An integrated machine learning model for indoor network optimization to maximize coverage

Ahmed Wasif Reza, Abdullah Al Rifat, Tanvir Ahmed


Indoor network optimization is not a simple task due to the obstacles, interference, and attenuation of the signal in an environment. Intense noises can affect the intelligibility of the signal and reduce the coverage strength significantly which results in a poor user experience. Despite having numerous shortcomings and vulnerabilities in the network, still many strategies have been implemented over the years to unravel the indoor coverage problem and come up with a robust solution. Most of the existing works are associated with finding the location of the devices via different mathematical and generic algorithmic approaches, but very few are focused on implying machine learning algorithms. The core purpose of this research is to introduce an integrated machine learning model to find maximum indoor coverage with a minimum number of transmitters. The users in the indoor environment also have been allocated based on the most reliable signal strength and the system is also capable of allocating new users. K-means clustering, K-nearest neighbor (KNN), Support Vector Machine (SVM), and Gaussian Naïve Bayes (GNB) have been used to provide an optimized solution. It is found that KNN, SVM, and GNB obtained maximum accuracy of 100% in some cases. However, among all the algorithms, KNN performed the best and provided an average accuracy of 93.33%. K-fold cross-validation (Kf-CV) technique has been added to validate the experimental simulations and re-evaluate the outcomes of the machine learning models.


Coverage maximization; Gaussian naïve Bayes; Indoor network optimization; K-means clustering; K-nearest neighbors; Machine learning; Support vector machine;



  • There are currently no refbacks.

Creative Commons License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.

shopify stats IJEECS visitor statistics