[WEEK 4 — Wi-Fi Based Indoor Positioning]

Team Members: Burak Emre Ozer, Huzeyfe Kocabas

This week we spent most of our time writing the progress report, therefore, we want to talk a little about our report and basis model of the project.

We started the report by an introduction to the Indoor Positioning System briefly. Then we explain the related works and clarified the approaches we used.

Finally, we shared our experimental results. We tried different machine learning methods such as K-Nearest Neighbor Algorithm, SVM, Gradient Boosting Decision Tree, and Random Forest to find the most effective one.

error distance (m)

According to the error distance (m) of each algorithm, we discovered that the Random Forest algorithm is the most effective one.

progress_report.pdf

Basis model

We analogized our experimental results so the basis model of the project is the Random Forest model. We will do our improvements in this model.

Map of the UJI Riu Sec Campus and zoom on the Tx Buildings. Pink refers to the ESTCE — Tx building on the UJI Campus map (left). On the Tx building zoom (right): red refers to TI building, green corresponds to TD building and blue stands for TC building. On the interior of TI building, the blue point is the reference point.

Last, while we are training these models, we think that it will not be the right way to use all the distant buildings data to predict a location, this approach will affect the results badly. So next week we will limit the data and use only the TI building that we see above red one. Salut!