VTIL: A multi-layer indoor location algorithm for RSSI images based on vision transformer

Abstract

As WiFi technology becomes more widespread and integrated, precise location tracking within complex indoor environments is gaining significance in contemporary public spaces.Nevertheless, challenges persist in indoor scenarios, including issues such as noise and multi-path effects stemming from the degradation of the Received Signal Strength Indicator (RSSI). Additionally, obstacles causing signal shading contribute to a decrease in the accuracy of indoor location tracking. Within this paper, we introduce an innovative indoor localization algorithm termed Vision Transformer Indoor Localization (VTIL). This algorithmleverages RSSI and Vision-Transformer (ViT) technologies to enhance indoor positioning accuracy. Initially, the RSSI fingerprints undergo normalization by scaling them to the maximum and minimum values. To mitigate the impact of noise and irrelevant features, the Principal Component Analysis (PCA) algorithmis then employed for effective feature extraction.Secondly, the RSSI fingerprint library is converted into an RSSI gray image library. Then the RSSI gray image is divided into several small blocks, and the position-coding input is performed in the form of a sequence. The ViT model divides the weight ratio of each block in the RSSI image to alleviate the impact of multi-path effects. According to the experimental results using public datasets, our approach achieves a noteworthy 37.26% reduction in the average distance estimation error when compared to existing indoor fingerprint localization algorithms.