This project compares two methods for indoor localization of a target node: Triangulation (using 3 Access Points) and Square Method (using 4 Access Points). By leveraging regression models, we aim to determine which method provides more accurate localization results in different environments.
Indoor localization is vital for location-based services in environments where GPS is ineffective. This project focuses on localizing a target node inside buildings using two distinct approaches: Triangulation and Square Method, to determine which is more efficient and accurate under varying conditions.
- Programming Languages: Python
- Libraries:
- Data Science & Machine Learning: Numpy, Pandas, Scikit-learn
- Algorithms: Regression Models, Triangulation, Square Localization Methods
- Data: Indoor datasets from IIT Jammu with varying conditions (obstacles and open spaces)
- Triangulation (3 APs): Localizes the target node using signals from three access points.
- Square Method (4 APs): Localizes the target node using signals from four access points.
-
Room 1: (Nescafe area, IIT Jammu) - Contains obstacles with 17 datasets.
- Outcome: Triangulation produced better results with higher accuracy and lower error metrics (MSE, MAE).
-
Room 2: (Open space, IIT Jammu) - 1.5k+ datasets without obstacles.
- Outcome: Square Method showed higher error rates and lower accuracy.
- Triangulation method consistently outperformed Square Method, achieving better accuracy and lower error metrics in both obstructed and open environments.
- Triangulation was especially effective in the presence of obstacles, while the Square method showed limitations in open spaces.
- Triangulation is the more reliable method for indoor localization.
- Square Method is less accurate and is recommended only in certain conditions where environmental factors are ideal.
To run the code and experiment with the datasets:
Ensure Python 3.x is installed. You will need the following libraries:
numpy
pandas
scikit-learn