AI and robots are the new battery development dream team

Nov 03, 2023

gmast3r/iStock

By subscribing, you agree to our Terms of Use and Policies You may unsubscribe at any time.

In a world with an ever-increasing demand for electrically-powered devices, portable power supplies have become essential. From electric vehicles to Internet of Things devices, it could be argued that the modern world could function without batteries.

But, sadly, energy densities, charging times, and lifespans of the most common batteries are limited. To this end, many researchers are working hard to find the "Holy Grail" of battery technology — a cheap, long-lasting, sustainable, and power-dense solution. This is no easy feat, but some hope that employing robots and artificial intelligence could supercharge, to borrow a phrase, the entire process.

But first, let's see why developing new batteries is so challenging.

Battery development is a complex field with several challenges to overcome. At present, the main issues tend to include, but are not limited to:

Innovations in materials science, engineering, and manufacturing are continually being pursued to address these issues. To this end, it is hoped that robotics and AI can significantly aid in accelerating these efforts and overcoming the challenges faced in battery development.

IURII BUKHTA/iStock

To get an appreciation of the potential impact AI and robotics could have on the battery industry, we already have a test case from Carnegie Mellon University. In September of 2022, a team of researchers combined cutting-edge AI and robotics technologies to drastically speed up the research process for creating new forms of better batteries.

Since efficient ion conductivity is crucial for designing lithium-ion batteries, researchers are constantly searching for ways to optimize it. The Carnegie Mellon team may have found a novel way to speed this up dramatically.

In their paper, published in Nature Communications, the researchers highlighted successfully pairing a specific type of robot with an AI learning system to develop advanced, non-aqueous liquid electrolytes.

To expedite this process, the researchers created a robot named "Clio" that utilized the necessary ingredients to craft electrolyte samples based on instructions.

A computer with a deep learning AI program named "Dragonfly" was incorporated into the system alongside "Clio" and electrolyte sensors.

"Dragonfly" analyzed the data and proposed suggestions for improvement. "Clio" then implemented these suggestions to create a new sample. The process was repeated over two days until the electrolyte quality improved. The mechanical duo was halted once the researchers deemed the products were ready for testing.

During the experiment, the scientists observed that their paired AI system functioned as intended, gradually enhancing the electrolyte samples.

The most exceptional outcome yielded an improvement of 13% compared to the current top-performing batteries on the market. The researchers now intend to enhance further their system's capabilities to assess various objectives and increase speed.

This research and development area has already expanded outside of academia, with companies like Chemix using a similar approach to develop advanced batteries for the EV market. Whether through university research or commercial development, it is hoped that the discovery of new and innovative materials can be accelerated through automation and machine learning. This could lead to improved batteries, more efficient photovoltaic systems, and other advancements.

And that is your lot for today.

While new battery development is complex and intensive, robotics and artificial intelligence might expedite this process to produce safe and reliable future batteries. The work of researchers at Carnegie Mellon, among others, has already shown some interesting results, and it will be exciting to see what this embryonic field will come up with in the future.