AI Enhances Indoor Positioning with RIS Intelligent Surfaces

Researchers in the UK have made significant strides in enhancing indoor positioning systems through the development of a machine learning AI technique. Engineers from the University of Glasgow and Australia have successfully utilized reconfigurable intelligent surfaces (RIS) in combination with AI algorithms to amplify signals from satellite navigation systems and cellular networks.

Traditionally, signals from these systems are often too weak to be reliably detected indoors. To address this challenge, the team at the University of Glasgow devised RIS sheets that were strategically placed on indoor walls and ceilings. These surfaces were designed to intercept wireless signals from external sources and intelligently reflect, redirect, and focus them as needed to enhance overall performance.

The implications of this innovative work are far-reaching, with potential applications ranging from aiding emergency services in locating individuals trapped in smoke-filled buildings to facilitating device-assisted navigation for visually impaired individuals in public spaces. Additionally, this technology could streamline the process of finding optimal spots for making mobile phone calls indoors.

During the experimental phase, the researchers set up a 1.3m-square RIS equipped with 4,096 passive elements within a designated space at the University of Glasgow. This setup was paired with two universal serial radio peripherals, with one serving as a signal receiver and the other as a transmitter.

In the initial phase of the experiment, the team focused on configuring the RIS to effectively reflect signals from the transmitter to the receiver by steering the beam across nine different positions and conducting signal tests at each location. Subsequently, various machine learning algorithms were employed to analyze the unique 'fingerprints' of the RIS-optimized wireless signals at each position to determine the most accurate algorithm for pinpointing the signals.

One algorithm emerged as the top performer, demonstrating an 82.4% accuracy rate in accurately identifying the receiver's location. Professor Qammer Abbasi from the University of Glasgow's James Watt School of Engineering highlighted the significance of this advancement in indoor positioning technology, emphasizing the potential of RIS to revolutionize location-finding indoors.