Indoor Position System Based on Internet of Things and Support Vector Machine

Abstract

Currently, the utilization of position tracking technology has extended to various aspects, primarily harnessing the Global Positioning System (GPS). However, a significant limitation arises in the accuracy of GPS when employed within indoor environments. In response to this challenge, the concept of the Indoor Positioning System (IPS) has been introduced and developed. IPS enables the tracking of positions within buildings and offers not only improved accuracy but also enhanced security, safety, and greater efficiency. The research described in this context aims to formulate solutions for tracking and identifying locations within indoor spaces, with a specific focus on the 7th floor of a civil building. The approach employed involves the implementation of access point infrastructure strategically placed within each room. The dataset utilized to support the analysis consists of 80 location points within these indoor spaces, meticulously divided into 40 points for the LIG2 room, 40 points for the LPY4 room, and the connecting corridor area. In the pursuit of achieving this objective, the study adopts the Support Vector Machine (SVM) method. Through the utilization of this technique, the SVM model is instructed utilizing Received Signal Strength Indicator (RSSI) data from microcontrollers. Through a series of trials and evaluations, the SVM method applied in this research has yielded promising outcomes. The accuracy rate reaches 79%, signifying that the proposed system is capable of accurately predicting locations based on the collected data.