Indoor Localization using RSSI based Supervised Machine Learning Approaches

Wireless indoor localization has attracted significant interest from industrial and academic domains due to its crucial role in location-based applications. However, localization is not an inherent feature of the ZigBee devices; it is possible to employ wireless technology capabilities for localization. This paper uses machine learning models to map the indoor localization system between the received signal strength indicator (RSSI) and the node position. A system map is created using a fingerprint technique containing an RSSI (input)position(output) pair. To study the influence of RSSI on the localization algorithm, we proposed framework employs four different Supervised Machine Learning models, including K-Nearest Neighbour (KNN), Feed-Forward Neural Network (FF), Cascade-Forward Neural Network (CF), and Long-Short - Term Memory-Recurrent Neural Network (LSTM-RNN). The results showed that $\mathrm{Wi}-\mathrm{Fi}$ technology achieved the best performance in the CF model over all.