Radar RFI Suppression via CNN-Based Reference Antenna Subtraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional RFI suppression methods in radar systems, such as temporal blanking and spatial filtering, are ineffective for nonstationary interference and do not utilize prior information or physics, leading to incomplete RFI removal, especially in complex electromagnetic environments.
Innovation Solution
A deep learning approach using reference antennas to predict and remove RFI signals by establishing non-linear mappings between primary and reference antennas, employing a convolutional neural network to subtract predicted RFI from primary antenna signals, thereby enhancing radar signal quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional temporal blanking or spatial filtering methods are used for RFI suppression, then the system structure remains simple, but the RFI suppression effectiveness is insufficient especially for nonstationary interference
Solution Approach 1:
The patent introduces reference antennas as intermediary elements that specifically detect RFI signals without receiving signals of interest. These reference antennas serve as mediators between the interference sources and the primary antennas, enabling the deep learning model to learn RFI patterns separately and subtract them from the primary antenna signals, thereby achieving effective RFI suppression while preserving target information
Solution Approach 2:
The patent replaces traditional mechanical signal processing methods (temporal blanking, spatial filtering) with a deep learning-based neural network system. The neural network automatically learns non-linear mappings between reference and primary antenna signals, substituting manual thresholding and subspace projection with adaptive, data-driven processing that achieves superior RFI suppression performance
2Reliability
If traditional blind signal processing methods are used, then the processing approach is simple, but prior information and physics are not utilized leading to incomplete RFI removal
Solution Approach 1:
The patent performs preliminary action by training the deep learning model using RFI data collected in the absence of signals of interest. This pre-training phase allows the system to learn RFI characteristics and patterns before actual radar operation, enabling the model to predict and remove RFI more effectively during signal acquisition without interfering with target detection
Solution Approach 2:
The patent implements feedback by using the trained deep learning model to continuously predict RFI signals based on reference antenna inputs and subtract them from primary antenna signals. The system effectively uses the learned RFI patterns as feedback to cancel interference in real-time, improving the completeness of RFI removal while preserving useful signal information
Data Source
AI summary
A system for suppressing radio frequency interference (RFI) or electromagnetic interference (EMI) in a radar or magnetic resonance imaging system that scans target(s) with energy and receives signals that reflect or echo from the target(s). The system involves obtaining RFI/EMI data in the absence of signals of interest in the scanning system using at least one primary antenna and a plurality of reference antennas. The reference antennas are designed and arranged to detect RFI/EMI signals but not signals of interest in the scanning system. Simultaneously a model, e.g. CNN, is trained with the RFI/EMI data to determine the non-linear signal mappings among primary and reference antennas. The trained model is applied to predict the RFI/EMI received by the primary antenna in the presence of signals of interest. Finally, the RFI/EMI signals received by the primary antenna are removed by subtracting out the predicted RFI/EMI.