June 2019
Intermediate to advanced
308 pages
7h 21m
English
The AE classifier we will be using will have an AE part consisting of an encoder and a decoder. The following AE architecture is used to determine the floor and building location where the Wi-Fi is located. The input to the AE are signal strengths detected in a scan. Then, one value for each visible network is considered an RSSI record. The output of a decoder is the reconstructed input from the reduced representation, as shown in the following diagram (source: Low-effort place recognition with Wi-Fi fingerprints using deep learning, Michał N. et al., arXiv:1611.02049v1):

Read now
Unlock full access