Skip to main content

Building robots with senses

Concordia professor Wen Fang Xie is programming industrial robots to see, feel and move with almost perfect accuracy
March 22, 2018
|
By Joseph Leger


Most industrial robots perform relatively simple, repetitive tasks while stationary or moving along a fixed path. Wen Fang Xie, professor in Concordia’s Department of Mechanical, Industrial and Aerospace Engineering, is determined to change that.

Wen Fang Xie Wen Fang Xie hopes her work will eventually lead to robots capable of performing any manual labour job. Photo: Concordia University

“My major project is to equip industrial robots with vision and tactile sensors,” Xie explains. “I have a colleague working on developing the vision while I’m focusing on the controls. The robots we are developing will be able to find and identify different objects, move to that location and then manipulate the objects.”

While this may sound simple, it actually is a daunting task — Xie has been working on the project since she arrived at Concordia in 2003. She is assisted by a small group of talented graduate and postgraduate students.

The team’s goal is to produce a robot that can move independently through a changing environment, locate and identify different objects and materials and then manipulate them regardless of their shape, texture or size — and this has to be repeatable countless times without error or loss of accuracy.

To achieve such consistently exact repetition, the team mounted an external “eye” or sensor able to track the robot and send it real-time corrections and location data. The sensor also plays an important part during the manufacturing portion of the tasks, allowing the robot’s limbs to “see” each other and remain calibrated while performing independent motions.

Robot Current industrial robots are mostly fixed in place and perform relatively simple, repetitive tasks. Photo: Concordia University

“Current models are static, not very dynamic, and for each point they have different accuracies,” Xie explains. “But with our external sensors we can accurately measure the position of the robot and repeatedly guide its controller to reach the correct positon at the right angle and orientation with high precision.”

Xie and her team have successfully achieved repeatable accuracy of 15 micrometres. To put that in perspective, the average strand of human hair has a diameter of 75 micrometres.

“We work in collaboration with a team at ETS [École de technologie supérieure] who are working on the static calibration,” she says. “When they saw our results they couldn’t believe it. They hired one of my students who graduated last year and it took them a year and a half to validate our results using a variety of different methods.”

Although Xie’s research focuses specifically on industrial manufacturing robots, the technology could have significant impact across different industries, including for self-driving cars, drones and domestic and medical robots.

She envisions a world where physical labour — even those jobs requiring a high-level of skill — are performed by robots, freeing humans to focus on more intellectual or creative pursuits.

“I want to equip robots with all five senses so they can replace human beings and do the jobs that are very tedious,” Xie says. “That’s my dream.”

#CUpride

Related links



Back to top

© Concordia University