Deep User Identification based on Motion Patterns

Motivation.

Smart environments require to identify users in a continuous, effective, and unobtrusive manner. However, existing approaches are only able to work when users execute specific activities, and typically their performance degrades as more users are considered.

Goal.

Devise and implement a machine learning approach to continuously identify users in indoor environments.

Requirements.

- Eagerness to learn and work hard

- Programming experience

- Some hands-on experience in machine learning