Error Bound Estimation for Fingerprinting-Based Indoor Positioning Systems Using Monte Carlo Simulations

Inhalt:
In many indoor positioning applications, it is crucial to accurately estimate the user’s (or object’s) position and quantify the uncertainty of that estimate. Traditionally, metrics such as the average positioning error over a set of evaluation samples have been used to represent the global error. Recently, a method that applies error propagation theory was introduced to show how the uncertainties in RSSI measurements propagate through the calculations. This method represents the uncertainty using circles, and later, ellipses, centered on the estimated position. However, a limitation of this approach is its reliance on the assumption that a distance-based regressor must be used for error-bound estimation. This paper proposes a novel Monte Carlo-based methodology that can estimate the error bounds (using both circles and ellipses) without being constrained to a specific regressor. Experimental results demonstrate that the proposed approach outperforms current state-of-the-art methods.