Automotive Radar Interference Mitigation Using STFT Signal Reconstruction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing CS-based interference mitigation algorithms for automotive radar systems fail to achieve high-performance interference removal under high interference power and long interference duration conditions in real-time, degrading target detection performance and posing safety risks, especially for autonomous vehicles.
Innovation Solution
A real-time compressed sensing-based interference mitigation framework that exploits the sparsity of interference signals in the frequency-time domain using a GPU-implemented CoSaMP algorithm, transforming return signals to the frequency-time domain, reconstructing interference signals, and removing them to isolate the intended signal, with parallel computation to meet real-time requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing CS-based interference mitigation algorithms are used, then interference removal is attempted, but they fail to achieve high-performance interference removal under high interference power and long interference duration conditions in real-time
Solution Approach 1:
The patent segments the interference mitigation process into distinct stages: (1) transforming the return signal to frequency-time domain using short-time Fourier transform, (2) determining the reconstructed interference signal based on sparsity in frequency-time domain, (3) transforming the reconstructed interference signal to time domain, and (4) removing the interference signal from the return signal. This segmentation enables real-time processing by breaking down the complex interference removal task into manageable computational steps that can be executed efficiently.
Solution Approach 2:
The patent transforms the signal from the time domain to the frequency-time domain using short-time Fourier transform, where the interference signal exhibits sparsity. By operating in this transformed domain, the algorithm can effectively separate and remove interference that would be difficult to handle in the original time domain, achieving high-performance interference removal while maintaining real-time processing capability.
2Measurement precision
If interference mitigation is performed to isolate the intended signal, then target detection performance improves, but computation time increases
Solution Approach 1:
The patent changes the domain parameter from time domain to frequency-time domain by applying short-time Fourier transform. This parameter change enables the algorithm to exploit the sparsity of interference signals in the frequency-time domain, allowing for efficient interference removal that improves target detection performance while keeping computation time manageable for real-time applications.
Solution Approach 2:
The patent creates a reconstructed interference signal by copying and processing the interference components in the frequency-time domain. By reconstructing the interference signal and then subtracting it from the original return signal, the system isolates the intended signal with high precision, improving target detection while maintaining computational efficiency through the structured approach.
3Speed
If real-time processing is achieved through parallel computation, then processing speed improves, but system complexity increases
Solution Approach 1:
The patent employs periodic operations including the short-time Fourier transform and inverse transform that can be efficiently implemented using parallel computation. The periodic nature of these transformations allows for optimized parallel processing architectures that improve processing speed while managing system complexity through standardized, repeatable computational patterns.
Data Source
AI summary
Real-time vehicle radar interference mitigation concepts are described. In one embodiment, a method to mitigate interference for a vehicle radar includes detecting a target object in a vicinity of a vehicle. The detection of the target object generates an intended signal. The method further includes detecting an interference signal in the vicinity of the vehicle and receiving a return signal that includes the intended signal and the interference signal. The method further includes transforming the return signal to a frequency-time domain return signal using a short-time Fourier transform, determining a reconstructed interference signal in a frequency-time domain, transforming the reconstructed interference signal to a time domain, and removing the interference signal in the time domain from the return signal to isolate the intended signal based on the reconstructed interference signal in the time domain. The method further includes employing a parallel and computational efficient structure to deliver real-time output.