ChemOS: An orchestration software to democratize autonomous discovery

Area of Science:

• Materials Science
• Artificial Intelligence
• Robotics

Background:

• Traditional materials discovery is slow, taking up to two decades for technologies to reach the market.
• The capital-intensive nature of current research necessitates disruptive strategies for innovation.
• Self-driving laboratories offer a path to revolutionize experimentation through automation and AI.

Purpose of the Study:

• To address the lack of adequate software solutions hindering self-driving laboratory development.
• To propose and develop ChemOS, a software package for deploying and operating self-driving laboratories.
• To demonstrate the versatility and functionality of ChemOS across various applications.

Main Methods:

• Development of ChemOS, a portable, modular, and versatile software package.
• Implementation of structured layers for self-driving laboratory deployment and operation.
• Integration of automated equipment and remote control capabilities.
• Operation at various autonomy levels, including researcher feedback loops.

Main Results:

• ChemOS facilitates seamless integration of automated equipment.
• The software enables remote operation of automated laboratories.
• Demonstrated flexibility and broad functionality through five diverse applications.
• Successful execution on different automated equipment highlighted software capabilities.

Conclusions:

• ChemOS provides essential software infrastructure for advancing self-driving laboratories.
• The software package accelerates the materials discovery pipeline.
• ChemOS supports adaptable autonomy, enhancing researcher interaction and efficiency.