Back to results
US 11,742,883 B2Granted
Adaptive narrowband interference rejection for satellite navigation receiver
AssigneeDeere & Company
InventorsWei Yu, Mark P. Kaplan, Richard G. Keegan, David M. Li
Life of this patent
Estimating the term…
Abstract
A controller is configured to control the adaptive notch filter and to execute a search technique (e.g., artificial intelligence (AI) search technique) to converge on filter coefficients and to recursively adjust the filter coefficients of the adaptive notch filter in real time to adaptively adjust one or more filter characteristics (e.g., maximum notch depth or attenuation, bandwidth of notch, or general magnitude versus frequency response of notch).
Claims19
A receiver system with interference rejection, the receiver system comprising:
- an antenna for receiving a radio frequency signal;
- a downconverter for converting the radio frequency signal to an intermediate frequency signal;
- an analog-to-digital converter for converting the intermediate frequency signal or an analog baseband signal to a digital baseband signal;
- a selective filtering module for filtering of the digital baseband signal consistent with a target receive bandwidth;
- a narrow band rejection filter for rejecting an interference component, wherein the selective filtering module comprises:
- an adaptive notch filter supporting an infinite impulse response mode;
- and
- a controller for controlling the adaptive notch filter, the controller configured to execute a Steiglitz-McBride (SM)-modeled estimator or a least mean squares algorithm estimator to estimate filter coefficients of the adaptive notch filter, or the controller configured to execute cumulatively both the Steiglitz-McBride (SM)-modeled estimator and the least mean squares (LMS) algorithm estimator, the SM-modeled estimator or least mean squares algorithm estimator being storable in a data storage device and the controller configured to recursively adjust the filter coefficients of the adaptive notch based on the estimated filter coefficients;
- wherein the SM-modeled estimator or the LMS estimator is configured to conduct a mini batch gradient decent search, wherein the SM-modeled estimator or LMS estimator or a filter coefficient updater is configured to avoid a covariance matrix update/propagation by disregarding correlation between two NBI components for a scenario or case wherein there are two simultaneous NBI components that interfere with the received radio frequency signal or its related respective baseband signal that is derived from the received radio frequency signal.
6 more elements
A receiver system with interference rejection, the receiver system comprising:
- an antenna for receiving a radio frequency signal;
- a downconverter for converting the radio frequency signal to an intermediate frequency signal;
- an analog-to-digital converter for converting the intermediate frequency signal or an analog baseband signal to a digital baseband signal;
- a selective filtering module for filtering of the digital baseband signal consistent with a target receive bandwidth;
- a narrow band rejection filter for rejecting an interference component, wherein the selective filtering module comprises:
- an adaptive notch filter supporting an infinite impulse response mode;
- and
- a controller for controlling the adaptive notch filter, the controller configured to execute a Steiglitz-McBride (SM)-modeled estimator or a least mean squares algorithm estimator to estimate filter coefficients of the adaptive notch filter, or the controller configured to execute cumulatively both the Steiglitz-McBride (SM)-modeled estimator and the least mean squares (LMS) algorithm estimator, the SM-modeled estimator or least mean squares algorithm estimator being storable in a data storage device and the controller configured to recursively adjust the filter coefficients of the adaptive notch based on the estimated filter coefficients;
- wherein the notch filter is physically realized as an infinite impulse response filter that uses an artificial intelligence (AI) model to converge on filter coefficients more quickly than a least minimum squares technique, wherein the filter coefficients are determined based on a mini-batch stochastic gradient descent search.
6 more elements
Dependencies are inferred from the claim text.