Sensing Assisted Localization Services for Indoor Environments

Indoor localization is a critical component of various applications, including assisted living, personnel monitoring, and asset tracking. Traditional localization methods relying on specialized sensors such as LiDAR, ultrasound, and 3D cameras offer high precision but suffer from high costs and limited interoperability. To address these challenges, this paper explores the Integrated Sensing and Communication (ISAC) paradigm, leveraging signal-based modalities focusing on received signal strength indication (RSSI). These measurements, inherently embedded in wireless communication packets, enable cost-effective and vendor-agnostic localization without the need for additional hardware. However, practical deployment remains challenging due to signal degradation from obstacles, multipath effects, and reflections. This paper presents an end-to-end localization framework utilizing COTS IoT devices and advanced RSSI processing techniques to enhance measurement reliability. By integrating filtering mechanisms and machine learning models, the proposed solution improves distance estimation, categorizes line-of-sight (LoS) and non-line-of-sight (NLoS) conditions, and enhances localization accuracy. Additionally, an open and extendable edge-to-cloud infrastructure supports scalability and real-time processing. Experimental evaluation demonstrate the effectiveness of this approach in various aspects such as increase of RSSI reliability (increase of up to 82 % regarding standard deviation, drastic reduction of outliers' detection and fluctuation of more than 90 % to distances up to 2 m), accurate LoS-NLoS classification up to 99 % and overall localization increased accuracy more than 83 %.