Blind Measurement of Receiver System Noise.

Area of Science:

• Electrical Engineering
• Communications Engineering
• Metrology

Background:

• Tightly packaged receivers present challenges for direct noise measurements.
• Receiver outputs (user data) are often indirect and processed, hindering validation.
• Existing methods rely on models and standards, lacking physical measurement validation.

Purpose of the Study:

• To develop and validate novel measurement techniques for receiver noise figure (NF).
• To complement and support existing system noise models with physical measurements.
• To enable accurate NF estimation from receiver user data, independent of protocol.

Main Methods:

• Developed a semiparametric model for full-stack receiver noise response.
• Employed systematic perturbation of input signals and excess noise levels.
• Introduced novel test capabilities: programmable excess noise source and attenuator.
• Utilized a regression procedure to estimate NF from user data and input conditions.
• Combined traditional and Monte Carlo methods for uncertainty estimation.

Main Results:

• Demonstrated a novel technique for measuring receiver noise figure (NF) using user data.
• Successfully implemented precision programmable excess noise source and directive programmable attenuator.
• Validated the technique with consumer wireless networking and geolocation equipment.
• Achieved verification through repeatability testing and comparison with Y-factor measurements.

Conclusions:

• The proposed measurement techniques provide a means to physically validate receiver noise performance.
• The blind approach is robust to protocol and implementation details.
• Accurate noise figure estimation and uncertainty quantification are achievable.
• The methods are applicable to diverse communication equipment.