Indoor localization is considered as one of the basic problems when the GPS is unavailable in indoor environments like hotels, hospitals, shopping malls or multi-storey buildings. Solutions with only Bluetooth Low Energy (BLE) suffer from noise, battery consumption and limited accuracy, while solutions using cameras are unusable under low light conditions and when affected by occlusion. In order to address these individual shortcomings, this paper proposes and applies a low-cost hybrid Indoor Positioning System (hybrid-IPS) which combines the advantages of BLE trilateration and ArUco vision (ArUco marker detection). There are five ESP32 microcontrollers used: four beacons set up in the corners of the room and one scanning receiver. The received RSSI is a noisy indication and is smoothed by using per-beacon Kalman filtering, the distance is calculated according to the log-distance path loss model and the raw x, y coordinates are estimated by using a least-squares trilateration algorithm on the ESP32 receiver. The output is exposed as a JSON endpoint over Wi-Fi, and further stabilised using a 2D Kalman filter. On the PC side, a Python FastAPI server launches two concurrent asynchronous processes: periodically polling BLE position from ESP32 every 2 seconds, as well as real-time ArUco marker detection using the IP Webcam program and the phone camera. The weighted Kalman fusion module uses a 60/40 split between BLE and vision, with the resulting position broadcast to a browser client by WebSocket at a frame rate of 3 fps. The experimental results over 20 test positions in a 4 × 4 metre room demonstrate that BLE-only positioning has a mean error of 1.42 m, while the hybrid fused positioning system improves this to 0.38 m, representing a 73% improvement. Given its cost of around INR 2000, the entire system is viable and feasible for use in GPS-denied environments.

Indoor Positioning System; Bluetooth Low Energy; RSSI; Trilateration; Kalman Filter; ArUco Markers; Sensor Fusion; ESP32; OpenCV; FastAPI