Fingerprint-based Positioning Method over LTE Advanced Pro Signals with GAN training contribute

Localization based on wireless signals has been introduced using a positioning method that relies on satellite navigation positioning systems, such as the Global Positioning System (GPS). The main issue with a passive localization system, such as this, is that although we would only gather anonymous information, we still need the user's approval. In this paper, a fingerprint-based localization approach based on Long-Term Evolution (LTE) Advanced Pro Signal is proposed. The authors analyzed received signal strength indication (RSSI) related to the communication between the base station and the user equipment (UE). Using a machine learning approach (i.e., a generative adversarial network - GAN), the proposed system allows understanding the correlation between RSSI and the position of the user in the space. The state of art related to fingerprinting-based localization has been drawn up as a costly and time-consuming offline training phase. The GAN will generate new models (RSSI, localization), and the experimental results show improvement in accuracy when compared with the effort to perform the offline phase. The results indicate great accuracy, 0.4 meters and 10 meters, respectively indoor, and outdoor environments case and the measurement campaign was fast and time saver.