Yuan Xing and Anne Schmitz, both assistant professors in the Engineering and Technology Department, both incorporate a neural network machine learning process as a way for AI to learn their tasks. Each of their individual research projects focus on a different topic, but they represent two of the many different AI-based research initiatives happening at the university level.
For Xing, he said his research tests the application of AI to differentiate and evaluate fresh fruit and produce.
“This project is to improve the smart manufacturing system,” Xing said. “We’re going to rely on an AI model to control multiple robots. It is also implemented as an automatic control system, or automation, in smart manufacturing.”
By using this technology and conducting research on it, Xing said their hope is to create a more accurate system to sort fruits by category or freshness.
“We can save human labor, so we don’t need a human to literally check every product … We can also speed up the production,” he said.
On top of that, Xing also said the system could increase the sustainability of produce as the more accurate measuring of the AI model can lead to less produce discarded and wasted by that means.
Xing said the current production line does not have a fully automated AI detection system in place, but there is an opportunity to implement the model in industrial-level robots with a prototype created through his research.
One concern that people may have with this technology is that it could lead to jobs being lost to artificial intelligence, but Xing said he does not necessarily believe that will be the case.
“That is a very important thing people need to discuss,” Xing said. “Because, some people may think that if robots can replace human labor, then people are losing jobs. I will say if they use a robot, you can train human labor to learn how to manipulate the robot … The labor no longer needs to do the job manually, they can control the AI model or maintain the robot to do the job.”
Xing also said he is seeing more and more of his colleagues take up AI in their research, especially in his department where they look at solving modern engineering problems.
“AI is really popular, especially in Wisconsin,” he said. “I think that is because more and more people are using ChatGPT. ChatGPT is very popular; however (the technique used) is only a very small part of artificial intelligence … There is a lot of AI technology different from the one in ChatGPT; you can use them to improve the system performance in engineering systems.”
With Anne Schmitz’s research, she said her focus has been on adapting AI to create structures and product designs like ones found in nature.
“My work is on lattice structures,” she said. “Mother nature is really good at minimizing materials while keeping things nice and strong. I’m looking at how we can use that to create foam structures.”
As an example, Schmitz said one area of her research has been focused on creating foam inserts for football helmets which better protect athletes’ heads while in a game.
“I am looking at how we can optimize these geometries,” she said. “First, I teach it a bunch of geometries and how they act, and then once I have taught it I am trying to use it now to say, ‘Let’s say I want to improve a football helmet. What is the right geometry? How can I make it absorb the impact that I need for a football helmet?’”
But that concept surrounding lattice structures and the AI’s ability to form geometric models can extend to other potential applications as well, including bicycle seats or even efficient airless tire models.
Through her experience and working in academia, Schmitz sees AI as a tool that can be used for efficiency, whether it is with her research in creating models or with students using it to help study for a test.
While AI technology is becoming more and more popular, Schmitz said we are still quite a ways away from potentially seeing the application of this tool everywhere. With continued research, these professors are measuring the potential of its application and the success of their models in order to use it for practical application.
©2024 the Leader-Telegram (Eau Claire, Wis.). Distributed by Tribune Content Agency, LLC.
