This project is aimed at recognizing human activities in the presence of uncertainty in the behavioral data. The approach used in this project involves training machine learning models on a synthetic dataset generated by a developed simulator.