Towards Metric-Agnostic Trajectory Forecasting
2026-07-01 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionRobotics
AI summaryⓘ
The authors studied how autonomous cars predict the movements of other vehicles and found that current methods try to train models specifically for certain evaluation rules, which can conflict. They suggest a new way to train models that doesn't focus on any single metric but produces a full range of possible future paths with probabilities. Then, they use special policies to pick the best predictions for different evaluation rules. They tested this idea by changing a top model called DONUT to use their method, and it performed very well on a popular self-driving dataset.
Trajectory forecastingAutonomous drivingProbabilistic modelingPredictive distributionArgoverse 2Waymo Open Motion DatasetTrajectory Distribution Evaluation (TraDiE)DONUTMotion prediction benchmark
Authors
Markus Knoche, Daan de Geus, Bastian Leibe
Abstract
Accurate trajectory forecasting of surrounding traffic participants is a core capability for autonomous driving, enabling vehicles to anticipate behavior and plan safe maneuvers. We observe that current state-of-the-art forecasting models on Argoverse 2 and the Waymo Open Motion Dataset tailor their training objectives to the different benchmark metrics. Because these metrics encourage conflicting behavior, we propose a paradigm change for trajectory forecasting: training models with metric-agnostic probabilistic objectives and treating metric optimization as a downstream task applied to the predictive distribution. Concretely, we introduce Trajectory Distribution Evaluation (TraDiE) policies, metric-specific policies that map a predictive distribution to the set of $K$ trajectories and confidences required by trajectory forecasting metrics. We evaluate this framework by introducing DONUT-NLL, which adapts the training objective of the state-of-the-art trajectory forecasting model DONUT to directly optimize the predictive distribution. Using our policies, DONUT-NLL achieves state-of-the-art results on all metrics of the Waymo motion prediction benchmark.