Automated Lab Testing Pipeline - Transforming Materials Research with Robotics & ML

Timeline: 6 months (Summer 2024 - Fall 2024)

Role: Test Automation Intern

Team: PYC Lab

---

🎯 Project Overview

What: Built a fully autonomous compression testing pipeline integrating robotic arm, testing equipment, and computer vision for high-throughput materials research
Why: Eliminate tedious manual testing that required constant researcher attention and accelerate data collection for machine learning workflows
Impact: Achieved 15 samples/hour throughput (2-3x improvement over manual processing) while collecting high-quality data for convolutional neural network training

---

🔧 Technical Implementation

Key Technologies & Tools

• Primary Stack: Python (orchestration), C++ (hardware interface), OpenCV (computer vision), Pandas (data management)
• Hardware Integration: UE5 robotic arm, Instron compression testing machine, Arduino microcontroller, camera for data matrix scanning
• New Skills/Tools: Serial communication protocols, multi-language system integration, lab automation workflows
• Notable Features:
  • Autonomous sample handling and positioning
  • Real-time data collection and formatting for ML pipelines
  • Computer vision-based sample tracking with data matrix scanning
  • Synchronized multi-device communication and timing control
  • Automated data export and organization

Architecture Highlights

Designed a full-stack automation system where Python scripts orchestrate the entire workflow through serial communication with Arduino, which interfaces directly with lab hardware. The system integrates computer vision for sample identification, robotic manipulation for physical handling, and automated data processing—creating a seamless pipeline from sample placement to ML-ready datasets.

---

🚧 Challenges & Problem-Solving

Challenge: Multi-System Integration & Hardware Communication

Solution: Developed a state-machine approach in Python to orchestrate timing between robotic arm, compression testing, and data collection while creating seamless interfaces between Python control logic and C++ hardware drivers. Implemented custom serial communication protocols for bidirectional data exchange and timing synchronization across all system components.
Skills Demonstrated: Systems integration, multi-language programming, hardware-software interface design, cross-disciplinary problem-solving

Challenge: Real-Time Error Handling & System Recovery

Solution: Built robust error detection and recovery mechanisms to handle hardware failures, communication timeouts, and sample positioning errors during autonomous operation. Implemented failsafe procedures and automatic retry logic to maintain system reliability without manual intervention.
Skills Demonstrated: Fault-tolerant system design, debugging complex multi-component systems, reliability engineering

---

📊 Results & Impact

• Throughput Improvement: 15 samples/hour (2-3x faster than manual processing)
• Process Automation: Eliminated need for constant researcher supervision of repetitive testing
• Data Quality: Automated data collection and formatting directly compatible with ML training pipelines
• Research Contribution: System contributed to ongoing lab research paper on materials property prediction

What’s Next

The automated pipeline established a foundation for high-throughput materials testing that could be expanded to other testing protocols. Future enhancements could include multi-sample batch processing and integration with additional characterization equipment.

---

💡 Key Takeaways

Technical Skills Developed:

• Lab automation and robotics integration
• Computer vision with OpenCV for industrial applications
• Serial communication and hardware interface programming
• Multi-language system architecture (Python + C++)

Professional Skills Strengthened:

• Cross-disciplinary collaboration (software + materials science)
• Systems thinking and integration challenges
• Research methodology and data pipeline development
• Problem-solving in hardware-constrained environments

---

🔗 Links

• Research Lab: [Lab Website]