KMITL

Permanent URI for this communityhttps://dspace.kmitl.ac.th/handle/123456789/1

Browse

Settings

Author: search.filters.author.Nimnaul, TanapolSubject: search.filters.subject.Indoor positioning system

Search Results

Now showing 1 - 2 of 2
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Item,
    Enhancing Wi-Fi-based Fingerprint Technique for Indoor Positioning System
    (2024-01-01)
    Nimnaul, Tanapol
    ;
    Bureetes, Natchapong
    ;
    Siriwat, Siwat
    ;
    Wongwirat, Olarn
    This paper addresses the enhancement of the Wi-Fi-based fingerprint technique for an indoor positioning system applied in an experimental area. The conventional Wi-Fi-based fingerprint technique utilizes a k-nearest neighbor (k-NN) algorithm for position estimation. The k-NN algorithm is a simple and intuitive classification algorithm based on distance metric, i.e., Euclidean distance (ED), but often demonstrates limited accuracy. To mitigate this constraint and enhance positioning precision, advanced machine learning algorithms in artificial neural networks (ANNs) have been introduced. Although ANN algorithms are considered highly reliable, they are complex and resource-intensive algorithms, resulting in less suitable for a small-scale area that requires simple indoor positioning applications. In contrast, the random forest (RF) algorithm offers comparable positioning accuracy while being more computationally efficient, making it a favorable choice for such scenarios. The work in this paper enhances the accuracy of the Wi-Fi-based fingerprint technique for indoor positioning systems by adopting the RF algorithm over the k-NN alternative for position estimation accuracy. The number of received signal strength (RSS) data selected from appropriate access points (APs) in the area chosen by a feature selection method is a pivotal factor influencing accuracy improvements. The experimental results express the direct correlation between increased RSS data and accuracy improvement for both algorithms. Significantly, the application of the feature selection method using the information gain ratio augments the positioning accuracy specifically for the RF algorithm.
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Item,
    Improving Machine Learning-Based Wi-Fi Fingerprint Technique with Feature Selection and Grid Search Methods
    (2024-01-01)
    Nimnaul, Tanapol
    ;
    Wongwirat, Olarn
    This paper presents the enhancement of the WiFi-based fingerprint technique for an indoor positioning system applied to a real experimental area. In typical Wi-Fi-based fingerprint techniques, classification algorithms such as k-nearest neighbor (k-NN), decision tree (DT), and random forest (RF) are used for position estimation. However, these algorithms do not perform well with high-dimensional and large datasets. They also face limitations related to overfitting and uninformative features in datasets. This paper overcomes these challenges by deploying feature selection based on filter methods. This process removes uninformative features from the datasets before feeding them to construct a training model. The grid search method is also employed to perform hyperparameter tuning, which is used to construct the outperforming models. The experimental setup involved collecting received signal strength indicators (RSSIs) from access points (APs) in the real indoor environment to create the radio map. The accuracy performance of the proposed methods was tested by employing the feature selection method on the radio map dataset and using the grid search method to find optimized hyperparameters for constructing the training models based on the k-NN, DT, and RF algorithms. The accuracy results were compared with those of non-feature selection and default hyperparameters used for the three algorithms. The computational results demonstrate that the RF algorithm outperforms the other two algorithms. Furthermore, performance is improved when using grid search and feature selection as proposed.