About this Abstract |
| Meeting |
2027 TMS Annual Meeting & Exhibition
|
| Symposium
|
AI-Enabled Materials Processing: Integrating Accelerated Experimental Workflows and Processing-Aware Machine Learning
|
| Presentation Title |
Low-Cost Open-Source Camera Modules for Continuous Monitoring of Self-Driving Laboratories |
| Author(s) |
Seth David Leavitt, Yanghuang Liu, Jonathan Woo, Sterling Baird |
| On-Site Speaker (Planned) |
Seth David Leavitt |
| Abstract Scope |
As laboratories move toward higher autonomy, continuous equipment monitoring becomes essential for debugging, provenance, and safety — echoing recent calls for aviation-style "black box" recording in self-driving laboratories (SDLs). We present a fully open, low-cost (<$100/module), many-camera, multi-site livestreaming infrastructure for SDL operations. Unlike task-specific computer vision for real-time process control, each equipment module receives a dedicated passive camera with tunable frame rate and resolution and failsafe auto-restart for unattended operation. Dozens of modules deployed across multiple institutions have achieved >99% uptime while collecting tens of thousands of hours of video. Streams linked to timestamps and experiment metadata provide a visual audit trail complementing structured records; the footage also yields heuristic utilization estimates and, prospectively, robot-learning training data. Cameras capture only operators' hands and forearms, a privacy-by-design choice for shared facilities housing furnaces, synthesis platforms, and characterization instruments. All hardware, firmware, and documentation are openly available (https://ac-training-lab.readthedocs.io/en/latest/devices/picam.html). |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Process Technology, Sustainability, |