Adaptive Control in Autonomous Driving via Real-Time Recurrent RL

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Lemmel, Julian, Resch, Felix, Farsang, Mónika, Hasani, Ramin, Rus, Daniela, Grosu, Radu
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910227483328512
author Lemmel, Julian
Resch, Felix
Farsang, Mónika
Hasani, Ramin
Rus, Daniela
Grosu, Radu
author_facet Lemmel, Julian
Resch, Felix
Farsang, Mónika
Hasani, Ramin
Rus, Daniela
Grosu, Radu
contents We study online fine-tuning of pretrained control policies for autonomous driving using Real-Time Recurrent Reinforcement Learning (RTRRL), a memory-efficient algorithm that updates policy parameters at every time step without backpropagation through time. We extend RTRRL to support LrcSSM, a recently proposed nonlinear diagonal state-space model, and combine offline behavioral cloning with online RTRRL fine-tuning to adapt policies to distribution shifts at deployment. We validate the approach in the CarRacing simulation and on a 1:10-scale RoboRacer platform equipped with an event camera, where a pretrained policy is fine-tuned online during real-world line-following. To our knowledge, this is the first demonstration of online RL fine-tuning with event-camera observations on standard (non-spiking) hardware in closed-loop control. LrcSSM-based policies improve fastest and most consistently across both settings.
format Preprint
id arxiv_https___arxiv_org_abs_2602_02236
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Adaptive Control in Autonomous Driving via Real-Time Recurrent RL
Lemmel, Julian
Resch, Felix
Farsang, Mónika
Hasani, Ramin
Rus, Daniela
Grosu, Radu
Robotics
Machine Learning
Neural and Evolutionary Computing
Systems and Control
We study online fine-tuning of pretrained control policies for autonomous driving using Real-Time Recurrent Reinforcement Learning (RTRRL), a memory-efficient algorithm that updates policy parameters at every time step without backpropagation through time. We extend RTRRL to support LrcSSM, a recently proposed nonlinear diagonal state-space model, and combine offline behavioral cloning with online RTRRL fine-tuning to adapt policies to distribution shifts at deployment. We validate the approach in the CarRacing simulation and on a 1:10-scale RoboRacer platform equipped with an event camera, where a pretrained policy is fine-tuned online during real-world line-following. To our knowledge, this is the first demonstration of online RL fine-tuning with event-camera observations on standard (non-spiking) hardware in closed-loop control. LrcSSM-based policies improve fastest and most consistently across both settings.
title Adaptive Control in Autonomous Driving via Real-Time Recurrent RL
topic Robotics
Machine Learning
Neural and Evolutionary Computing
Systems and Control
url https://arxiv.org/abs/2602.02236