Ph.D. student Durgakant Pushp guides an autonomous vehicle outside the Multidisciplinary Engineering and Sciences Hall on the IU Bloomington campus. Photo by Chris Meyer, Indiana University
By sea, by air or by land, the greatest achievements in self-driving technology will not be limited to cars, and they will not take place on wide open highways. In the Multidisciplinary Engineering and Sciences Hall on the north edge of the Indiana University Bloomington campus, IU's Vehicle Autonomy and Intelligence Lab, or VAIL, is charting the future of unmanned vehicles with a small fleet of miniature boats, copters and wheeled vehicles that use artificial intelligence to navigate uncertain terrains - choppy waters, cloudy skies and bumpy forest floors - without the benefit of GPS guidance systems.
Led by Lantao Liu, associate professor of intelligent systems engineering at the IU Luddy School of Informatics, Computing and Engineering, VAIL researchers are pushing the frontier of autonomous technology. Like their vehicles, they are navigating new ground without a map.
"We aren't interested in self-driving cars," Liu said. "There are plenty of companies already working on that technology. We're interested in the next challenge, which is off-road autonomy; it's about creating systems that efficiently and accurately navigate environments for which there isn't existing data. That requires designing new algorithms that can learn and adapt, and it's a challenge. It's a really big challenge."
But big challenges offer big rewards. Liu said the potential applications for vehicles that can navigate rough, unknown environments on the fly are vast, including conducting search and rescue; exploring extreme environments, such as the bottom of the ocean and outer space; and monitoring hazardous environments or weather conditions.
"We're not looking to replace the things that people already do," Liu said. "We're looking to expand the boundaries of what's possible. Autonomous vehicles can accomplish tasks that people can't due to physical or safety constraints, such as entering small, remote or dangerous spaces."
There's also the prospect of freeing up time from rote tasks that require significant mental effort without significant personal benefit, such as piloting vehicles over long distances, he added.
"Autonomous technology really has the potential for transformative impact across numerous domains," said Durgakant Pushp, a Ph.D. student who joined Liu's lab in 2021. "Robotics with AI has the capacity to revolutionize transportation, health care, manufacturing and many other industries, making processes more efficient and safer.
"From self-driving cars to health care diagnostics, robotics and AI hold the promise of improving our daily lives and addressing complex global challenges. Witnessing the evolution of these technologies, and their continued integration into society, is both thrilling and motivating."
Among the many projects in Liu's lab is a partnership with Naval Surface Warfare Activity, Crane Division, a naval base in southern Indiana, to train autonomous watercraft to map the surface of the ocean floor with remote sensors. Another project, supported in part by a grant from Amazon, is testing autonomous water vehicles with sensors that can detect pollutants, such as agriculture runoff, in local lakes and rivers. These pollutants, which are most common in rural areas, can trigger catastrophic algae blooms that kill plants and wildlife. Autonomous remote sensors on the water could alert local authorities before these substances reach dangerous levels, Liu said.
In addition to watercraft, Liu's lab is working on technology to traverse uneven surface terrain and obscured skies. The applications of these technologies include sending teamed robots into cramped areas to conduct inspection and repairs . For example, Liu said, NASA used an autonomous vehicle called a Skycrane to lower the Perseverance rover onto the surface of Mars in 2021.
To test their technology, Pushp and other student researchers at VAIL often venture out to local woods, forests, farm fields and country roads to test ground robots of different sizes and speeds for various autonomous tasks, including perception, navigation, control and communication. They also send prototype devices out to explore waterways, reservoirs and lakes near the Bloomington campus, such as Griffy Lake and Lake Monroe, and drones above the forests and farms for environmental and agricultural uses .
Although the test vehicles are small, the AI technology they use to intelligently maneuver is essentially the same, whether it's used to power a small drone or a large naval submersible or a high-end space rover, Liu said.
"You need a lot of complicated mathematics and algorithms to make this work, but the unifying concept is very simple: It's autonomy," Liu said. "The vehicles might differ in shape, design or capabilities, but they all deploy the same core technology."
In terms of student interest, Liu said VAIL currently engages over a dozen Ph.D. and undergraduate students, and some projects have attracted over 100 applicants. Pushp was drawn to VAIL out of a desire to gain the skills needed to work in one of the world's top emerging technology fields, he said.
"My ultimate goal is to work as a research scientist at a top robotics research organization," Pushp said. "With what I'm learning at IU, I plan to work on exciting projects that make self-driving and AI tech even better. This could mean making self-driving cars safer, improving other autonomous technology or solving other big real-world problems."
