Recording, Replay, Validation: A PCIe-Based Approach to AI-Enabled ADAS

The speech of Thomas Schöpfner explores how PCIe-based architectures enable high-bandwidth vehicle and sensor data acquisition – from a few megabytes to several gigabytes per second – and how the same measurement hardware can subsequently be used for data replay and Hardware-in-the-Loop (HIL) testing.

A key focus is the role of this data in AI-assisted ADAS development. Recorded sensor and vehicle data can be used for training-data preparation, model evaluation and systematic error analysis. Repeatable replay makes it possible to compare different AI model versions under identical conditions, investigate challenging driving scenarios and analyse how changes to a model influence the behaviour of the vehicle controller.

By combining recording, targeted data injection, replay, automated testing and HIL integration, VCARSYSTEM aims to create a continuous feedback process from real-world observations to AI model improvement and system-level validation.

📅 October 13, 2026 | 1:20–1:40 pm
📍 Holiday Inn Munich-Westpark, Munich
🎯 Session Block: Autonomous Mobility Technology

Want to hear directly from Thomas Schöpfner and other experts shaping autonomous and AI-driven mobility?

#eMove360 #ADAS #AutonomousDriving #ArtificialIntelligence #AutomotiveAI #HardwareInTheLoop #HIL #AutomotiveTesting #PCIe #VehicleDevelopment #VCARSYSTEM

[DISPLAY_ULTIMATE_SOCIAL_ICONS]