Autonomous Driving for All: The 4th Autoware Open Source Tutorial

Duration: Half-day (3.5 Hours)
Format: Hybrid (Lecture + Hands-on Workshop)
Date: 22 June 2026 Detroit, MI, United States

Organizers

• Dr. Simon Thompson | TIER IV, Inc. (Tokyo, Japan), simon.thompson@tier4.jp
• Dr. Johannes Betz | Technical University of Munich (Munich, Germany), johannes.betz@tum.de
• Prof. Rahul Mangharam | University of Pennsylvania (Philadelphia, USA), rahulm@upenn.edu
  

Abstract

As autonomous driving (AD) research matures, the barrier to entry remains high due to the complexity of integrating perception, planning, and control stacks. This tutorial lowers that barrier by introducing Autoware, the world’s leading open-source autonomous driving software platform.

Designed for researchers, students, and industry engineers, this session provides a comprehensive entry point into the Autoware ecosystem (based on ROS 2). Attendees will move beyond theory to practice through a guided, hands-on workshop covering the installation, configuration, and execution of the Autoware stack in simulation. Furthermore, the tutorial will bridge the gap between AI development and vehicle deployment by demonstrating an open MLOps pipeline for training and deploying models within the stack. By the end of the session, participants will have a functional AD environment on their laptops and the knowledge to extend the platform for their specific research needs.

Keywords: Open-source Autonomous Driving, Autoware, ROS 2, Education, Simulation, MLOps, Vehicle Navigation.

Motivation and Learning Objectives

Research in autonomous vehicles often requires reinventing the wheel—building basic infrastructure before novel algorithms can be tested. Autoware provides a standardized, modular foundation that accelerates research.

Key Learning Outcomes:

1. Modern Architecture for Autonomous Vehicles: Understand the core nodes of Autoware (Perception, Localization, Planning, Control) and how they interact via ROS 2.
2. Autonomous Vehicle Deployment: Gain proficiency in setting up the development environment (Docker/Source) and launching simulations.
3. MLOps for Autonomous Vehicles: Learn the end-to-end workflow for training perception models on public data and deploying them to the vehicle stack.
   

Target Audience

• Graduate Students & Academic Researchers looking for a validated platform to test novel algorithms without building a car from scratch.
• Industry Engineers seeking to understand open-source standards in AD.
• Prerequisites: Basic familiarity with Linux (Ubuntu) and ROS/ROS 2 concepts is recommended.
• Requirements: Participants should bring a laptop (NVIDIA GPU recommended, Ubuntu 22.04+) to participate in the hands-on installation.
  

Proposed Agenda (3.5 Hours)

Technical Committee Support

Correspondent TC: IEEE ITSS Technical Committee on Self-Driving Automobiles (SDA-TC).

---