Abstract

Positioning is required for a variety of applications in indoor environments where global navigation satellite systems (GNSS) are unavailable. Wi-Fi infrastructure is widely available in buildings, but conventional positioning relies on the received signal strength indicator (RSSI), which degrades from propagation effects or requires fingerprinting surveys. The IEEE 802.11mc Fine Timing Measurement (FTM) protocol addresses these limitations by using signal round-trip time (RTT) to estimate distances. Although FTM support is available in several low-cost microcontrollers, performance in real-world deployments is challenged by environmental and hardware-related factors.

In this thesis the design, implementation and experimental evaluation of a complete indoor positioning system built entirely with ESP32-S2 development boards is presented. The system consists of five fixed devices that act as anchors and one mobile initiator that performs ranging sessions and sends measurements to a server. The positioning pipeline consists of calibration, outlier rejection, multilateration and kinematic filtering.

Experimental evaluation in a realistic indoor environment shows the progressive performance of each pipeline stage. The mean positioning error is 3.70 m without calibration, and after per-anchor linear calibration, the error is reduced to 0.73 m, which is an 80.2% reduction. Weighted least-squares multilateration with residual-based outlier rejection further reduces the error to 0.64 m, with 90% of estimates below 1.27 m. The integration of a Kalman filter, optimized for pedestrian motion, improves position stability and jitter. Finally, by adding one more anchor in the shadowed region, the mean error is reduced to 0.28 m (a 56.3% improvement). The results demonstrate the feasibility of low-cost 2.4 GHz microcontroller platforms for sub-meter indoor positioning, provided that systematic biases and the anchor geometry are properly addressed.

Keywords

Indoor positioning, Wi-Fi RTT, IEEE 802.11mc, FTM, ESP32, multilateration