Research Projects

Research Projects 

Activity Recognition through Smart Glasses

Faculty advisor(s): Dr. James Kiper (CSE)

Undergraduate Researcher(s): N/A

Graduate Researcher(s): Ms. Mrittika Raychoudhury

Project description: Falls are a major health risk for older adults and patients, as they diminish the quality of life and are a major cause of pain, disability, loss of independence, and premature death. Older adults tend to use different types of glasses for different activities (e.g., near vision glasses for reading, distance glasses for walking). Forgetting to change near vision glasses into distance glasses when switching to a different activity, or vice-versa, is a major cause of falls. We plan to address this problem by detecting when a user switches to a new activity and alerting them to change glasses when a relevant activity transition takes place. We plan to use inertial sensors (accelerometers and gyroscopes) to capture user movement data, to determine what activities users are engaged in, and to alert them when switching glasses is necessary.

Project goals:

  • Develop a head-mounted inertial measurement unit (IMU) based system for detecting Activities of Daily Living (ADL)
  • Recognize at least 10 ADLs through multiple supervised and unsupervised algorithms
  • Detect relevant activity transitions and alert user
  • Receive feedback and improve the system

Funding Source: N/A

Work Produced: MS thesis

Mapping the kinematics of dog gait

Faculty advisor(s): Dr. Mark Walsh (KNH) (Assisting Dr. Jessica Sparks with this project)

Undergraduate Researcher(s): Currently Paul Filonowich (Rotates through seniors who take this on as their senior project)

Graduate Researcher(s): Tess McGuire

Project description: We have created a dog model in our motion capture software. We attach reflective markers to joint landmarks on a dog (Dollie) and have her walk through our calibrated space where her motion is captured by 9 infrared cameras. Our motion capture software allows us to create an avatar of Dollie so we can see all of the displacements and velocities of the markers and the angles of Dollie’s joints.

Project goals:

  • To learn enough about dog gait to allow us to make a prosthetic device to assist a dog who is missing a leg.

Funding Source: None

Work Produced: Senior project, possibly undergraduate research symposium