Indoor scenario-based UWB anchor placement optimization method for indoor localization

Hao Pan,Xiaogang Qi,Meili Liu,Lifang Liu
DOI: https://doi.org/10.1016/j.eswa.2022.117723
IF: 8.5
2022-11-01
Expert Systems with Applications
Abstract:Currently, ultra-wideband (UWB)-based localization systems are attracting much attention due to their high ranging accuracy. It is well known that the relative sensor–source geometry and non-line-of-sight (NLOS) measurement can severely impact the accuracy of the location estimate. To improve the localization accuracy, we propose an anchor placement optimization method that is based on the corresponding indoor scenario. Specifically, concrete walls affect the ranging error and constrain the placement of UWB sensors. An indoor scenario modeling method is presented for generating NLOS ranging paths and dividing the search space. Furthermore, a stochastic/deterministic combination ranging model is proposed based on the through-the-wall path and Gaussian process regression technique. We propose a heuristic differential evolution algorithm for searching for an optimal anchor placement that is based on minimizing the Cramer–Rao lower bound (CRLB) and predicting the ranging error. In addition, a region division method and a region combination method are proposed for reducing the search space, optimizing anchor placement in parallel in different regions, and providing full line-of-sight (LOS) coverage regions to the greatest extent possible. Finally, field experiments demonstrate the effectiveness of the proposed indoor scenario modeling method and ranging model. The proposed indoor scenario-based UWB anchor placement method achieves decimeter-level localization accuracy. A thorough comparison confirms the localization performance improvement attributed to the proposed anchor placement method.
computer science, artificial intelligence,engineering, electrical & electronic,operations research & management science
What problem does this paper attempt to address?