Networking and Computing with Smart Surfaces

Smart Surfaces based wireless networking and computing (ScatterMIMO)

Increasingly wireless networks are densifying and distributing themselves with different technologies, with the key bottleneck as the deployment of access points/base-station. For example, to mmwave 5G will require 100x more base-station (gNB) deployment to achieve similar coverage to its sub-6 GHz counterpart owing to significantly higher penetration loss at mmwave. To make matters worse, each base station/access point deployment requires a fiber optical backhaul, which doesn’t exist today, making it near-impossible and cost-prohibitive.

My team has introduced new passive smart surfaces that can compute and shape the ambient wireless signals to overcome this challenge. Specifically, we built ScatterMIMO, a smart surface using tiny antennae to reflect signals back like a typical backscatter. My work shows that if you combine them, they could be as powerful as an access point in terms of reflected power. Furthermore, the exciting part is that it requires smart placement of smart surface and requires only 50 tiny antennas to re-radiate the same power as an AP. My work shows for the first time a smart surface can create distributed MIMO – which typically requires another WiFi access points with synchronization, active power and knowledge of data upfront – all of these are provided by the power of backscattering and low-complexity channel estimation and smart surface optimization – solving the fundamental problem with smart surfaces. Our work has shown that they can double the data-rate of communication by adding new MIMO spatial multiplexing that also reduces unwanted interference and works with COTS WiFi without requiring any protocol change, and costs just 5 dollars.

The paper creates a new direction in creating simple, smart surfaces that are scalable and improve the performance of a variety of applications, from wireless communication to building wireless sensing. It was covered in the news and media. In the next few years, we plan to build surfaces that can shape and compute on existing wireless signals to reduce and manage cross-technology interference while improving the data rate. Furthermore, we plan to scale the smart surfaces to get inputs from users, interact with users by touch or gestures in the air, all detected via the existing connectivity eco-system.