cs.LGSep 3, 2026

A Low-Cost, Open Platform for End-to-End Autonomous Driving on a Miniature Ackermann Vehicle

Authors: Gustavo Claudio Karl Couto, Eric Aislan Antonelo, Gabriel George Zipperer

Abstract

This paper presents a low-cost, open experimental platform for research in end-to-end autonomous driving with miniature Ackermann vehicles. The platform combines a physical vehicle, a printed urban track, data collection tools, trajectory registration, and a Webots digital twin, enabling controlled experiments that connect simulation-based autonomous-driving methods to real-world execution. As a first baseline, we implement command-conditioned behavior cloning, in which a neural policy receives an on-board camera image and a high-level navigation command and outputs steering and speed. The system is evaluated both on the physical vehicle and in simulation. In real closed-loop experiments, the learned policy follows lanes and executes commanded turns, reaching a mean cross-track error of 6.1 cm with respect to the reference route, close to the 4.7 cm observed in human demonstrations. In the digital twin, camera field of view has a strong effect on performance, reducing the mean cross-track error from 35.6 to 3.3 cm when widened from 58 to 120 degrees. Using the digital twin to generate synthetic driving data and a learned sim-to-real image translator to reduce the appearance gap, we further show that a higher-capacity policy trained on this synthetic data combined with real demonstrations is the only configuration that completes all four track routes in closed loop, whereas the compact baseline and the same network trained on real data alone complete fewer. These results establish the open platform as a practical testbed for sim-to-real studies and provide an initial command-conditioned imitation-learning baseline; we release it to support reproducible research.

Explore similar work

CardsList
Apr 27, 2026cs.RO

TEACar: An Open-Source Autonomous Driving Platform

Intelligent Transportation Systems (ITS) increasingly rely on vision-based perception and learning-based control, necessitating experimental platforms that support realistic hardware-in-the-loop validation. Small-scale platforms for autonomous racing offer a practical path to hardware validation, but often suffer from limited modularity, high integration complexity, or restricted extensibility. This paper presents TEACAR, a 1/14- to 1/16-scale autonomous driving platform designed with modular mechanical architecture, hardware abstraction, and ROS 2-based software. The system adopts a four-layer deck structure that physically decouples sensing, computation, actuation, and power subsystems, improving structural rigidity while simplifying reconfiguration. We constructed and comprehensively evaluated the prototype of TEACAR. Its mechanical stability, structural characteristics, and software performance were quantified based on three CNN-based steering controllers. Inference latency, power consumption, and system operating time were measured to evaluate computational capability and robustness. Our experiments demonstrated that TEACAR offers a scalable, modular, and cost-effective testbed for ITS research, education, and development. Our project repository is available on GitHub.
Zhongzheng Zhang, Maxwell Ruyle, Andrew Kappes +5
Autonomous DrivingAutomotive Domain
AskPDF
Jun 5, 2026cs.RO

Test-Time Trajectory Optimization for Autonomous Driving

End-to-end planners for autonomous driving typically generate a set of candidate trajectories, score each one, and return the highest-scoring candidate. However, the scorer is applied only after the proposals are generated and cannot influence the set of trajectories: a weak set of candidates limits planning performance regardless of the scorer's quality. We instead treat the scorer as a learned trajectory-level reward function and search for trajectories that maximize it. Our method, TOAD, runs the Cross-Entropy Method at test time, warm-started from the planner's proposals. It requires no retraining and is plug-and-play for existing planners. Across six base planners, TOAD improves results on NAVSIM-v1 (94.7 PDMS), NAVSIM-v2 (56.3 EPDMS), and the closed-loop HUGSIM benchmark. The code will be made publicly available via the project page: https://valeoai.github.io/TOAD/.
Yihong Xu, Eloi Zablocki, Yuan Yin +4
Autonomous DrivingGradient-Based Trajectory OptimizationTrajectory-Level Credit AssignmentScorerCross-Entropy Losses
AskPDF
Jun 11, 2026cs.LG

Human-like autonomy emerges from self-play and a pinch of human data

Self-play reinforcement learning has recently emerged as a way to train driving policies without any human data. It uses cheap, large-scale simulations to substitute expensive, large-scale human driving demonstrations. A key limitation of this approach is that policies trained through pure self-play can learn effective but alien driving conventions incompatible with people. Previous works attempt to mitigate such behavioral misalignments through extensive reward engineering and domain randomization, which are brittle and labor-intensive. Instead of completely discarding human demonstrations, our method treats them as a regularization objective on top of a minimal safe goal-reaching reward. Like the spice in a good stew, we find that a little human data goes a long way: our method uses only 30 minutes of human demonstrations, 2500x fewer than comparable imitation learning approaches. Resulting policies coordinate with held-out human trajectories and complete training in 15 hours on a single consumer-grade GPU. Videos and full source code are available at https://spiced-self-play.com/.
Daphne Cornelisse, Julian Hunt, Zixu Zhang +4
Self-PlayNaturalistic Driving DataRandomMisalignment
AskPDF