INDOOR LOCALIZATION SYSTEM BASED ON WEARABLE DEVICE, BLUETOOTH LOW ENERGY (BLE) BEACON, AND ARTIFICIAL INTELLIGENCE

• Main Author: Baejah
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/84556
• Institution: Institut Teknologi Bandung

Indoor localization is a technology used to quickly respond to events that may occur in cases of dementia. Dementia is the leading cause of death in the elderly, not because of Alzheimer's itself, but due to vulnerability to other diseases that arise because of Alzheimer's, such as infections and falls. Therefore, monitoring the activities of the elderly at home is crucial to minimize deaths due to falls. The users of this system are those who need to know the position of patients as detected objects. This research aims to position patients using Bluetooth Low Energy (BLE) IBKs 105 as a Received Signal Strength Indicator (RSSI) transmitter, which is a simple and affordable solution. A wearable device consisting of Esp32 is used by patients as an RSSI scanner from BLE Beacons, which then sends the RSSI data to an edge device in the form of a Raspberry Pi to be processed into a decision regarding the patient's location. In this experiment, the data processing method employs boosting algorithms (XGBoost and LightGBM) compared with other machine learning and deep learning algorithms. This model produces a classification performance metric namely F1-Score, precision, and recall with an accuracy of 91,000%. Model optimization is carried out to enhance accuracy and computational efficiency, resulting in a real-time room prediction decision process of 1,300 seconds.