Robotic service dog replacement in development

A robot mimicking the functions of service dogs is under development by a research group that includes Tech professor Charlie Kemp and his associate, graduate student Hai Nguyen.

Most service dogs require extensive investment and care to render them viable in assisting the disabled. Even then, there is no guarantee the canine will pass the final tests. Additionally, the Georgia Canines for Independence states that of the 1,200 people who apply for service dogs annually, only 500 are available. A robotic servant would be a great boon since it would not require the same amount of training as service dogs so it could potentially help many more people.

According to the Tech researchers, most research has been on “wheelchair mounted arms,” while “relatively little work has explored the possibility of assistive robots with autonomous manipulation capabilities that move independently from the user.”

Kemp and Nguyen’s research focuses mainly on helping the motor impaired, i.e. those who are unable to or cannot easily move or use their appendages dexterously.

Much of the inspiration for the robot, named El-E, comes from Kemp and Nguyen witnessing a demonstration given by the Georgia Canines for Independence. Kemp states they “were amazed at what the dog could do” and decided service dogs “seemed a great model to go by.”

El-E bears little resemblance to its canine counterpart, instead looking like a vertical cylinder on wheels with an arm extension and a motor on the bottom. The arm connects to a hand, and while it is not as flexible as a human’s, it is capable of mimicking many of the capabilities of a dog’s mouth. It opens doors, drawers and microwaves, provided there is a towel latched onto those handles from which it can grip. Kemp states this simplifies the programming since “instead of the robot having to recognize many different door handles, all it has to do is recognize one type of door towel.”

Users can give the robot up to 71 commands, all similar to what is given to service dogs; they then use a laser pointer to direct the robot to the desired location. Kemp recently presented his findings at the IEEE International Conference on Biomedical Robotics and Biomechatronics at Scottsdale, Ariz. El-E opened and closed drawers and doors with a 90 and 80 percent success rate, respectively. Besides feedback from scientific peers, Kemp was also looking for reactions from the same people his work would help. He remembers one specific example at the presentation.

“There was a disabled person in the audience and some people asked her what she thought of the work. I had already talked with a lot of patients but I wasn’t aware of how they really thought of it. But after seeing the presentation, she had really glowing praise,” Kemp said.

Students interested in robots may also consider applying to work with Kemp and Nguyen on the project. In fact, Nguyen is a third-year graduate student who stumbled onto the project after attending a class delivered by Kemp. Their research staff now consists of multiple other graduate students, including a few undergraduates from varying majors.

When will these robots be put in production? Kemp said he “will be disappointed if it doesn’t come out in at least ten years. But it depends mostly on the funding [they] get; not necessarily [him] specifically, but in the area of robotics in general.”

More information on this and other projects in the Healthcare Robotics lab can be found at http://healthcare-robotics.com