Exploiting 2-D Representations for Enhanced Indoor Localization: A Transfer Learning Approach

Indoor localization (IL) systems predominantly depend on 1-D signal measurements, such as the received signal strength indication (RSSI) from Bluetooth or Wi-Fi access points (APs). Such methods, however, grapple with issues like interference from other APs and environmental challenges. To address these, this article introduces an innovative IL technique employing a classification system bolstered by transfer learning (TL). Instead of relying solely on 1-D signals, we transform them into images using techniques like spectrograms, scalograms, or Gramian angular fields. These transformed images feed into our classification system using a TL approach. We tested our method on two public datasets, achieving remarkable accuracy rates of 99% with the GoogleNet model and 98% with the SqueezeNet architecture. These figures underscore the efficacy of our technique for IL, marking a notable advancement over existing strategies.