Constantin Selzer
PhD Student · Intelligent Vehicles Lab · Munich University of Applied Science
I develop unified deep learning frameworks that jointly predict traffic participant behavior and plan safe trajectories for autonomous vehicles. My research bridges the gap between trajectory prediction and motion planning in dense urban traffic, using transformer architectures and aerial imagery.
Research Interests
Trajectory Prediction
Forecasting multi-agent behavior in complex urban environments with multimodal distributions
Motion Planning
Goal-conditioned planning under uncertainty with dynamic feasibility constraints
Unified Prediction & Planning
Closing the gap between prediction and planning in autonomous driving pipelines
Aerial Imagery
Leveraging drone-based datasets for comprehensive urban traffic understanding
News
June 2026 Paper accepted at IEEE IV 2026: PlanTRansformer — Unified Prediction and Planning with Goal-conditioned Transformer
September 2024 Presented DeepUrban at IEEE ITSC 2024 in Edmonton, Canada
2024 Participated in Tacheles Conference 2024 on Generative Models for Autonomous Driving
February 2023 Started PhD at the Intelligent Vehicles Lab, Munich University of Applied Science