Autonomous robots set to make food deliveries from Ann Arbor restaurants this fall

This fall, Ann Arbor residents will be able to have food delivered from select restaurants to their homes by an autonomous robot called the REV-1. That's thanks to an Ann Arbor company called Refraction AI – and a food delivery that went wrong with unexpectedly inspiring results.

 

Refraction cofounders Ram Vasudevan and Matthew Johnson-Roberson are both University of Michigan (U-M) professors with extensive research experience in autonomous vehicles and robotics. Two years ago they placed an order from Ann Arbor restaurant Miss Kim through a popular food delivery service and were disappointed when their order arrived with items missing. They called the restaurant to resolve the issue and were surprised when its owner, Ji Hye Kim, showed up personally at their door a half-hour later with the missing items.

 

Kim apologized profusely, explaining that the restaurant had been having a bad night – and on her birthday, no less.

 

"We were hearing this really sad story and we thought, 'Hey, we work in the autonomous vehicle space. This seems like a really great opportunity to start addressing some real problems,'" Vasudevan says.

 

So he and Johnson-Roberson set about developing the REV-1. The three-wheeled vehicle travels between restaurants and delivery locations primarily using bike lanes. It carries food in a large storage compartment that can be opened by restaurant staff or customers using a unique passcode entered on a keypad.

 

Vasudevan describes the REV-1 as an exciting opportunity to immediately get an autonomous vehicle on the road. Fully autonomous cars are still a long way off, but the REV-1's slow speed of 12-15 miles per hour and light weight of 100 pounds make it a safe and viable project right now.

 

"We'd been working in the space for a fairly long time and working on the large self-driving vehicle problems," Vasudevan says. "We wanted to see if there was a way that we could bring some of those things that we'd been thinking about into the real world, without having to worry about solving the full self-driving car problem."

 

Given Kim's pivotal role in the conception of Refraction, Vasudevan says she was "a really natural partner" to work with in piloting the REV-1. Refraction has also partnered with Ann Arbor's Belly Deli on the pilot project. During the pilot, the REV-1 has been used to make deliveries to employees of Refraction, Miss Kim, and Belly Deli during lunch hours. Safety bikers have followed the REV-1 on its deliveries to observe and ensure that everything goes well.

 

This fall, the REV-1 will officially begin full-fledged commercial deployment. Anyone within a roughly three-mile radius of one of Refraction's Ann Arbor restaurant partners will be able to place an order for delivery by the REV-1, and the robots will eventually begin making their deliveries without a bicycle escort. The robots will be made available to partnering restaurants for a per-delivery fee.

 

Vasudevan says he's been pleased to find that both restaurants and consumers seem excited about the REV-1. He intends to offer both groups not just a technological novelty, but also financial savings.

 

"We want to charge significantly less than what typical delivery companies are doing," he says. "Our goal is to try to bring the customer fee down by at least half, and to do something similar when it comes to the restaurant side as well."

 

Vasudevan says he's particularly excited to be launching the REV-1 in Ann Arbor. He notes that similar last-mile delivery robots have been developed by other companies, but they're mostly on the West Coast.

 

"That's a really exciting market, but it doesn't allow most of the U.S. to see one of these robots or realize the benefits of them," he says. "We thought that deploying in a place like Ann Arbor, where there's a young community that's more willing to engage with this type of thing, and also making it accessible to places that have winter weather, was really important. So Ann Arbor fit that set of requirements in a really nice way."

Signup for Email Alerts