Online Aggregation of Trajectory Predictors

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Tong, Alex, Sharma, Apoorva, Veer, Sushant, Pavone, Marco, Yang, Heng
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866916622381350912
author Tong, Alex
Sharma, Apoorva
Veer, Sushant
Pavone, Marco
Yang, Heng
author_facet Tong, Alex
Sharma, Apoorva
Veer, Sushant
Pavone, Marco
Yang, Heng
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2502_07178
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Online Aggregation of Trajectory Predictors
Tong, Alex
Sharma, Apoorva
Veer, Sushant
Pavone, Marco
Yang, Heng
Robotics
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.
title Online Aggregation of Trajectory Predictors
topic Robotics
url https://arxiv.org/abs/2502.07178