"Robotic Map Generation"
It didn't pay his bills or iron his shirts, but it won him distinction with one of the most novel and fun soar projects of the past year: a spry robot. Toward the end of the spring semester, Wesley Moser '08 was a familiar sight in the Haupert Union Building in the company of his colorful creation, often trailed by curious students and faculty.
It was a robot he had designed to tell a computer program where it had been. As it ambled around a flat terrain, steering clear of obstacles, its sensors relayed its path electronically to the computer, which then displayed its route as a line-map on a screen.
The little wanderer’s frame is built of lego blocks—yellow, blue, red, and green—prized artifacts of Wesley’s childhood. To provide mobility, two miniature tractor tires anchor the rear, and steering wheels are mounted on the front. Its innards are packed with sensing and computing equipment. A short, revolving antenna in the forward section uses a sonar type mechanism to survey the terrain ahead of the moving object. It sends what it finds to the electronic processor behind it, which keeps track of everything the sonar senses and all motion that has occurred. The information eventually appears on the computer screen.
It has the look of a small delivery truck with a forward “cab” and a rear cargo container crammed with technological gadgetry. To Wesley, it is an achievement even he sometimes finds hard to believe, and a source of sheer delight. “Anything I’d built before I’ve had to push,” he noted.
The research started when Ben Coleman, assistant professor of computer science, asked if he’d like to do a project in the field. From the shelves of the department he found the abandoned remains of a robot-like object that had been used for an experiment a decade ago. That touched off his lego affinity and his curiosity about robots. From that germ of an idea, the concept gradually took shape.
The first goal was to keep the robot from bumping into things. This involved installing sensors that would measure the distance to objects. As the robot approached an object, it would recognize a reduction in the distance, and turn to avoid the object. No more crashes.
Finding a means that allowed the device to map its direction proved even tougher, Wesley said. “The robot can’t see where it is,” he said, “so that presented all kinds of challenges. It involved lots of math. All the robot does know is how fast it can go.” Where and how far had to be determined by ingenuity. Persistence and know-how would pay off when he was able to make the biggest leap, translating the data gathered from charting the robot’s location and devising a program that would show its track on a computer screen.
“The roughest part was the robot telling the computer what it did,” he said. Math again came to the rescue as he turned data from “the real world to the images of a computer world.” In its finished form, the robot scoots over a surface within an enclosed area and its meanderings appear as crooked lines on a computer monitor.
Wesley is majoring in both mathematics and computer science. He counts as great good fortune his opportunity to learn hands-on skills from his father, an electrician with whom he has worked in partnership. He would like to learn more about robots, but he’s not yet ready to embark on the frontier of artificial intelligence. It’s a bit “scary,” he said, “but then again I don’t understand how it works.”
Learn more about the SOAR program...
