AI boosts RIS intelligent surfaces for indoor positioning
Researchers in the UK have developed a machine learning AI technique to boost indoor positioning systems.
Engineers from University of Glasgow and Australia have used reconfigurable intelligent surfaces (RIS) and AI algorithms to boost the signals from satellite navigation systems and cellular networks.
Usually these signals are too weak to be picked up indoors. The team at Glasgow developed sheets of RIS that were placed on walls and ceilings indoors to intercept wireless signals from outside and intelligently reflect, redirect and focus them as required to improve performance.
The work could have a wide range of future applications, from helping emergency services quickly find people trapped in smoke-filled buildings to offering device-assisted navigation through public spaces for blind and partially-sighted people.
It could also help finally eliminate the need to move around indoors to find the best spot to make a mobile phone call.
The researchers set up a 1.3m-square RIS containing 4,096 passive elements in a space at the University of Glasgow. They paired it with two universal serial radio peripherals where one acted as a receiver of wireless signals and the other as a transmitter.
In the first phase of the experiment, the researchers configured the ability of the RIS to reflect signals from the transmitter to the receiver effectively by steering the beam between nine different positions and sending test signals at each location.
In the second phase, the team used a series of different machine learning algorithms to analyse the unique ‘fingerprints’ of the RIS-optimised wireless signals at each location, testing which algorithm was capable of pinpointing the signals most accurately.
One algorithm clearly outperformed the others, proving itself capable of accurately determining the location of the receiver 82.4% of the time.
- €15m for the next generation of RIS reconfigurable sufraces
- First ETSI report on RIS Reconfigurable Intelligent Surfaces
- Trio conduct 6G reconfigurable intelligent surfaces trials
“While GPS works very well outdoors, helping us to use mapping apps to find our way efficiently on foot or in a vehicle, it works considerably less well in indoor environments. Positioning communication signals can be weakened by thick walls or interfered with by other electronic signals, reducing the accuracy of GPS,” said Professor Qammer Abbasi, of the University of Glasgow’s James Watt School of Engineering.
“RIS has the potential to greatly improve active location-finding indoors. It can do that by being aware of the communications signals being sent and received from devices like mobile phones at any given time, which allows them to precisely locate the device and its user.
“That opens up a wide range of possible applications, including tapping into RIS’ ability to focus signals to improve call quality by directing signals straight to mobile phones’ antennae, no matter where their user moves indoors.
“Our research is an important step forward in fine-tuning RIS technology’s ability to perform indoor localisation tasks in future comms networks.”
RIS is also a key technology for next generation 6G systems.
“This research shows that RIS can be used to shape and direct wireless signals in ways that have a lot of exciting future applications once RIS devices evolve and become more widely-adopted across communications networks,” said researcher Dr Syed Tariq Shah.
“The system we’ve prototyped could help with to develop improved management of crowds of people carrying mobile phones at large public events, or enable warehouse managers to keep better track of stock tagged with wireless transmitters. It could be applied to any situation where active positioning of signals from wireless devices is required.”
Researchers from the University of New England, Australia, also contributed to the paper, which is titled ‘Coded Environments: Data-Driven Indoor Localisation with Reconfigurable Intelligent Surfaces’ and published in Communications Engineering.
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
