- Implemented and compared traditional localisation and basic machine learning models to determine the indoor location of a device using Wi-Fi received signal strength and obtained accurate results for a 1m x 1m grid layout
- Trained Neural Networks, Random Forests and obtained accurate results for a grid mapping of 1m^2
- Used modified DBSCAN algorithm and implemented Tries for more efficient data storage and retrieval
- Trained three classifiers using \textbf{random forests}, \textbf{CNN} and \textbf{logistic regression} with TensorFlow and performed offline classification by running the trained models on the device