5 papers
CANMOT: Class-Aware Noise Modeling for Multi-Object Tracking in Autonomous Driving
Timo Osterburg, Stefan Schütte, Torsten Bertram
Kalman filter (KF)-based multi-object tracking (MOT) remains a strong baseline for autonomous driving due to its strong performance, computational efficiency and interpretability.…
Learned Non-Maximum Suppression for 3D Object Detection
Timo Osterburg, Stefan Schütte, Torsten Bertram
Post-processing is a critical stage in LiDAR-based 3D object detection, where dense and overlapping proposals must be filtered for compact and reliable perception. This work introd…
HiLO: High-Level Object Fusion for Autonomous Driving using Transformers
Timo Osterburg, Franz Albers, Christopher Diehl +2
The fusion of sensor data is essential for a robust perception of the environment in autonomous driving. Learning-based fusion approaches mainly use feature-level fusion to achieve…
LoRD: Adapting Differentiable Driving Policies to Distribution Shifts
Christopher Diehl, Peter Karkus, Sushant Veer +2
Distribution shifts between operational domains can severely affect the performance of learned models in self-driving vehicles (SDVs). While this is a well-established problem, pri…
Energy-based Potential Games for Joint Motion Forecasting and Control
Christopher Diehl, Tobias Klosek, Martin Krüger +3
This work uses game theory as a mathematical framework to address interaction modeling in multi-agent motion forecasting and control. Despite its interpretability, applying game th…