Autonomous Learning Across Heterogenous, Low-Cost, Modular Self-Driving Labs

Autonomous research laboratories (also known as self-driving labs) accelerate the scientific process - letting scientists fail smarter, learn faster, and spend less resources in their studies. The user first defines their goals and machine learning then controls automated research equipment to perform experiment design, execution, and analysis in a closed loop, homing in on the user goals. A common challenge that impacts both traditional and autonomous materials research is the choice between measurement techniques (or experiments) that are high speed but low fidelity (e.g., surrogate experiments) and those that are high fidelity but low speed. Similarly, different autonomous systems may integrate measurement techniques of different speed and fidelity. We demonstrate an on-the-fly autonomous research campaign across multiple autonomous research systems. Demonstration is performed across a set of Legolas systems – a low-cost modular, self-driving lab. A pH-meter equipped Legolas works hand-in-hand with a camera-equipped Legolas to map color dyed acid and base mixtures to pH and color, accelerating the discovery of the Henderson-Hasselbach relationship as well as the relationship between color and pH.