Publications

Published Papers

1. 

   PB-MOT: Pose-aware Association Boosted Online 3D Multi-Object Tracking Bo Pang, Yang Xu, Jiming Chen, and Liang Li 2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2025 [Abs] [Paper] [Project Page] [Video Demo] [Benchmark]

   Robotic and autonomous driving platforms necessitate efficient 3D Multi-Object Tracking (MOT) that harmonizes geometric precision, motion robustness, and computational efficiency. Traditional 3D MOT approaches face critical challenges: geometric similarity metrics (e.g., IoU-based) degrade at long ranges with high computational costs, while distance-based methods fail to capture object orientation and shape; the effects of occlusion and the intricate relative ego-object motion degrade tracking performance in dynamic scenes. To this end, we propose PB-MOT, an online framework integrating two key innovations: ego-motion-compensated state estimation that decouples dynamic interactions; and a rotated ellipse association algorithm unifying pose and shape-aware matching with adaptive distance constraints. Evaluations on the KITTI benchmark show that our PB-MOT achieves state-of-the-art performance with a HOTA score of 81.94%, while running at an impressive 2,402.76 FPS on CPU. This enables real-time, high-fidelity perception and tracking for resource-constrained robotic systems.

Manuscripts

1. 

   PB-MOT++: Extending Pose-aware Association to Trajectory-Centric Offline Optimization Bo Pang, Yang Xu, Jiming Chen, and Liang Li Under revision after first-round review at IEEE Transactions on Intelligent Transportation Systems (T-ITS). [Abs] [Project Page] [Video Demo] [Benchmark]

   Reliable perception of dynamic traffic participants is essential for intelligent transportation systems and autonomous driving. While recent advances in 3D multi-object tracking (3D MOT) have substantially improved benchmark performance, trajectory fragmentation and identity inconsistency remain challenging under prolonged occlusion or sparse observations. Existing offline methods partially alleviate these issues through global matching or sequence-level refinement, yet many remain detection-centric and offer limited capability for active trajectory reconstruction. To address this problem, we propose PB-MOT++, a computationally efficient trajectory-centric offline optimization framework that extends our previous pose-aware tracker, PB-MOT. Instead of repeatedly operating on frame-wise detections, PB-MOT++ directly optimizes trajectories through an evidence-first four-stage pipeline: boundary recovery, topology repair, gap completion, and kinematic refinement. By separating structural restoration from geometric smoothing, the proposed framework effectively reduces error propagation and improves long-range trajectory consistency. Extensive experiments on the KITTI benchmark demonstrate the effectiveness of PB-MOT++. Our method achieves state-of-the-art performance with 83.16% HOTA and 87.23% AssA, while maintaining 2351.08 FPS on a CPU-only platform. These results show that high-fidelity offline trajectory optimization can substantially improve tracking quality with minimal computational overhead.