A Low-Cost, Open Platform for End-to-End Autonomous Driving on a Miniature Ackermann Vehicle
2026-09-03 • Machine Learning
Machine LearningArtificial IntelligenceRobotics
AI summaryⓘ
The authors created a low-cost, small-scale experimental setup to study self-driving car technology using tiny cars on a printed urban track. Their system includes a real car, a digital simulation twin, and tools to collect and compare data. They tested a simple AI model that drives based on camera images and commands, showing it can follow lanes and turns accurately both in real life and simulation. They found that a wider camera view improves driving performance in simulation, and training with both real and simulated data helps the AI complete more routes successfully. The authors provide this platform openly to help researchers test how well simulated training translates to real-world driving.
autonomous drivingAckermann vehiclebehavior cloningcommand-conditioned controlsim-to-real transferdigital twincross-track errorimitation learningneural policycamera field of view
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.