(Georgia Institute of Technology, 2013-03)
Park, Hae Won; Howard, Ayanna M.
In a recent conference on assistive technology in special education and rehabilitation, over 54 percentage of the sessions were directly or indirectly involved with tablets. Following this trend, many traditional assistive technologies are now transitioning from standalone devices into apps on mobile devices. As such, this paper follows this trend by discussing transforming a tablet into an HRI research platform where our robotic system engages the user in social interaction by learning how to operate a given app (task) using guidance from the user. The objective is to engage the robot within the context of the user's task by understanding the task's underlying rules and structures. An overview of the HRI toolkit is presented and a knowledge-based approach in modeling a task is discussed where previously learned cases are reused to solve a new problem.
(Georgia Institute of Technology, 2013-02)
Park, Hae Won; Howard, Ayanna M.
Over 3 million individuals in the US have a disability in their hands and/or forearms and thus have difficulties in effecting pinch and swipe gestures needed for tablet interaction. In this paper, a forearm mountable mobile interface, TabAccess (controller for Tablet Accessibility) is introduced. The objective is to provide an input interface for individuals with limited manipulation skills an alternative way to interact with touchscreen tablet applications. We believe that by combining TabAccess with mobile computers, effective education and entertainment opportunities could be delivered to persons lacking fine motor skills. For translation of gross motor gestures into touchscreen-based gestures, a methodology was developed to convert raw sensor data retrieved from the sensors into press and swipe gestures. The proposed device recognizes different gestures generated by a combination of sensors with hidden Markov models. This paper presents the design specifications of TabAccess, and discusses the training and testing results with three diverse applications - a music player, a robot controller, and a communication app.
(Georgia Institute of Technology, 2012)
Roberts, Luke; Park, Hae Won; Howard, Ayanna M.
Rehabilitation robots in home environments has the potential to dramatically improve quality of life for individuals who experience disabling circumstances due to injury or chronic health conditions. Unfortunately, although classes of robotic systems for rehabilitation exist, these devices are typically not designed for children. And since over 150 million children in the world live with a disability, this causes a unique challenge for deploying such robotics for this target demographic. To overcome this barrier, we discuss a system that uses a wearable arm glove input device to enable interaction with a robotic playmate during various play scenarios. Results from testing the system with 20 human subjects show that the system has potential, and a user specific device calibration algorithm is proposed to improve the performance of the system.