Machine Mentality: UTA Inquiry Magazine

Spring 2016

Inquiry Magazine Archive

Winter 2016: Energy Evolution

From carbon dioxide conversion to landfill mining, researchers at UTA are seeking viable alternative energy options.
Spring 2016: Premium Blend

Found in everything from space shuttles to dental fillings, composite materials have thoroughly infiltrated modern society. But their potential is still greatly untapped, offering researchers ample opportunity for discovery.
Fall 2015: Collision Course

Within the particle showers created at the Large Hadron Collider, answers to some of the universe’s mysteries are waiting.
Spring 2015: Almost Human

Model systems like pigeons can help illuminate our own evolutionary and genomic history.
Fall 2014: Small Wonder

UT Arlington's tiny windmills are bringing renewable energy to a whole new scale.
Winter 2014: Overdue for an Overhaul

The stability of our highways, pipelines, and even manholes is reaching a breaking point.
2012: Mystery solved?

Scientists believe they have discovered a subatomic particle that is crucial to understanding the universe.
2011: Boosting brain power

UT Arlington researchers unlock clues to the human body’s most mysterious and complex organ.
2010: Powered by genetics

UT Arlington researchers probe the hidden world of microbes in search of renewable energy sources.
2009: Winning the battle against pain

Wounded soldiers are benefiting from Robert Gatchel’s program that combines physical rehabilitation with treatment for post-traumatic stress disorder.
2007: Sensing a solution

Tiny sensors implanted in the body show promise in combating acid reflux disease, pain and other health problems.
2006:Semiconductors: The next generation
Nanotechnology researchers pursue hybrid silicon chips with life-saving potential.
2005: Imaging is everything
Biomedical engineers combat diseases with procedures that are painless to patients.

Machine Mentality

Improving real-time decision-making in controllers

Frank Lewis

New technology developed at the UTA Research Institute (UTARI) could allow everything from airplane autopilots to vehicle emission controls to make informed, real-time decisions independently.

Frank Lewis, a professor in the Electrical Engineering Department and at UTARI, developed an innovative process that allows a device to make control decisions in reaction to a set of variables that changes based on each previous decision. Called Integral Reinforcement Learning (IRL), it involves a batch process that requires taking in a set of data before updating the control law.

"Optimal feedback controllers allow a device to use the minimum energy necessary while saving time and fuel," Dr. Lewis explains. "The advantage of using adaptive Integral Reinforcement Learning is that a device now can have optimal controls by looking within the system and calculating changes in real-time rather than offline, where changes can only be made when the device is no longer in use."

Lewis, along with Draguna Vrabie from United Technologies and Kyriakos Vamvoudakis from the University of California at Santa Barbara, was issued a patent to develop new technology for IRL to enable decisions to be made continuously in real time and online, providing greater autonomy and a speedier response.

According to Dr. Vamvoudakis, adaptive integral reinforcement may someday allow machines to adapt to, and work in, increasingly difficult situations: "Efficiency will be defined by the potential to adapt autonomously in complex environments to enable capabilities beyond human limits," he says.

Spring 2016

Think Fast

Machine Mentality

Improving real-time decision-making in controllers

Explore
More articles from this issue

Drained

Fluid Decisions

Funded

Grants

Inquiry Magazine Archive

Think Fast

Machine Mentality

Improving real-time decision-making in controllers

ExploreMore articles from this issue

Explore
More articles from this issue