With the increasing maturity and popularity of wireless network techniques, indoor Wi-Fi positioning will inevitably become a significant application in indoor location-based services. In this circumstance, there is normally no control over the number of access points (APs) and the diversity of the Wi-Fi signal distribution, which may significantly deteriorate the positioning effectiveness as well as the system efficiency. To address this issue, we first adopt the fuzzy information entropy-based fuzzy rough set to conduct redundant APs reduction. Second, we calculate the Wasserstein distance between the signal distribution at the target position and the one at each Reference Point (RP) by the Wasserstein distance method. Third, the multisource information fusion method based on the Dempster–Shafer evidence theory is exerted to construct the matching RPs set. Finally, the abundant experiments and results in a realistic indoor Wi-Fi environment testify that the proposed method is able to preserve satisfactory localization performance as well as reduce the computation overhead of localization.

随着无线网络技术的日趋成熟和普及，室内Wi-Fi定位将不可避免地成为室内基于位置的服务中的重要应用。在这种情况下，通常无法控制接入点（AP）的数量和Wi-Fi信号分布的多样性，这可能会大大降低定位效率以及系统效率。为了解决这个问题，我们首先采用基于模糊信息熵的模糊粗糙集进行冗余AP的约简。其次，我们通过Wasserstein距离方法计算目标位置的信号分布与每个参考点（RP）的信号分布之间的Wasserstein距离。第三，运用基于Dempster-Shafer证据理论的多源信息融合方法来构造匹配的RP集。最后，大量的实验和实际的室内Wi-Fi环境中的结果证明，该方法能够保持令人满意的定位性能，并减少定位的计算开销。