Luddy researchers partner with Astemo Americas to improve AI perception for autonomous vehicles

Indiana University's Luddy School is partnering with Astemo Americas to use generative AI for training self-driving vehicle vision systems. The one-year project targets edge cases-rare weather, unusual roads-that real-world datasets rarely capture.

Categorized in: AI News Science and Research
Published on: May 15, 2026
Luddy researchers partner with Astemo Americas to improve AI perception for autonomous vehicles

Luddy researchers partner with Astemo on automotive perception AI

Computer science professors at Indiana University's Luddy School are working with automotive supplier Astemo Americas to develop generative AI methods for training vehicle vision systems. The one-year funded project aims to improve how self-driving and advanced driver assistance systems recognize their surroundings in rare or challenging scenarios.

Byung-Cheol Min, principal investigator and professor of computer science and intelligent systems engineering, said the research focuses on synthetic data generation. Real-world driving datasets rarely capture edge cases-unexpected weather, unusual road conditions, or uncommon obstacles-that autonomous systems must handle safely.

"More robust perception can contribute to safer intelligent driving technologies, especially in situations that are rare, unexpected, or difficult to model using conventional data collection alone," Min said.

Addressing data gaps with AI

The project combines Min's expertise in robotics and AI with Jiangpeng He's work in computer vision and deep learning. He serves as co-principal investigator and assistant professor of computer science.

The team will test whether synthetic and augmented visual data can train perception models more effectively than conventional methods alone. They'll also develop evaluation techniques to measure whether augmented data actually improves real-world performance.

He said the collaboration addresses a fundamental problem: "Data scarcity and imbalance in automotive perception." Generative AI can create training examples for scenarios that are difficult or expensive to capture in practice.

Research, not product development

Min emphasized the work is exploratory. The goal is to generate foundational methods and practical insights, not to announce a specific product or timeline.

"This project is about improving how we train and evaluate perception systems so they can operate more reliably in complex, real-world conditions," he said.

The research supports student researchers and expands Luddy's portfolio of applied AI work. Min said the partnership with Astemo Americas connects academic research to real automotive challenges while advancing fundamental computer vision and machine learning techniques.

Learn more about generative AI and LLM applications or explore AI research courses for professionals working in this space.


Get Daily AI News

Your membership also unlocks:

700+ AI Courses
700+ Certifications
Personalized AI Learning Plan
6500+ AI Tools (no Ads)
Daily AI News by job industry (no Ads)