Abstract—Nowadays, WiFi based Indoor Positioning System is receiving a lot of attention not just because of its low cost but also because it provides a better Location Based Services (LBS) such as assets tracking and indoor navigation . However, the accuracy and consistency of WiFi positioning system cannot meet the demands of practical applications. To solve these issues, this paper proposes different machine learning algorithms to provide an improved WiFi based positioning system. In this paper we enhance the algorithm by successfully tuning the parameters used to train the classifiers which improves the location performance. As the Received Signal Strength (RSSI) values were collected using different devices verifies the superiority of the proposed algorithms in terms of robustness to various challenging factors. Finally, extensive experiments are carried out and comparison of algorithms based on accuracies with existing classifiers to predict the location of the user in a building, floor or a precise position on the floor was made.
The research paper was selected for UKSim 2017 conference.
- Project : Research
- Category : Machine Leanrning
- Role : First Author