Abstract: A WiFi-based RSSI indoor ranging algorithm based on wavelet transform and neural network is proposed. The method is to correct the RSSI data and path loss model by wavelet transform and neural network. Using wavelet decom position and single-reconstruction reconstruction method, only a single-branch reconstruction of the approximate part of the low-frequency, discarding the high-frequency details, and using the neural network to train the path loss model in a speci fic environment. The example shows that the maximum ranging error, minimum ranging error and average ranging error of the algorithm are 1.206, 0.037 and 0.692 m. The average ranging error is 1.846 and 0.469 m compared with the path loss model and BP neural network model.

Key words: wavelet decomposition and single-branch reconstruction, neural network, path loss model, RSSI, indoor