Reduction of Signal Strength Data for Fingerprinting-Based Indoor Positioning

Indoor Positioning Services (IPS) estimate the location of devices, frequently being mobile terminals, in a building. Many IPS algorithms rely on the Received Signal Strength (RSS) of radio signals observed at a terminal. However, these signals are noisy due to both the impact of the surrounding environment such as presence of other persons, and limited accuracy of strength measurements. Hence, the question arises whether raw RSS data can undergo transformation preserving its information content, but reducing the data and increasing the resilience of the algorithms to inherent noisiness of raw RSS data. This work evaluates the use of binning applied to RSS data and proposes the way the reduction of RSS data can be attained. Results show that the proposed way of using RSS binning causes little or no positioning accuracy degradation. Still, it yields significant RSS data reduction. Interestingly, in some cases data reduction results even in accuracy gains.