New App from Purdue Researchers Makes Programming Robots Easy
Many experts will tell you that it won’t be long before we have access to affordable and simple robots performing practical tasks in our homes and businesses. The problem, some say, isn’t the hardware but the software: how to program and control these machines to do what you want effectively. Researchers think they’ve found the answer in the form of our smartphones.
A team of Purdue University researchers, led by professor Karthik Ramani, has created an app called VRa (short for “virtual robotic assistant”) that uses augmented reality to choreograph exactly how a robot should move and complete tasks. The app, with the help of a phone’s camera, allows users to finger-draw or walk along a path across a given space that the robot will follow.
The robot can be programmed to perform complex functions and multitask along this path—from watering plants, vacuuming, and picking up litter to transporting items to returning to its battery charger for a refresh. The app provides options for completing these duties, including a repeat function and the setting of time limits. After preprogramming via the app, the phone is placed into a dock attached to the robot to direct the tasks; in essence, the phone and VRa app serve as the brain and eyes for the robot.
The app also enables a robot to create an Internet of Things network in which the robot wirelessly interacts with another device or object after that object’s QR code is scanned and the phone is docked to the robot. Additionally, the app allows the robot to be monitored and controlled remotely. VRa uses special “simultaneous localization and mapping” algorithms, similar to those employed in drones and self-driving cars.
Ramani says the app may be instrumental in helping develop tomorrow’s smart factories, which will use augmented reality and AI to improve efficiency and enhance worker productivity. He’s also hopeful that future versions of the app can use speech recognition technology as another means of controlling the robot.
“The unique aspect of this app is its ability to transfer human motion—where you go and what you do—enabled by computer vision algorithms and location awareness of the robot,” says Ramani. “It becomes a simpler and more direct way to program a robot. In a roundabout way, the phone becomes the robot itself.”
This is important, “as the programmer need not be a sophisticated computer programmer or AI expert. It can be a common person,” says Ramani, noting that these innovations were funded in part by a grant from the National Science Foundation’s Future of Work at the Human-Technology Frontier (FW-HTW) program.
Ramani and his team have been working on human-robot interactions for years; one of their most noteworthy and acclaimed recent inventions was commercialized into a hand-controlled robotic kit called Ziro.
Purdue, which has patented the VRa technology, is now developing versions of the app for factories. Ramani indicates there are plans to make the app available for commercial use and, eventually, to license it for consumer use. “The app is designed for novice users and is intended to be useful for anyone—children, consumers and businesses,” he says, adding that young students, hobbyists, and others without advanced mechanical skills can build a simple robot on wheels today that can be controlled in the near future by VRa.
“As robot costs continue to come down, many people have the ability to build a basic motor-driven base that serves as the robot,” says Ramani. “Potentially anything on wheels with an interface can be controlled. But they need this app and their phone to function as its eyes and brains.”