Image Credit: Refraction AI
In Ann Arbor, MI, three-wheeled autonomous food delivery robots are on the road. Two University of Michigan Engineering Professors co-founded Refraction AI with the intent to create a cost-effective and contactless food delivery service. The result is the REV-1, a sleek silver bullet that navigates the streets of Ann Arbor at roughly 15 miles per hour, delivering food from 11 local restaurants.
Unlike food delivery robots roaming other college campuses, the REV-1 cannot drive on the sidewalk. The small size and weight of the REV-1 robot and its level of power allow the robot to be considered an e-bike, so the REV-1 can operate in bike lanes. In addition, because of its speed and design, it can operate in traditional car lanes, something other food delivery robots cannot do. The REV-1 can also be used in different types of weather, including snow, and is robust enough to conquer the city’s potholes. Recently, Refraction AI began testing the REV-1 robot in Austin’s Central Business District in partnership with six local restaurants there.
Revved Up for Delivery
Unlike a traditional LiDAR (Light Detection and Ranging) system that uses lasers to measure ranges, the REV-1 uses integrated ultrasound sensors, millimeter wave radar, and cameras. The technology helps the robot to better navigate the street, stopping for pedestrians and traffic control signs when necessary. It also allows for a short stopping distance, which helps prevent accidents and makes REV’s road journey easier. Artificial intelligence and an integrated GPS helps the car-like robots navigate through the city.
The REV-1 is battery powered, and it can operate for about 12 hours before needing to recharge. The robots recharge at Refraction AI’s headquarters in downtown Ann Arbor. The inside compartment of the vehicle, where food is placed, is equipped with ultraviolet sterilizing lights that clean the inside compartment between uses.
When a customer places an order, it triggers the REV to go to the restaurant and provides the delivery address. An employee enters a security code on the REV’s exterior touchscreen, which opens the compartment, and the customer is notified that the food is en route. When the REV arrives at the destination, the customer enters a specialized access code to open the compartment and collect their goods. The entire process is contactless, something many people look for amid the pandemic.
Refraction AI is still working out the nuts and bolts of REV-1. A Refraction AI team member riding an electric scooter follows the robot as it delivers orders. The team member monitors the robot’s activity and notes any issues. Refraction AI can then use this data to improve the REV-1. In the future, the company hopes that the REV-1 could be used for pharmaceutical, grocery, and other store deliveries.
Most delivery robots have a smaller form factor and restricted delivery paths. Starship Technologies has multiple robot fleets across the country. These robots are smaller and slower than the REV-1 but have been very popular on college campuses. Purdue University and the University of Wisconsin-Madison are among the universities using Starship robots for food delivery. The Starship robots use cameras, GPS, ultrasonic sensors, and 3D radar to navigate their environments. Like REV-1, they are connected to a mobile app that allows customers to track their robot and unlock it upon its arrival. Kiwibot has developed a similar autonomous robot for food delivery on college campuses.
Other companies are creating similar delivery robots. FedEx created the FedEx SameDay Bot to deliver a variety of small cargo in the US. Amazon launched Scout, an autonomous delivery robot that delivers packages to residents in the United States. Delivery robots like these offer a convenient and contactless method to get food, groceries, and more.