Method and Apparatus for Efficient Access Point Discovery Based on Learning CSI in Dense Wi-Fi Networks

INV-20082
 
Background
Consider a building where Access Points (APs) are deployed to provide WiFi connectivity. When a Wi-Fi device such as a smartphone, tablet, or laptop enters the building, before being able to communicate with the Internet, it needs to establish a connection with the AP. Legacy procedures require the client to scan the neighborhood in order to find a set of viable APs, before choosing the most appropriate AP to establish the connection. The process requires a handshake of several low data rate packets between the APs and the client, which wastes the spectrum resource. As Wi‑Fi technology progresses towards achieving gigabits of network throughput, several hundreds of APs are deployed to meet the bandwidth needs of end-users. With APs deployed at that scale, the legacy procedure of discovery requires too much time. The problem worsens when the client is mobile. With the proliferation of dense network deployments, there is a strong need for an intelligent AP that takes proactive actions to prevent unnecessary discovery delays and improve spectrum utilization by avoiding unneeded transmissions. 
 
Technology Overview
This invention introduces the CSIscan framework, a spectrum-efficient and low-latency Wi‑Fi access points (AP) discovery method for Wi‑Fi clients. CSIscan exploits the broadcast nature of Wi‑Fi channels by embedding discovery related information within an AP’s ongoing regular transmissions. The AP does this by intelligently distorting the transmitted OFDM frame by inducing perturbations in the preamble, and these injected ‘bits’ of information are detected via changes in the perceived channel state information (CSI). A deep learning framework allocates the optimal level of distortion on a per-subcarrier basis that keeps the resulting packet error rate to less than 1%. Existing clients perceive no changes in their ongoing communication, while potential new clients quickly obtain discovery information at the same time. 
A demonstrated result shows that CSIscan reduces the overall Wi‑Fi latency from 150 ms to 10 ms and improves spectrum utilization with ∼ 72% reduction in the probe traffic. It also shows that CSIscan delivers up to 40 discovery information bits in the outgoing Wi‑Fi packet in an indoor environment.
 
Benefits
- Achieves a 72% increase in AP spectrum efficiency 
- Significantly improves the speed of AP discovery at Wi‑Fi clients
- Achieves Wi‑Fi latency reduction from 150 ms to 10 ms
- Reduces power consumption at Wi‑Fi clients
- Extracts AP discovery information at the physical layer without any MAC layer processing 
- AI-powered algorithm determines an optimal selection of AP bounds for i) phase shifts ii) number of subcarriers
 
Applications
- Wi‑Fi networks for consumers and businesses, including enabling enhanced WiFi services for use cases in education, retail, travel and hospitality, healthcare, banking, financial services, insurance and Information Technology (IT) segments
- IoT systems
- Enhanced Wi‑Fi router for internet service providers such as Comcast, Verizon, etc.
 
Opportunity
- License
- Partnering
- Research collaboration
Patent Information:
For Information, Contact:
Mark Saulich
Associate Director of Commercialization
Northeastern
m.saulich@northeastern.edu
Patent #
Inventors:
Kaushik Chowdhury
Kunal Sankhe
Dheryta Jaisinghani
Links
Department
Copyright© Northeastern University. All rights reserved. Powered by Inteum.