On Learning the Tail Quantiles of Driving Behavior Distributions via Quantile Regression and Flows

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Tee, Jia Yu, De Candido, Oliver, Utschick, Wolfgang, Geiger, Philipp
Natura: Preprint
Pubblicazione: 2023
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866909363318292480
author Tee, Jia Yu
De Candido, Oliver
Utschick, Wolfgang
Geiger, Philipp
author_facet Tee, Jia Yu
De Candido, Oliver
Utschick, Wolfgang
Geiger, Philipp
contents Towards safe autonomous driving (AD), we consider the problem of learning models that accurately capture the diversity and tail quantiles of human driver behavior probability distributions, in interaction with an AD vehicle. Such models, which predict drivers' continuous actions from their states, are particularly relevant for closing the gap between AD agent simulations and reality. To this end, we adapt two flexible quantile learning frameworks for this setting that avoid strong distributional assumptions: (1) quantile regression (based on the titled absolute loss), and (2) autoregressive quantile flows (a version of normalizing flows). Training happens in a behavior cloning-fashion. We use the highD dataset consisting of driver trajectories on several highways. We evaluate our approach in a one-step acceleration prediction task, and in multi-step driver simulation rollouts. We report quantitative results using the tilted absolute loss as metric, give qualitative examples showing that realistic extremal behavior can be learned, and discuss the main insights.
format Preprint
id arxiv_https___arxiv_org_abs_2305_13106
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle On Learning the Tail Quantiles of Driving Behavior Distributions via Quantile Regression and Flows
Tee, Jia Yu
De Candido, Oliver
Utschick, Wolfgang
Geiger, Philipp
Machine Learning
Robotics
Towards safe autonomous driving (AD), we consider the problem of learning models that accurately capture the diversity and tail quantiles of human driver behavior probability distributions, in interaction with an AD vehicle. Such models, which predict drivers' continuous actions from their states, are particularly relevant for closing the gap between AD agent simulations and reality. To this end, we adapt two flexible quantile learning frameworks for this setting that avoid strong distributional assumptions: (1) quantile regression (based on the titled absolute loss), and (2) autoregressive quantile flows (a version of normalizing flows). Training happens in a behavior cloning-fashion. We use the highD dataset consisting of driver trajectories on several highways. We evaluate our approach in a one-step acceleration prediction task, and in multi-step driver simulation rollouts. We report quantitative results using the tilted absolute loss as metric, give qualitative examples showing that realistic extremal behavior can be learned, and discuss the main insights.
title On Learning the Tail Quantiles of Driving Behavior Distributions via Quantile Regression and Flows
topic Machine Learning
Robotics
url https://arxiv.org/abs/2305.13106