On-Chip Active Resonator Based Spectrum Sensing

INV-22107

Background

Precise navigation and timing needs for both military and civilian applications are primarily enabled by GPS. While GPS can provide navigational accuracy down to a few meters, that accuracy degrades in a multi-path environment, yet various DoD applications involving drones and mobile platforms require navigational accuracy down to centimeters. Further, GPS signal is weak at the receiver and its carrier frequency is known, which makes it prone to jamming and spoofing attacks, which can disrupt the GPS based industry and pose a risk to our critical infrastructure. A potential solution for the continuous operation of critical civilian and military applications in GPS denied environments is the use of cognitive radios (CRs) for navigation and timing.

Cognitive radios detect unoccupied channels in a frequency band by sensing the spectrum for the absence of a primary user. However, traditional radio circuit designs are not well-suited for CRs because they rely on a frequency selective approach that is difficult to implement in CR applications. Additionally, this architecture is not suitable for precision navigation as the mismatch between clocks is not tolerable. To compensate for the phase offset and error between two clocks, the width of the received symbol and its phase relative to the local clock needs to be calculated, but this can be affected by fading and interference, resulting in timing errors.

Technology Overview

The proposed invention is an energy detection based cognitive communication where spectrum sensing is performed using a programmable, high-quality (Q), on-chip active resonator. The sensing system can sweep 1-20GHz frequency band in 1µs to identify narrow-band channels available for communication. Further, it uses the active resonator to enable transmission and reception of RF signal within a µs of identifying available RF channels. A machine learning based phase noise correction technique uses the envelop of the received signal to train against fading. The learned model can extract the feature map from the received envelope and predict the required phase correction within 10 clock cycles. The anticipated relative clock error is less than 10ps.

By enabling fast and precise two-way time transfer between two wireless nodes at any arbitrary time, the invention can significantly improve the current state of precision navigation. This feature can also enable navigational accuracy down to centimeters, as well as provide jamming and spoofing proof PNT infrastructure, while efficiently using spectrum in cognitive communication. The time transfer technique will also be useful for fast synchronization in next generation 5G communication applications.

Benefits

  • The invention has a low power, fast spectrum sensing technique.
  • A fast start-up cognitive radio communication system to communicate within 1µs of channel selection
  • A novel radio transmission technique using an on-chip resonator

Applications

  • Cognitive/Agile Communication
  • Low Power spectrum sensing

Opportunity

Seeking licensee and/or industry partner

Patent Information: