Wireless indoor localization based on multispectral waterfall maps

Wireless indoor localization is becoming increasingly important with continuing technological advancements and the new demands of the business model. This paper proposes a wireless localization method based on the waterfall maps of GSM(Global System for Mobile Communications) spectrum and Wi-Fi spectrogram. The method adopts the naive Bayes classifier to classify various locations based on the frequency characteristics of those locations. The data in different frequency bands can be combined to perform grid anchor positioning based on frequency characteristic classification. Meanwhile, the indoor experiment with an anchor spacing of 1.2 m achieves a localization accuracy of over 90%. The experimental results demonstrate that combining the frequency bands of phones and Wi-Fi can be an effective strategy for indoor wireless localization.