This paper presents an analysis of indoor positioning systems based on Bluetooth Low Energy (BLE) beacons. BLE-based systems are commonly used in asset tracking solutions due to their low power consumption, cost-effectiveness, and ability to provide real-time location data within indoor environments. The research investigates the accuracy and reliability of trilateral positioning methods under various environmental conditions, focusing on electromagnetic interference impacts. Five distinct testing scenarios were conducted with controlled placement of beacons across diverse environments: server rooms with high interference, corridors with moderate infrastructure interference, and open spaces with minimal interference. Testing was performed using multiple mobile devices to ensure consistency of results. An ensemble learning approach combining transformer architectures with specialized positioning networks was implemented, achieving a prediction confidence of 0.974 while maintaining an acceptable error rate. The findings indicate that the primary factors affecting accuracy are not merely physical obstacles but rather the predictability and stability of the signal propagation environment. The proposed methodology offers significant improvements for indoor positioning in environments where traditional GPS signals are unavailable or unreliable.