AV4EV: Open-Source Modular Autonomous Electric Vehicle Platform for Making Mobility Research Accessible

TL;DR

AV4EV presents an open-source, one-third-scale autonomous electric go-kart platform designed to democratize mobility research by coupling a modular mechatronic stack with flexible sensing and a ROS2-based software framework. The system supports manual, remote, and autonomous modes and employs GNSS-RTK EKF localization, raceline optimization via MINCURVE, and adaptive pure pursuit along with boundary-detection–driven reactive control (Follow-the-Gap). Key contributions include a detailed five-subsystem mechatronics design, a versatile sensing suite, and a dual-track software approach that combines pre-mapped racing with reactive execution, all packaged with comprehensive open-source resources. The platform achieved competitive success (2023 Purdue Grand Prix) and is intended to accelerate cost-effective algorithmic development and verification in AV and EV communities, bridging the gap between full-scale vehicles and reduced-scale prototypes. Overall, AV4EV offers a practical, cost-effective, and extensible reference for universities and labs to advance autonomous mobility research through shared design, software, and tutorials.

Abstract

When academic researchers develop and validate autonomous driving algorithms, there is a challenge in balancing high-performance capabilities with the cost and complexity of the vehicle platform. Much of today's research on autonomous vehicles (AV) is limited to experimentation on expensive commercial vehicles that require large skilled teams to retrofit the vehicles and test them in dedicated facilities. On the other hand, 1/10th-1/16th scaled-down vehicle platforms are more affordable but have limited similitude in performance and drivability. To address this issue, we present the design of a one-third-scale autonomous electric go-kart platform with open-source mechatronics design along with fully functional autonomous driving software. The platform's multi-modal driving system is capable of manual, autonomous, and teleoperation driving modes. It also features a flexible sensing suite for the algorithm deployment across perception, localization, planning, and control. This development serves as a bridge between full-scale vehicles and reduced-scale cars while accelerating cost-effective algorithmic advancements. Our experimental results demonstrate the AV4EV platform's capabilities and ease of use for developing new AV algorithms. All materials are available at AV4EV.org to stimulate collaborative efforts within the AV and electric vehicle (EV) communities.

Figures (5)

• Figure 1: Go-kart platform overview with Steer-by-Wire System (SBWS) including its hand wheel (HW) and road wheel (RW) components, Throttle-by-Wire System (TBWS), and Electronic Braking System (EBS). The sensors and computing units mounted on the double-deck rear shelf are enumerated from top to bottom as follows: (1) Ouster LiDAR, (2) OAK-D camera, (3) Onboard laptop, (4) Main Control System (MCS), (5) Sepentrio GNSS, and (6) IMU, concealed from the main view perspective, is positioned on the lower deck.
• Figure 2: Sensing (left) and motor (right) power system with connections and devices.
• Figure 3: Software pipeline for go-kart autonomous driving capabilities: GNSS-based adaptive pure pursuit (red), camera-based follow-the-gap (green), go-kart mechatronics execution (blue).
• Figure 4: Waypoints collection and raceline optimization at Purdue Grand Prix racing track, which spans a distance of 434 meters.
• Figure 5: Grass boundary detection. (a) Raw camera input. (b) Filtered grass mask. (c) The BEV of the grass mask. Green lines indicate the angles for searching grass distances. (d) The converted depth data is plotted as green dots and overlaid onto the BEV image of the camera input.