/Indoor-Positioning-System

Wi-Fi Indoor Positioning System based on K-Nearest Neighbors Algorithm

Primary LanguageRMIT LicenseMIT

Wi-Fi Indoor Positioning System (IPS)
based on
K-Nearest Neighbors (KNN) Algorithm

#WiFi fingerprinting   #WiFi indoor positioning   #K-nearest neighbor algorithm

📦Indoor_positioning_system
 ┣ 📂img                               //Data Visualizations
 ┣ 📂lib                               //Supplementary Materials
 ┣ 📂src                               //Source Code
 ┃ ┣ 📂clean_data
 ┃ ┣ 📂raw_data
 ┃ ┣ 📄Step.1_Data_Cleaning.qmd
 ┃ ┣ 📄Step.1_Data_Cleaning.R
 ┃ ┣ 📄Step.2_Data_Analysis.qmd
 ┃ ┣ 📄Step.2_Data_Analysis.R
 ┃ ┣ 📄Step.3_Data_Visualization.qmd
 ┃ ┣ 📄Step.3_Data_Visualization.R
 ┃ ┗ 📄Step.99_Final_Complete_Code.R
 ┣ 📄LICENSE
 ┗ 📄README.md

Problem

Identify the physical location of indoor devices that are connected to the network.


Goals

  • Create a model that takes a set of signal strengths of the relevant access points to a connected device.
  • Predicts the physical location of that device.
  • Quantify the accuracy and precision of the model.

Results

RSSI Heat Map of All APs and Angles RSSI Heat Map based on Fast Thin Plate Regression RSSI Heat Map based on Kriging Method

k=1 k=3 k=5

Average Error Distance Median Error Distance

2.517842 m

1.902775 m