Journal Title
Journal ISSN
Volume Title
School of Engineering |
Master's thesis
Unless otherwise stated, all rights belong to the author. You may download, display and print this publication for Your own personal use. Commercial use is prohibited.
Authors
Date
Department
Major/Subject
Mcode
Language
en
Pages
41
Series
Abstract
Object detection, tracking, and classification are fundamental tasks for smart urban environments and autonomous driving systems. Covering large outdoor spaces with LiDAR sensors typically requires either many sensors with overlapping fields of view or fewer sensors with non-overlapping fields of view. The latter approach is more cost-effective but introduces a handover problem: when a tracked object exits one sensor's field of view and then enters another field of view, maintaining its identity is crucial.
This thesis proposes a method for tackling such scenarios by using a constant velocity motion model with a Kalman filter operating in prediction-only mode: updating the state estimate forward in time without measurement corrections until the object re-enters a field of view of one of the sensors. The method is implemented as a ROS2 node and evaluated on real-world data collected specifically for the research. The results are discussed and improvements for robustness are proposed.