Statistical Positioning Quality Metrics for Common Received Signal Strength-Based Positioning Techniques

This paper presents the development of positioning quality metrics for common Received Signal Strength (RSS) based localization systems like those based on Wireless Local Area Networks (WLAN), Bluetooth or ZigBee. Therefore, common algorithms are investigated and the necessary statistical foundations are elaborated. A confidence area and an associated probability, which are reflected by the metric, are utilized for estimating the quality of the position estimate. It is shown that the algorithm in place strongly influences these statistical parameters. Therefore, for four frequently employed techniques the metrics are elaborated and the modifications for other variants are explained. With the help of manifold simulations we show the meaningfulness of our strategy. As an example, the error of our approach is below 5% for Proximity, a frequently applied algorithm. Moreover, experiments demonstrate that our results can be utilized even in sophisticated multi-path environments.