Human activity recognition using ML and AI methods

Project Type: Academic

Main Challenge:

Extracting information from data became useful in various fields. Human activity can be recognized from the raw sensors data and can be used for functional, and behavioral health assessment, remote monitoring of patients, smart home activity recognition, sports analytics, fitness tracking, transportation analytics. Human Activity Recognition refers to human daily activity recognition in an automated fashion using data mining and machine learning approaches.

Funding: Internal

DA&AI tools:

• Adversarial Autoencoders
• Transfer learning

Publications:

• Balabka, D., Shkliarenko, D., 2021. Human activity recognition with AutoML using smartphone radio data, in: Proceedings of the 2021 ACM International Symposium on Wearable Computers. ACM, Virtual USA, pp. 346-352. https://doi.org/10.1145/3460418.3479377
• Balabka, D. (2019). Semi-supervised learning for human activity recognition using adversarial autoencoders. 10.1145/3341162.3344854.

Student theses:

• Balabka, D. (2019). Human Activity Recognition with Smartphone Sensors using machine Learning Algorithms