Refraction AI autonomous vehicle REV-1
Refraction AI, founded by University of Michigan researchers, has developed a three-wheeled autonomous vehicle called REV-1 that is providing delivery services from local restaurants in Ann Arbor.
Photo: Refraction AI

Video Friday is your weekly selection of awesome robotics videos, collected by your Automaton bloggers. We’ll also be posting a weekly calendar of upcoming robotics events for the next few months; here’s what we have so far (send us your events!):

ICRA 2020—1 June 2020—[Virtual Conference]
RSS 2020—12–16 July 2020—[Virtual Conference]
CLAWAR 2020—24–26 August 2020—Moscow
ICUAS 2020—1–4 September 2020—Athens

Let us know if you have suggestions for next week, and enjoy today’s videos.

Refraction AI, a University of Michigan startup that began delivering food in late 2019, says its pilot deployment of five “Rev-1” robots is doing four times as many runs since the COVID-19 crisis began. The small fleet of delivery robots helps keep employees and patrons safer by limiting human-to-human contact while also helping restaurants save money on delivery services due to the lower cost of Refraction AI’s service.

There are five Ann Arbor restaurants and approximately 500 customers participating in the pilot using a custom-designed app. The battery-powered robots have a top speed of about 15 miles per hour and operate mainly in bike lanes. Their low speed enables them to use inexpensive camera-based navigation systems, making their cost feasible for a delivery service.

Refraction AI ]

This video highlighting autonomous drone research from the Multi-robot Systems Group at FEE-CTU in Prague is awesome because the first half is full of drones being awesome, and the second half is full of drones, uh, being less awesome. My favorite clip starts at 4:25.

[ CTU ]

Typically, a LiDAR is installed at the front side of a mobile robot to reduce interference with other equipment. Due to such an installation, about half of the measurement area of LiDAR can’t be used because the robot’s body blocks it. In addition, there is an unmeasurable area where the LiDAR cannot measure close to the robot. Therefore, in this research, the authors aim to expand the measurement area of LiDAR by installing mirrors in the blind area. By installing mirrors in the blind area of LiDAR, the measurement area can be expanded without disturbing the regular measurement. Based on the method, the robot enabled to obtain surface information of stairs during stair climbing, which was typically difficult to obtain.

[ Paper ]

The Lake Kivu Challenge was designed to take the world’s best drones and pit them against the greatest challenges that Africa has to offer. It brought together 10 flying teams from 7 countries on the shores of Lake Kivu in Karongi, Rwanda, to conduct 68 Beyond Visual Line of Site flights, taking drones more than 20km across water. The goal of the Lake Kivu Challenge was to create real use cases for drones to perform critical safety functions in Africa, from delivering emergency medical supplies, to collecting critical medical tests, to locating survivors after disasters.

[ ADF 2020 ]

RoboTiCan is providing telepresence robots to hospitals in Israel, including Soroka hospital in Beer Sheva and Hadassah hospital in Jerusalem.

[ RoboTiCan ]

Some very creative research on wheeled platforms from CSIRO Data61.

[ CSIRO ]

Take a look at the first unmanned Loyal Wingman aircraft prototype, one of three prototype aircraft for the Royal Australian Air Force. These aircraft are designed to extend and protect existing aircraft, and they’re the foundation for the global Boeing Airpower Teaming System aircraft to be developed for other global forces.

[ Boeing ]

The Matrice 300 RTK, DJI’s flagship commercial drone platform, features a flight time of up to 55 minutes, 6 Directional Sensing & Positioning, support for up to 3 payloads simultaneously, and more. Built to reinvent the way you work, the M300 RTK combines industry-leading airborne intelligence with unrivaled reliability.

It has 55 minutes of flight time. Wow.

[ DJI ]

Robotic insertion tasks are characterized by contact and friction mechanics, making them challenging for conventional feedback control methods due to unmodeled physical effects. Reinforcement learning (RL) is a promising approach for learning control policies in such settings. However, RL can be unsafe during exploration and might require a large amount of real-world training data, which is expensive to collect. In this paper, we study how to use meta-reinforcement learning to solve the bulk of the problem in simulation by solving a family of simulated industrial insertion tasks and then adapt policies quickly in the real world. We demonstrate our approach by training an agent to successfully perform challenging real-world insertion tasks using less than 20 trials of real-world experience.

