A WiFi Indoor Localization System Based on Robot Data Acquisition and Deep Learning Model

Abstract

With the proliferation of smartphones and the rapid development of the Internet of Things (IoT), Location-Based Services (LBS) have garnered increased attention, particularly in the context of Indoor Positioning Systems (IPS). Mainstream positioning technologies perform poorly indoors due to obstructions, necessitating specialized IPS. A significant advantage of WiFi-based positioning is its ability to operate without additional hardware, leveraging existing mobile devices. However, WiFi-based positioning faces challenges, including the labor-intensive and costly process of manual fingerprint data collection and the limitations of traditional feature extraction methods in meeting accuracy demands. This paper proposes a novel deep learning-based indoor positioning method utilizing robotic WiFi fingerprint data collection. By employing a robotic platform for automated data collection and a Deep Long ShortTerm Memory (DLSTM) model to extract complex temporal features of WiFi data, it enhances both the efficiency and accuracy of indoor positioning. Experimental results demonstrate significant improvements in positioning accuracy compared to traditional methods.