Organizations: Neuromorphic Automation and Intelligence Lab (NAIL), School of Artificial Intelligence and Robotics, Hunan University, Changsha, China. · State Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, China.
Event cameras are neuromorphic vision sensors with pixels that independently and asynchronously respond to brightness changes at microsecond resolution, supporting state estimation under rapid motion, motion blur, and high dynamic range (HDR) illumination. This also enables capture of high-frequency blinking patterns from light-emitting diodes (LEDs). Based on this, we propose a real-time event-based framework for LED tracking and six-degree-of-freedom (6-DoF) relative pose estimation using pulse-width modulation (PWM)-encoded LED markers. Multiple LEDs share a common blinking frequency and encode identities through distinct PWM duty cycles, reducing frequency-resource requirements and computational cost. Our dual-rate asynchronous tracking framework combines low-rate LED detection with high-rate raw-event updates, maintaining robust LED identity association and accurate pixel-level tracking under rapid motion. LED identities and positions establish reliable 2D-3D correspondences for real-time metric 6-DoF relative pose estimation. Extensive evaluations on a self-collected dataset demonstrate robustness under large motion flow and transient image-plane LED overlap. Furthermore, we develop a coordinated robotic platform comprising a quadrotor and a Mecanum-wheeled mobile robot, demonstrating the proposed framework's potential for closed-loop drone guidance through hovering and target-following experiments under time-varying HDR illumination, with onboard deployment on an NVIDIA Jetson Orin NX. The dataset and implementation will be released to facilitate further research.
Figures & tables
Method
LED
Asynchronous
High-Speed
Real-Time
Identification
Tracking
Validation
( ≥0.7 M events/s)
ACE [ 25 ]
✗
✓
✗
✗
HASTE [ 26 ]
✗
✓
✗
✗
EKLT [ 27 ]
✗
✗
✗
✗
DeepEvT [ 28 ]
✗
✗
-
✓
jAER [ 29 ]
✗
✓
✓
✓
TABLE I: Comparison of Event-Based Tracking Methods
Fig. 0: Overview of the proposed event-based relative pose estimation framework. The system follows a detect-track-estimate architecture, consisting of PWM-encoded LED detection and identification, asynchronous Kalman filter tracking, and 6-DoF pose estimation.
Sequence
Motion Type
LEDs
Main Characteristics
single_led
General motion
1
Large optical flow
rotation_slow
Rotation
4
Slow rotational motion
rotation_normal
Rotation
4
Moderate rotational motion
rotation_fast
Rotation
4
Large optical flow
translation_z
Translation
4
Translation along the optical axis
6dof_normal
6-DoF
4
Moderate 6-DoF motion
TABLE II: Summary of the Self-Collected Indoor Dataset
Fig. 1: LED tracking results for the evaluated methods on part sequences. Each subplot shows the x -coordinate tracking results of the 60% duty-cycle LED.
Sequence
HASTE [ 26 ]
jAER [ 29 ]
AEB-Tracker [ 13 ]
Ours †
Ours
single_led
failed
14.39
15.59
148.81
13.37
rotation_slow
155.15
134.73
86.78
6.58
8.28
rotation_normal
156.69
8.64
59.08
32.72
7.93
rotation_fast
110.75
13.99
13.21
149.59
13.06
translation_z
failed
65.17
32.54
10.30
7.12
6dof_normal
120.01
17.16
6.93
16.29
6.66
TABLE III: Comparison of LED Tracking Results on the Self-Collected Dataset. [ RMSE : px] Best and second-best results are highlighted and underlined , respectively
Fig. 2: Experimental platforms for relative-pose-based drone guidance. (a) The mecanum-wheeled car carrying the PWM-encoded LED marker. (b) The quadrotor equipped with a Prophesee EVK4 event camera and an NVIDIA Jetson Orin NX.
Fig. 3: Drone relative-position control using the estimated relative pose under three marker motion conditions: stationary, linear, and circular motion.
Method
Asynchronous update
Time / event [ns]
HASTE [ 26 ]
Patch update
217.0
jAER [ 29 ]
Rectangular Cluster Tracker update
32.1
AEB-tracker [ 13 ]
8-state EKF update
2139.0
Ours
4-state KF update
15.3
TABLE V: Computational Cost of the Asynchronous Tracking Update for Different Event-Based Trackers
Event-driven Perception for Robotics, Istituto Italiano di Tecnologia, Italy · Humanoid Sensing and Perception, Istituto Italiano di Tecnologia, Italy · University of Genoa, Genoa, Italy
Institute of Visual Computing, Graz University of Technology · MIS laboratory, University of Picardie Jules Verne · ICB laboratory, University of Burgundy Europe