Human Activity Recognizion (HAR)

Jan ‘24 - Apr ‘24, IIT Gandhinagar

• Utilized the UCI-HAR dataset, which comprises time-series data capturing the activities of thirty subjects engaging in six different activities classified as walking, sitting, standing, running up, running straight and running down.
• Employed the TSFEL library for feature extraction from the time-series data, followed by Principal Component Analysis (PCA) to reduce dimensionality. Developed and trained a Decision Tree model using the extracted features.
• Utilized the Physics Toolbox Suite app to collect activity data for testing the trained model’s performance.

Github Repo Link : Github
Group Members:

• Aadya Arora
• Guntas Singh Saran
• Hrriday V. Ruparel
• Jiya Desia
• Mumuksh Jain
• Siya Patil