9 papers
INTERACT: Interactive Planning for Autonomous Driving via Anchor-Conditioned Prediction and Trust-Region Refinement
Aron Distelzweig, Andreas Look, Faris Janjoš +3
Driving in dense urban traffic is interactive: whether a merge or an unprotected turn succeeds depends on how surrounding agents respond to the ego vehicle. Conventional planners p…
Comfort by Construction: Adaptive, Comfort-Bounded Action Spaces for Learned Driving Policies
Anna Rothenhäusler, Daniel Jost, Raghu Rajan +4
Data-driven driving simulators command accelerations and steering rates from a fixed grid without constraining the realized accelerations and jerks. As a result, reinforcement-lear…
PRISM: Privileged Probabilistic Latent Supervision for End-to-End Autonomous Driving Motion Planning
Volodymyr Havrylov, Faris Janjoš, Andreas Look +2
End-to-end autonomous driving (E2E AD) systems integrate perception, prediction, and planning into a single differentiable architecture. While these models show great promise, thei…
Beyond Self-Play and Scale: A Behavior Benchmark for Generalization in Autonomous Driving
Aron Distelzweig, Faris Janjoš, Andreas Look +7
Recent Autonomous Driving (AD) works such as GigaFlow and PufferDrive have unlocked Reinforcement Learning (RL) at scale as a training strategy for driving policies. Yet such polic…
Uncertainty Matters: Structured Probabilistic Online Mapping for Motion Prediction in Autonomous Driving
Pritom Gogoi, Faris Janjoš, Bin Yang +1
Online map generation and trajectory prediction are critical components of the autonomous driving perception-prediction-planning pipeline. While modern vectorized mapping models ac…
Driving is a Game: Combining Planning and Prediction with Bayesian Iterative Best Response
Aron Distelzweig, Yiwei Wang, Faris Janjoš +5
Autonomous driving planning systems perform nearly perfectly in routine scenarios using lightweight, rule-based methods but still struggle in dense urban traffic, where lane change…
Perfect Prediction or Plenty of Proposals? What Matters Most in Planning for Autonomous Driving
Aron Distelzweig, Faris Janjoš, Oliver Scheel +3
Traditionally, prediction and planning in autonomous driving (AD) have been treated as separate, sequential modules. Recently, there has been a growing shift towards tighter integr…
Stochasticity in Motion: An Information-Theoretic Approach to Trajectory Prediction
Aron Distelzweig, Andreas Look, Eitan Kosman +3
In autonomous driving, accurate motion prediction is crucial for safe and efficient motion planning. To ensure safety, planners require reliable uncertainty estimates of the predic…
Motion Forecasting via Model-Based Risk Minimization
Aron Distelzweig, Eitan Kosman, Andreas Look +3
Forecasting the future trajectories of surrounding agents is crucial for autonomous vehicles to ensure safe, efficient, and comfortable route planning. While model ensembling has i…