Multi-Value Learning Rate Approach for Indoor Real-Time Localization with Bluetooth Low Energy

This paper introduces an indoor realtime location system (RTLS) using Bluetooth Low Energy (BLE) beacons. The system consists of fixed transmitter beacons and a mobile receiver beacon (the target to be located). The RTLS works in two stages: first, it estimates the distance between the mobile beacon and each transmitter. Then, it calculates the mobile beacon's coordinates within the designated area. The gradient descent is used to calculate coordinates for a more accurate method than the trilateration. Using multiple learning rates can accelerate convergence to the desired error. This approach improves location accuracy, making it suitable for real-time applications.