Rancang Bangun Sistem Deteksi Posisi Objek dalam Rumah dengan Metode Support Vector Machine Berdasar Kekuatan Sinyal Wi-Fi

Abstract

Indoor Positioning System (IPS) is an object tracking technology that utilizes networks such as Wireless Fidelity (Wi-Fi) to determine the location of an object. IPS is closely related to the implementation of the Internet of Things (IoT) to carry out an order in a smart home. However, the weakness of IPS is the attenuation of the signal received when the tag or target moves to a room that borders another room, causing errors in tracking. The IPS implementation will be carried out based on the 2.4 GHz Wi-Fi signal emitted from the ESP32. The research will use the trilateration method which requires three sink nodes to receive signal strength, then a machine learning algorithm, namely Support Vector Machine (SVM), to classify rooms in three different scenarios, namely when the target is stationary, moving between rooms, and is on the edge room adjacent to another room. The results of the test show that the three scenarios provide different levels of accuracy. The accuracy of the system on the target scenario while still in the room reaches 100%, on the target moving room scenario reaches 86.15%, and on the target scenario that is at the edge of the room adjacent to another room reaches 80%.

Keywords

• indoor positioning system
• trilateration
• svm
• wi-fi