Photo provided by Zhen Liu A miniature geosystem set up to test the combination of AI and pumping systems, part of the early stages of research into the idea of intelligent infrastructure.

HOUGHTON — Technology is quickly changing the world. Soon, self-driving cars and smartphones could be joined by smart infrastructure, if one Michigan Tech professor has anything to say about it.

Zhen Liu, an environmental engineering professor with a focus on geotechnical engineering is currently exploring, and trying to get research funding for, the idea of intelligent infrastructure.

At the moment, infrastructure solutions are focused on the idea of resilient infrastructure, able to withstand more pressure and recover more quickly. Methods include compacting soil better for roads, installing more drainage systems or improving soil properties.

Liu feels the lifecycle of that research is nearing its end.

“I believe we are heading into a dead alley. The materials, they have physical limits,” Liu explained.

Liu’s looking to change this approach and implement artificial intelligence (AI) into infrastructure. The systems would learn and adapt to conditions like storms and disasters using AI systems rather than the traditional unchanging systems.

Liu’s background in geotechnical engineering is a new angle for AI, he said. Currently he is testing the idea, focusing on pumping systems and is using a miniature geosystem, pumps and a microprocessor to test the combination of AI and pumping systems and develop the appropriate algorithm.

The concept of intelligent infrastructure could apply to many areas but Liu is starting by focusing on what he knows. 

“I want to prove it in geosystems, the area I’m familiar with first, then extend the concept to other areas of infrastructure,” he explained.

Potentially it would be able to respond to scenarios like flooding.

“We’re trying to put some AI programs in it so we can let the pumping system to learn how to work,” Liu said.

The miniature geosystem is just the first step in research as the concept is just in the beginning stages.

“We are still in the stage of proof of concept,” Liu said.

Liu has submitted a proposal to the National Science Foundation (NSF) in hopes they will help provide funding for further research, which would speed up the development.

Liu is still quite early in the process of exploring intelligent infrastructure as a solution but feels the potential is there. Chances are, development will take years, with implementation five, ten or twenty years down the road.

Following the current lab testing, the idea will be taken into the field for small-scale experimentation.