Accuracy of a Single Position Estimate for kNN-Based Fingerprinting Indoor Positioning Applying Error Propagation Theory

Indoor positioning systems usually consider the average positioning error over a set of evaluation samples, or a quartile of that value, as the global error. However, they do not provide a metric for the uncertainty for each individual position estimation. In this article, we apply the error propagation theory (EPT) to the
${k}$
NN algorithm in Wi-Fi fingerprint-based indoor positioning. Our proposed method does not only retrieve the position estimate but also describes how the uncertainties of the RSSI measurements propagate through the calculations. We have validated our proposed method with two open-access datasets.