Welcome to the second monthly Hello Robot community update! If you’d like to see your work featured in a future newsletter or on the Stretch Community Repo, drop us a line at community@hello-robot.com.
The teams at FAIR and CMU, including Prof. Deepak Pathak, have released the Vision Robotics Bridge (VRB), allowing Stretch to accomplish manipulation tasks by learning affordances from human videos on the internet. Learn More
Hello Robot has launched a new series of video workshops aimed at helping users solve interesting problems with Stretch! Workshops are presented live via Google Meet, with recordings posted on Youtube. The next session around detecting and reaching to ArUco markers will take place on Wednesday, April 28th. Learn More
The team at the Allen Institute for Artificial Intelligence have published the Stretch Pick-and-Place Benchmark. In this work they demonstrate how agents trained with rich visual observations are able to generalize when faced with novel objects and visual diversity. Learn More
Prof. Maya Cakmak and team at Univ. of Washington have released new work on Evaluating Customization of Remote Tele-operation Interfaces for Assistive Robots. They expand on Stretch’s browser-based interface to allow individuals with severe disabilities to perform tasks for themselves. Learn More
Researchers at Oregon State, CMU, Meta, Washington, Georgia Tech, and Berkeley released Mod-IIN - a modular method for navigating to objects specified by images. Learn More
The INSPIRE lab at Spaulding Rehabilitation Hospital have shared an open-source demo of autonomous skills with Stretch, using ArUco tags to program behaviors like finding and retrieving objects on the floor. Learn More