Extraction and Characterization of Location-dependent UWB Radio Features with Practical Implications for Indoor Positioning

Abstract:
This paper regards the exploitation of a time domain Ultra Wideband (UWB) channel measurement campaign, which was carried out in the [3:7]GHz band in a typical indoor office environment. Relevant locationdependent radio features are extracted out of the Channel Responses (CRs) observed at four synchronized receivers spatially disseminated in the building, for various transmitter locations. These metrics are subsequently analyzed from the perspective of localization. First of all, we consider correlations between the extracted radio features and environmental variables (e.g. transmitter coordinates, transmitter-to-receiver distances, channel obstruction conditions). The analysis also comprises interpretations based on deterministic predictions in the very environment, emphasizing local shadowing effects due to the presence of metallic pieces of furniture. Then statistical models are particularized according to geographic sub-areas (or serving access points). Based on the previous results, practical implications are finally drawn for context-aware solutions based on Time Of Arrival (TOA) estimation or Received Signal Strength Indicator (RSSI), aiming at the self-adaptation of positioning algorithms to their local operating conditions.