A group of scientists from Georgia Tech created an innovative mobile application designed to make robot interaction much easier for everyone. Known as COBALT, this tool lets individuals without technical backgrounds manage robotic arms from distant locations simply by using their smartphones and a stable web link.

Developed in Georgia Tech’s People, AI & Robotics (PAIR) Lab, the setup converts phones into gesture-based controllers. By tilting and moving the device, users guide the robot's actions instantly. Functions like grasping, relocating, and dropping items are handled via touchscreen buttons, resembling a casual smartphone game more than complex machinery operation.

Ayush Agarwal, a doctoral candidate in the School of Interactive Computing at Georgia Tech and head of the COBALT project, explained that the design focuses on accessibility for newcomers. In trials, people from places like India, Indonesia, and Pakistan successfully directed lab robots in Atlanta with zero previous knowledge of robotics.

Experts see crowdsourcing as key to advancing robotic capabilities

The main objective of COBALT goes further than ease of use. The team thinks it could address a major issue in robotics by gathering vast amounts of practical data needed to enhance AI-driven robot performance.

Today's robots need huge datasets for reliable task execution. Assistant Professor Animesh Garg, leader of the PAIR Lab, notes that virtual simulations fall short for widespread use. They imagine a worldwide network of phone users contributing data through casual remote sessions with robots.

Garg highlighted the potential of engaging with billions of smartphone owners globally. Reducing entry barriers aims to build an extensive system that speeds up robot training and automation processes.

This approach also promises benefits for learning environments. Recently, the researchers showed COBALT to Atlanta's Midtown High School pupils, who operated robotic arms via phones. Such an easy interface might bring robotics into more schools lacking costly tools.

Robot gig economy could emerge soon

Additionally, the creators foresee COBALT enabling novel remote job opportunities. Garg mentioned a potential economy where individuals control helper robots in various settings like homes or industrial sites from afar.

Practically, this means robots handle routine work but call for human input on tricky issues. Remote users could step in briefly via phone before returning control to the automated system.

Agarwal noted that studies found phones more appealing than VR or other devices due to their natural feel and accurate data capture. WebRTC tech keeps delays low, like in video calls, for smooth operation over distances.

The COBALT study will be shared at the IEEE International Conference on Robotics and Automation in Vienna this week, highlighting both the platform and its potential for large-scale remote networks.