Previous Abstract Return to Session E3 Next Abstract

Session E3: All-source Intelligent PNT Method

Geodetic Network Optimization Based on a Multipath-Resolved Radio-Environment Map
Agnes Koller, Maximilian Schuh, Carlo Alberto Boano, Kay Römer, Klaus Witrisal, Graz University of Technology
Date/Time: Thursday, Sep. 22, 10:40 a.m.

Peer Reviewed

Highly accurate, robust, and reliable position determination is a topic attaining broad interest by the research community, since there are many scenarios and applications where current satellite-based positioning systems are not available or suitable, e.g. assisted living, robot navigation, and asset tracking. Robust and accurate indoor positioning can be provided by ultra-wideband (UWB) time-of-arrival measurements of the line-of-sight (LOS) component. However, in a UWB positioning system with many nodes, the problem occurs which measurements shall be conducted. To enable scalable real-time localization systems (RTLS) with an increasing amount of devices, reducing the number of nessessary rangings is of high importance. To reach a robust and accurate position solution, the determination of a useful LOS component is neccessary. To this end, a positioning system with a high time resolution is needed. Low-cost UWB receivers such as the DecaWave DW1000 use a high bandwidth and offer a fine-grained time resolution enabling cm-accurate ranging. In indoor environments, multipath interference has a significant influence on the position accuracy. To address this issue, multipath-resolved radio environment maps are formulated to improve the reliability of a multipath-assisted positioning system. The core idea of a multipath-resolved REM is to map the expected utility of position measurements to different anchor nodes (see Fig. \ref{fig:overview}) throughout an environment to be covered. Considering a positioning system with dozens of nodes, this also addresses another problem, i.e., which sensor configuration shall be used to reach a pre-defined accuracy criterion. Previous work did not consider including multipath-resolved environment maps for finding an optimal configuration of a wireless network in harsh environments for a robust measurement selection. This paper adresses the quantification of performance metrics to formulate a location-resolved environment model for harsh environments, derived from the Cramér-Rao lower bound on position-related signal parameters. This environment map enables the prediction of dependable position-related performance parameters and thus the real-time design of a balanced and optimized network configuration by a selection of measurements for precise positioning. To validate our theoretical results, experimental data is analyzed. In this work, we evaluate throughout practical measurements the selection of a minimum sensor configuration by geodetic network optimization strategies, based on a Cramér-Rao-lower-bound (CRLB) on signal level. The CRLB is used to determine a multipath-resolved-radio-environment map (REM) in terms of the effective Signal-to-Interference-plus-Noise-Ratio, which provides the ranging intensity information within diffuse multipath and also the measurement reliability due to the „visibility“ of an anchor node. Hence it represents the connection between accuracies of the ranging measurements and the reachable accuracy of the position measurement and can be used as an input for geodetic network design to find an optimum sensor configuration before conducting actual measurements. We found that linking performance metrics of positioning systems to an environment model by integrating it in geodetic network design enables a proper selection of available sensor nodes. The results, obtained from a simulation environment, indicate so far, that performance metrics of positioning systems can be mapped geometrically, exploiting the localization and channel estimation capabilities of the considered UWB system. Our preliminary results also show that decreasing the number of measurements based on a balanced observation plan does not decrease the accuracy and allows to reach the targeted performance bounds received from our multipath-resolved radio-environment map. Investigations about the potential of linking different disciplines (wireless communications, geodesy, statistical signal processing) will be performed and the results will be analyzed statistically. The developed algorithms will enable linking performance bounds for a proper measurement selection process to positioning algorithms. In summary, our findings suggest that the proposed radio environment map links analytical performance bounds to geodetic network algorithms for measurement selection in harsh environments. This approach increases the position reliability and accuracy enabling location-aware, resource-efficient indoor-positioning. The interdisciplinary approach of this paper (ToA-based localization, geodetic algorithms, statistical signal processing, and geomatics) will eventually lead to the design of a reliable positioning system by using fundamental performance metrics inside geodetic network optimization in presence of dense multipath.



Previous Abstract Return to Session E3 Next Abstract