Current browse context:
eess.SY
Change to browse by:
References & Citations
Electrical Engineering and Systems Science > Systems and Control
Title: GNN-PMB: A Simple but Effective Online 3D Multi-Object Tracker without Bells and Whistles
(Submitted on 21 Jun 2022 (v1), last revised 8 Feb 2023 (this version, v6))
Abstract: Multi-object tracking (MOT) is among crucial applications in modern advanced driver assistance systems (ADAS) and autonomous driving (AD) systems. The global nearest neighbor (GNN) filter, as the earliest random vector-based Bayesian tracking framework, has been adopted in most of state-of-the-arts trackers in the automotive industry. The development of random finite set (RFS) theory facilitates a mathematically rigorous treatment of the MOT problem, and different variants of RFS-based Bayesian filters have then been proposed. However, their effectiveness in the real ADAS and AD application is still an open problem. In this paper, it is demonstrated that the latest RFS-based Bayesian tracking framework could be superior to typical random vector-based Bayesian tracking framework via a systematic comparative study of both traditional random vector-based Bayesian filters with rule-based heuristic track maintenance and RFS-based Bayesian filters on the nuScenes validation dataset. An RFS-based tracker, namely Poisson multi-Bernoulli filter using the global nearest neighbor (GNN-PMB), is proposed to LiDAR-based MOT tasks. This GNN-PMB tracker is simple to use, and it achieves competitive results on the nuScenes dataset. Specifically, the proposed GNN-PMB tracker outperforms most state-of-the-art LiDAR-only trackers and LiDAR and camera fusion-based trackers, ranking the $3^{rd}$ among all LiDAR-only trackers on nuScenes 3D tracking challenge leader board at the time of submission.
Submission history
From: Liping Bai [view email][v1] Tue, 21 Jun 2022 11:01:49 GMT (4761kb,D)
[v2] Tue, 12 Jul 2022 02:22:10 GMT (5017kb,D)
[v3] Sat, 16 Jul 2022 16:18:23 GMT (5079kb,D)
[v4] Thu, 21 Jul 2022 08:15:48 GMT (5079kb,D)
[v5] Fri, 21 Oct 2022 02:40:52 GMT (5524kb,D)
[v6] Wed, 8 Feb 2023 09:43:51 GMT (5524kb,D)
Link back to: arXiv, form interface, contact.