From Zero to High-Speed Racing: An Autonomous Racing Stack

2025-12-07

AI Summary

High-speed autonomous racing faces significant challenges in precise localization, real-time perception, dynamic trajectory planning, and robust control—further constrained by limited track access and high hardware costs. To address these, we propose the modular Autonomous Racing System (ARS), developed over three iterative generations. We introduce the first publicly available multi-sensor dataset specifically designed for high-speed racing scenarios, encompassing both oval and road-course environments. The system integrates multi-source localization, dynamic object perception, model predictive control (MPC), and robust state estimation to enable end-to-end closed-loop autonomous driving. On real-world tracks, it achieves stable autonomous operation at speeds up to 260 km/h. Comprehensive cross-track performance evaluation demonstrates the method’s effectiveness and generalization capability in highly dynamic environments. Key contributions include: (1) a scalable, modular racing architecture; (2) the first high-speed multi-sensor racing dataset; and (3) empirical validation of robust, high-velocity autonomous control under realistic track conditions.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Computer Vision: Vision for Robotics & Autonomous Driving
• Multiagent Systems: Adversarial Agents

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

High-speed, head-to-head autonomous racing presents substantial technical and logistical challenges, including precise localization, rapid perception, dynamic planning, and real-time control-compounded by limited track access and costly hardware. This paper introduces the Autonomous Race Stack (ARS), developed by the IU Luddy Autonomous Racing team for the Indy Autonomous Challenge (IAC). We present three iterations of our ARS, each validated on different tracks and achieving speeds up to 260 km/h. Our contributions include: (i) the modular architecture and evolution of the ARS across ARS1, ARS2, and ARS3; (ii) a detailed performance evaluation that contrasts control, perception, and estimation across oval and road-course environments; and (iii) the release of a high-speed, multi-sensor dataset collected from oval and road-course tracks. Our findings highlight the unique challenges and insights from real-world high-speed full-scale autonomous racing.

Problem

• Develops an autonomous racing system for high-speed competition
• Addresses challenges in localization, perception, planning, and control
• Evaluates performance across different track types and conditions

Innovation

• Modular autonomous racing stack for high-speed competition
• Three iterations validated on tracks up to 260 km/h
• Released multi-sensor dataset from oval and road courses