JMU Engineering Students Test Self-Driving Cart

Nov 20, 2020

JMU X-Labs employee Kim Fisher takes a ride in the autonomous golf cart.
Credit Calvin Pynn

A team of engineers and students at James Madison University has outfitted a golf cart to drive itself.  And it’s not meant for the golf course.  The autonomous cart is designed to help transport elderly passengers safely across retirement communities. The group test drove the cart on campus this week, and WMRA’s Calvin Pynn took it for a ride.

[Sound of cart approaching]

From a distance, it looks like any normal golf cart, outfitted with a decorative fin on top and some decals on the hood. But as it gets closer, you can see network of cables, boxes, and sensors that, collectively, scream 'research equipment.'

AI VOICE: SELECT YOUR DESTINATION

On a golf course, you would drive the cart. But on JMU’s East campus, the cart drives you.

Reporter: Do I want to go to the café? I’m gonna say that.

NATHAN SPRAGUE: It’s more fun if you say it.

AI Voice: DRIVING TO CAFÉ.

And with a tap on the screen or a voice command, as the team prefers, it takes off.

Watch a video of Calvin riding in the JACart!

[Sound of golf cart driving]

The JMU Autonomous Cart – or JACart – research group developed this golf cart by outfitting it to sense objects and navigate a path on its own. It all started with an autonomous vehicles class in the Spring 2018 Semester through JMU’s X-Labs.

SAMY EL TAWAB: By the end of the class, we were able to make the cart drive itself autonomously without hitting anything in the parking lot of the JMU X-Labs.

The class won the Governor’s Technology Award for the cart they built that semester. Samy El Tawab is one of JACart’s faculty members.

El TAWAB: From there, we started thinking of how we can get some external funding, and how we can use it.

They got that funding last year in a grant through the Jeffress Trust Program and began developing the cart they’re currently testing. This plan for this one is to transport residents in retirement communities.

EL TAWAB: So, what we discovered that since we are in Shenandoah Valley, and there was a lot of retirement communities, I think retired people will be interested in something like this.

They began working with Bridgewater Retirement Community when they learned the residents were on board with the idea. Health metrics and other personal data have been a priority in developing how the cart will communicate with elderly passengers.

[WMRA's operating license is held by JMU's board of visitors, and both JMU and Bridgewater Retirement Community sponsor programs on WMRA.]

EL TAWAB: When I talk about a golf cart, it simplifies a lot of things. So, an autonomous driving car will need to take care of them with signals and the traffic. But when I talk about the golf cart, now I'm talking about the more simplified scenario, you know?

The team is still figuring out how to store that data, how long to store it, and how to keep it secure. JMU Engineering professor Jason Forsyth is helping develop wearable technology for the goal. 

JASON FORSYTH: We would like to make those measurements sort of as non-invasive as possible. I mean, we don't to give you the ability to take an autonomous car somewhere and then have you wear a jumpsuit of sensors. Right. Um, so can we measure heart rate? Can we measure your respiration rate? Is there something in the way that you look? Cause we have cameras that if we can detect that you're getting anxious in the cart, maybe we slow it down. Maybe we stop it. Maybe we just adjust its behavior to make you feel a little more comfortable.

A team of undergrads has been working on the project alongside JACart’s faculty members. Mridul Pareek is a Junior at JMU studying computer science, and it was his first time working on self-driving vehicles.

MRIDUL PAREEK: I'm 25 now, so I started a bit late, I'm from India. I worked as a game developer for like four years. So, I had a good grasp of the web technologies and like a bunch of other stuff. So that actually helped me with that.

That came in handy when it came to developing the cart’s controls.

PAREEK: Where I fit in is, I handle all the user interface, which you'll see. And there's an app where you can actually call the card and it talks to the server online. So, you just have to be connected to the internet, you can call the car from anywhere you want.

Josh Griffin is a senior also studying computer science. His job has been to work with the cart’s software.

JOSH GRIFFIN: It’s called Ross, the robotics operating system. So, it controls the navigation and the cart figuring out where it is. It's kind of the brain. Um, I'm still relatively new, but it's been a lot of fun so far.

Nathan Sprague shows off the equipment in the cart.
Credit Calvin Pynn

Nathan Sprague is another one of the professors on the team. He said the students did most of the heavy lifting.

SPRAGUE: 95% of the code, 95% of the electronics, is done entirely by the students. So, you know, any contributions that as faculty have made, we've just been sort of stepping in and helping them when they have questions for us.

Several people took turns riding in the cart as it navigated a short, fixed path along JMU’s East Campus quad. Griffin compared it to riding in a roller coaster.

GRIFFIN: You're completely out of the control loop, so you’re just along for the ride. And you know, when the steering wheel all of a sudden turns on its own, you're just like: “Oh!” A little nervous, but it’s like: “oh, this is working.”

The cart moved slowly at just about 10 miles per hour, but Griffin’s description isn’t far off. It handled smoothly, even when it had to move as two Starship delivery robots – the other self-driving vehicles on campus – veered close to its path.

(Sound of Starship robots passing by)

As the team was awarded another grant this past summer, they will continue working on the cart through next semester.

SPRAGUE: There's still a quite a bit of refinement to do in terms of tying up loose ends and in the functionality. Making everything a little cleaner, a little smoother, a little more reliable. And we're also kind of to the point where we can start doing the science that we want to do with the system.

Until it’s time to take to the paths of retirement communities, the cart will continue living in JMU’s EnGeo building.

AI VOICE: DRIVING TO HOME.