Dave Feil-Seifer: Designing expressive robots for speech-language therapy
Title
Designing expressive robots for speech-language therapy
Mentor
Department
Computer Science and Engineering
Biosketch
David Feil-Seifer, Ph.D., is a professor of computer science and engineering, graduate director of the Robotics M.S. program, and director of the Advanced Robotics Center of Nevada at the ҹɫÊÓƵ. He earned his Ph.D. and M.S. in computer science from the University of Southern California and his bachelor’s degree in computer science from the University of Rochester. He completed a postdoctoral appointment in the Social Robotics Lab at Yale University.
His research focuses on human-robot interaction, socially assistive robotics and robotics education. He studies how robots can communicate and collaborate with people in socially appropriate ways, particularly in health care, education and other settings where technology must respond to individual human needs. His current work includes developing socially assistive robots to support speech-language pathologists and children with communication challenges.
Feil-Seifer has mentored more than 100 undergraduate researchers and has received University and statewide awards for graduate advising and mentoring. Outside of work, he is a distance runner, skier, and cyclist and studies piano.
Project overview
Socially assistive robots are designed to help people through social interaction rather than physical assistance. For example, a robot might demonstrate an activity, provide encouragement or help a child practice a communication skill. For these robots to be useful in speech-language therapy, their facial expressions, voice and timing must be understandable, supportive and appropriate for the situation.
This project will investigate how one aspect of a robot’s expressive behavior affects how people interpret its responses. The student will work with the M social robot, which uses an animated face, speech and body movement to interact with people. With support from the research team, the student will focus on one manageable question, such as whether changing the intensity of a facial expression or the timing of a response makes the robot appear clearer, more encouraging or more socially appropriate.
The student will begin by reading and discussing introductory research about socially assistive robots, speech-language therapy and robot expression. The student will then learn to operate the M robot and its existing control software. Using existing software as a starting point, the student will create a small set of controlled robot behaviors in which one feature is changed while the remaining features are kept consistent. For example, the robot might deliver the same encouraging statement using several facial-expression intensities.
The student will document these behaviors and help conduct a structured laboratory evaluation using existing data, recorded demonstrations or adult laboratory participants under an approved research protocol. The student will organize the resulting observations or ratings, compare the experimental conditions and identify patterns in how the robot’s behavior was perceived. Involvement of speech-language pathologists or child participants will not be required for the student to complete the project.
Throughout the semester, the student will meet regularly with the faculty mentor and research team, maintain research notes and receive feedback on research design, data analysis and scientific communication. The student will present the results in a poster at Wolf Pack Discoveries. The results may also provide the foundation for a follow-up Nevada Undergraduate Research Award proposal.
Student Training Opportunities
- Operating the M social robot and its existing behavior-control software
- Reading research articles and preparing an annotated summary of relevant work
- Designing controlled comparisons of robot facial expressions, speech or interaction timing
- Organizing and analyzing observational or rating data
- Preparing a scientific poster and presenting research findings to a general audience
Required Qualifications
Applicants should be interested in one or more of the following areas: robotics, computer science, artificial intelligence, psychology, communication sciences, education, speech-language pathology or human-centered technology.
No previous research or robotics experience is required. Some prior programming experience, particularly with Python, is helpful but not required. The selected student should be willing to learn unfamiliar software, work carefully and consistently, maintain organized research notes, attend regular meetings and ask questions when assistance is needed.
Previous participation in the Nevada Summer Engineering Academy is a plus but is not required. Students from any major who are interested in learning about robotics and research are encouraged to apply.
Maximum hours per week
The PREP student will be expected to work on the project for a maximum of 10 hours per week.
Pack Research Experience Program information and application