ROS 2-Based LiDAR Perception Framework for Mobile Robots in Dynamic Production Environments, Utilizing Synthetic Data Generation, Transformation-Equivariant 3D Detection and Multi-Object Tracking

Lukas Bergs; Tan Chung; Marmik Thakkar; Alexander Moriz; Amon Göppert; Chinnawut Nantabut; Robert Schmitt

cs.ROApr 2, 2026

ROS 2-Based LiDAR Perception Framework for Mobile Robots in Dynamic Production Environments, Utilizing Synthetic Data Generation, Transformation-Equivariant 3D Detection and Multi-Object Tracking

Authors: Lukas Bergs, Tan Chung, Marmik Thakkar, Alexander Moriz, Amon Göppert, Chinnawut Nantabut, Robert Schmitt

Organizations: Chair of Intelligence in Quality Sensing (IQS), Laboratory for Machine Tools and Production Engineering (WZL), RWTH Aachen University, Campus-Boulevard 30, Aachen 52074, Germany · The Sirindhorn International Thai-German Graduate School of Engineering (TGGS), King Mongkut’s University of Technology North Bangkok (KMUTNB), 1518 Pracharat 1 Road, Wongsawang, Bangsue, Bangkok 10800, Thailand · Fraunhofer Institute for Production Technology (IPT), Steinbachstr. 17, Aachen 52074, Germany

Abstract

Adaptive robots in dynamic production environments require robust perception capabilities, including 6D pose estimation and multi-object tracking. To address limitations in real-world data dependency, noise robustness, and spatiotemporal consistency, a LiDAR framework based on the Robot Operating System integrating a synthetic-data-trained Transformation-Equivariant 3D Detection with multi-object-tracking leveraging center poses is proposed. Validated across 72 scenarios with motion capture technology, overall results yield an Intersection over Union of 62.6% for standalone pose estimation, rising to 83.12% with multi-object-tracking integration. Our LiDAR-based framework achieves 91.12% of Higher Order Tracking Accuracy, advancing robustness and versatility of LiDAR-based perception systems for industrial mobile manipulators.

Figures & tables

Figure 1: Framework overview. Icons by Freepik from www.flaticon.com.

Figure 2: Synthetic data generation visualization.

Figure 4: Tracklets visualization in RViz.

Figure 5: Experimental environment at WZL-IQS.

Input Factors	Level 1	Level 2	Level 3
Number of Objects	1	2
Robot Linear Movement	Stationary	0.25 m/s	0.5 m/s
Robot Angular Movement	Stationary	0.25 rad/s	0.5 rad/s
Occlusion	None	$<20\%$	$>40\%$
Initial Distance	2.5 m	3.5 m	4.5 m
Robot Linear Movement with Objects Movements (for mobile objects only, 6 levels)	Stationary - NL - NA	Stationary - PL - NA	Stationary - NL - PA

Table 1: Levels of input factors (PL: with object linear movement, PA: with object angular movement, NL: without object linear movement, NA: without object angular movement).

Number	Robot Linear	Robot Angular	Occlusion	Initial Distance
1–19	Stationary - NL - NA	Stationary	No	2.5 m
2–20	Stationary - NL - NA	0.25 rad/s	$<20\%$	3.5 m
3–21	Stationary - NL - NA	0.5 rad/s	$>40\%$	4.5 m
4–22	Stationary - PL - NA	Stationary	No	3.5 m
5–23	Stationary - PL - NA	0.25 rad/s	$<20\%$	4.5 m
6–24	Stationary - PL - NA	0.5 rad/s	$>40\%$	2.5 m

Table 2: OA matrix for mobile workstation and mobile storage units (Single mobile workstation: 1 - 18; Single mobile storage unit: 1 - 18; Two mobile storage units: 19 - 36).

