Multiband 3D SAR Fusion Using Dual-Domain Complex CNNs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing multiband radar systems face challenges in achieving high-resolution near-field synthetic aperture radar imaging due to limited bandwidth and implementation issues with commercially available mmWave radars, leading to degraded image quality and inability to handle intricate targets.
Innovation Solution
A hybrid, dual-domain complex value convolutional neural network (CV-CNN) architecture, referred to as kR-Net, alternates between the wavenumber-domain and spatial-domain to fuse radar data, using a Fourier-based algorithm for improved 3-D SAR imaging.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If ultrawideband transceivers are used to achieve greater bandwidths, then image resolution is improved, but cost and device complexity increase significantly
Solution Approach 1:
The patent divides the ultrawideband spectrum into multiple subbands, each processed by a separate radar transceiver operating at a lower bandwidth. Instead of using a single complex ultrawideband transceiver, the system segments the bandwidth into manageable subbands (e.g., 60-64 GHz and 77-81 GHz) that can be handled by commercially available radars, thereby reducing device complexity while maintaining overall high resolution through multiband fusion
Solution Approach 2:
The patent employs a nested structure where multiple subband radars are integrated within a unified multiband fusion framework. The individual subband signals are processed and fused together using a hybrid dual-domain CV-CNN network, creating a nested architecture where simpler subband processors are contained within a more complex fusion system, achieving ultrawideband performance through composition of narrower band components
2Measurement precision
If several radars operate at distinct subbands to improve sensing resolution, then image resolution is improved, but implementation complexity and data fusion difficulty increase
Solution Approach 1:
The patent introduces a hybrid dual-domain complex-valued convolutional neural network (CV-CNN) as an intermediary to facilitate data fusion between multiple subband radars. This intermediary network operates in both spatial and wavenumber domains to effectively integrate signals from different frequency bands, simplifying the fusion process and reducing implementation complexity while maintaining high sensing resolution
Solution Approach 2:
The patent transforms the radar data between different domains (spatial domain and wavenumber domain) using parameter changes. By applying Fourier transforms and operating in multiple domains, the system changes the representation parameters of the signal to make the fusion process more effective and manageable, thereby reducing implementation complexity while achieving high resolution
3Ease of manufacture
If commercially available mmWave radars are used, then ease of manufacture and cost are improved, but bandwidth and image resolution are limited
Solution Approach 1:
The patent merges the capabilities of multiple commercially available mmWave radars operating at different subbands to achieve effective ultrawideband performance. By combining several narrower band radars (e.g., 60 GHz and 77 GHz bands) and fusing their data through the hybrid dual-domain CV-CNN, the system achieves the high resolution equivalent to a single ultrawideband radar while maintaining ease of manufacture and lower cost
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
kR-Net achieves high-fidelity multiband signal fusion, outperforming traditional methods by providing superior resolution and robustness for both low and high-bandwidth targets, even in complex scenarios.
Implementation Method 1
A synthetic aperture radar image is reconstructed from the fused first and second radar data according to Fourier-based algorithm
Data Source
AI summary
Multiband radar fusion is provided. The method comprises collecting first radar data from a first radar having a first frequency and collecting second radar data from a second radar having a second frequency different from the first frequency. A hybrid, dual-domain complex value convolutional neural network (CV-CNN), fuses the first and second radar data. The CV-CNN alternates between operating in the wavenumber-domain and the spatial-domain of the first and second radar data. A synthetic aperture radar image is reconstructed from the fused first and second radar data according to Fourier-based algorithm.