Online Aggregation of Trajectory Predictors

TL;DR

This paper introduces a lightweight, model-agnostic online aggregation method for trajectory predictors that adaptively combines multiple models to improve prediction accuracy in autonomous driving, even across different environments.

Contribution

It develops an online convex optimization-based framework to dynamically aggregate diverse trajectory predictors, enhancing robustness and performance across varied datasets.

Findings

Outperforms individual predictors on NUSCENES dataset

Maintains high accuracy on out-of-distribution LYFT dataset

Adapts effectively to different city environments

Abstract

Trajectory prediction, the task of forecasting future agent behavior from past data, is central to safe and efficient autonomous driving. A diverse set of methods (e.g., rule-based or learned with different architectures and datasets) have been proposed, yet it is often the case that the performance of these methods is sensitive to the deployment environment (e.g., how well the design rules model the environment, or how accurately the test data match the training data). Building upon the principled theory of online convex optimization but also going beyond convexity and stationarity, we present a lightweight and model-agnostic method to aggregate different trajectory predictors online. We propose treating each individual trajectory predictor as an "expert" and maintaining a probability vector to mix the outputs of different experts. Then, the key technical approach lies in leveraging online data -- the true agent behavior to be revealed at the next timestep -- to form a convex-or-nonconvex, stationary-or-dynamic loss function whose gradient steers the probability vector towards choosing the best mixture of experts. We instantiate this method to aggregate trajectory predictors trained on different cities in the NUSCENES dataset and show that it performs just as well, if not better than, any singular model, even when deployed on the out-of-distribution LYFT dataset.