		IoU	Pos	Rot	DetA	HOTA
MW	D	40.03 %	1.78 $m$	68.99 ∘	49.38 %	-
MW	T	79.08 %	0.06 $m$	23.11 ∘	76.80 %	87.31 %
SW	D	72.20 %	0.71 $m$	102.49 ∘	66.80 %	-
SW	T	82.23 %	0.05 $m$	1.23 ∘	85.06 %	92.11 %
MSU	D	75.79 %	1.14 $m$	29.52 ∘	89.57 %	-
MSU	T	88.04 %	0.04 $m$	8.63 ∘	90.98 %	93.92 %

Table 3: Overall experiment results. MW - mobile workstation; SW – station-ary workstation; MSU - mobile storage unit; Ave – Average of three classes, D - Detection; T - Tracklet. IoU (Average IoU); Pos (Directional RMSE) Rot (Yaw RMSE). Full experimental details are provided in the [ 31 ] .

Figure 6: Bounding box overlap of SW experiment.

Figure 7: Directional and rotation errors in SW experiment.

Lidar-Based 3D Object Detection

Lidar-based 3D object detection locates and classifies objects in point clouds, supporting perception for autonomous vehicles, robots, and other mobile systems.

See all 54 papers

RRTrack: Robust and Recoverable Object 6D Pose Tracking for Dynamic Scenes
Jul 26, 2026Junyue Li, Ye Zheng, Yifan Chen +2Visual Object Tracking Object Pose Estimation
College of Computer Science and Technology, Zhejiang University, Hangzhou, China · Institute of Artificial Intelligence (TeleAI), China Telecom, China · College of Future Information Technology, Fudan University, Shanghai 200433, China
EllipseLIO: Adaptive LiDAR Inertial Odometry with an Ellipsoid Representation
May 20, 2026Rowan Border, Margarita ChliInertial Odometry Mobile Robotics
Vision for Robotics Lab (V4RL), University of Cyprus, Cyprus · ETH Zurich, Switzerland
Factorized Spatio-Temporal Convolutions for Human Pose Estimation from Planar Lidar
Jul 23, 2026Simone Arreghini, Mirko Nava, Nicholas Carlotti +2Relative Pose Estimation Human Pose Estimation
Dalle Molle Institute for Artificial Intelligence (IDSIA), USI-SUPSI, Lugano, Switzerland

Jul 26, 2026J/K move · Enter open · S save

RRTrack: Robust and Recoverable Object 6D Pose Tracking for Dynamic Scenes

Junyue Li, Ye Zheng, Yifan Chen, Zhe Sun+1

College of Computer Science and Technology, Zhejiang University, Hangzhou, China · Institute of Artificial Intelligence (TeleAI), China Telecom, China · College of Future Information Technology, Fudan University, Shanghai 200433, China

Visual Object Tracking Object Pose Estimation

Open paper AskPDF arXiv

Robust object 6D pose tracking is critical for robotic systems operating in dynamic and occluded scenes. Per-frame estimators are accurate but computationally expensive, while current trackers struggle with fast motion and complete occlusion due to their reliance on continuous visibility. To address these challenges, we present RRTrack, an efficient, recoverable object 6D pose tracker that enables robust tracking through fast motion and target disappearance--reappearance. RRTrack introduces a 2D--6D closed-loop tracking strategy that integrates memory-based video object segmentation (VOS) with 6D pose refinement. The 2D branch maintains target localization, and the 6D branch verifies geometric consistency before memory updates. In addition, a DINOv2-based dual-bank template matching module is developed to recover lost targets by jointly exploiting offline synthetic templates and online observation anchors while maintaining real-time efficiency. We also introduce a synthetic RGB-D benchmark comprising three robotic scenarios with fast motion and full occlusion. Experimental results on the synthetic benchmark demonstrate that RRTrack improves equal-subset mean ADD-S AR by 66.3% and ADD-S AUC by 65.7% over FoundationPose while achieving 55.2 FPS. Real-world experiments further validate the robustness of RRTrack under noisy sensing conditions. Project page: https://github.com/7kevin24/RRTrack

Abstract

Lidar-Based 3D Object Detection

Explore similar work

RRTrack: Robust and Recoverable Object 6D Pose Tracking for Dynamic Scenes

EllipseLIO: Adaptive LiDAR Inertial Odometry with an Ellipsoid Representation

Factorized Spatio-Temporal Convolutions for Human Pose Estimation from Planar Lidar