[ Github ]

Design, manufacturing, characterization and demonstration of soft sensorized foot module providing details on different surface condition or terrain by performing experiments which could help in building a better control of soft robots during locomotion.

Our work demonstrated that a sensorized foot module (SFM) could be used to built a crawling soft robot to achieve detecting different terrain during locomotion. This study provides a better understanding of the interactions between the foot and the terrain and opens up to new way to design soft robots able to locomote on unstructured terrains.

[ Paper ]

Thanks Saravana!

I can’t imagine that Webby Awards are all that big of a deal because I’ve never won one, but IHMC is up for one and they’ve got Valkyrie on their side.

[ IHMC Stem-Talk ]

Three questions about soft robotics grippers get answered. Many other questions about this video remain.

Although, long finger life is something I’m very concerned about, personally.

[ Soft Robotics ]

Mobility of wheeled or legged machines can be significantly increased if they are able to move from a solid surface into a three-dimensional space. Although that may be achieved by addition of flying mechanisms, the payload fraction will be the limiting factor in such hybrid mobile machines for many applications. Inspired by spiders producing draglines to assist locomotion, the paper proposes an alternative mobile technology where a robot achieves locomotion from a solid surface into a free space. The technology resembles the dragline production pathway in spiders to a technically feasible degree and enables robots to move with thermoplastic spinning of draglines.

This paper is from 2014—do we have robot spiders yet?

[ BIRL ]

Satisfying video of some beefy mobile robots swapping out the floor in a gym.

[ Kawasaki ]

SLAMcore are developing the world’s first, commmercial-grade, full-stack Spatial AI SDK for robot developers. Optimised for readily available hardware, you can now build solutions that deliver the accuracy and robustness you need at a price that makes commercial sense.

[ SLAMcore ]

Sometimes I feel bad for DJI that evil drones inevitably look like Phantoms.

[ Raytheon ]

Approaches for prototyping and testing interactions between autonomous vehicles and pedestrians, from the Australian Centre for Field Robotics at the University of Sydney.

[ ACFR ]

Chris Atkeson gave a talk to middle school students about helpful robots, meaning that it’s actually fun and relaxing to listen to.

[ CMU ]

The Conversation (0)

How the U.S. Army Is Turning Robots Into Team Players

Engineers battle the limits of deep learning for battlefield bots

11 min read
Robot with threads near a fallen branch

RoMan, the Army Research Laboratory's robotic manipulator, considers the best way to grasp and move a tree branch at the Adelphi Laboratory Center, in Maryland.

Evan Ackerman
LightGreen

This article is part of our special report on AI, “The Great AI Reckoning.

"I should probably not be standing this close," I think to myself, as the robot slowly approaches a large tree branch on the floor in front of me. It's not the size of the branch that makes me nervous—it's that the robot is operating autonomously, and that while I know what it's supposed to do, I'm not entirely sure what it will do. If everything works the way the roboticists at the U.S. Army Research Laboratory (ARL) in Adelphi, Md., expect, the robot will identify the branch, grasp it, and drag it out of the way. These folks know what they're doing, but I've spent enough time around robots that I take a small step backwards anyway.

The robot, named RoMan, for Robotic Manipulator, is about the size of a large lawn mower, with a tracked base that helps it handle most kinds of terrain. At the front, it has a squat torso equipped with cameras and depth sensors, as well as a pair of arms that were harvested from a prototype disaster-response robot originally developed at NASA's Jet Propulsion Laboratory for a DARPA robotics competition. RoMan's job today is roadway clearing, a multistep task that ARL wants the robot to complete as autonomously as possible. Instead of instructing the robot to grasp specific objects in specific ways and move them to specific places, the operators tell RoMan to "go clear a path." It's then up to the robot to make all the decisions necessary to achieve that objective.

Keep Reading ↓ Show less