Radar Image Generation via Compression Sensing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current radar technologies, such as the 2D MUSIC algorithm, face challenges with high computational complexity and noise, making them unsuitable for real-time high-resolution radar image generation, particularly in applications like autonomous driving.
Innovation Solution
The implementation of a compression sensing algorithm that generates a radar image by receiving signals from distributed radars, updating support vectors, and computing coefficients to produce high-resolution images with reduced computational requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the 2D MUSIC algorithm is used to acquire high-resolution radar images, then the image resolution is improved, but the computational complexity increases significantly
Solution Approach 1:
The patent transforms the radar image generation problem from the frequency domain to the time domain by changing the parameter representation. Instead of using frequency-domain spectral analysis (2D MUSIC), the invention uses time-domain signal processing with adaptive filtering and correlation operations, fundamentally changing the mathematical domain and computational approach while maintaining image resolution.
Solution Approach 2:
The patent replaces the complex iterative spectral estimation mechanism of the 2D MUSIC algorithm with a direct time-domain correlation and filtering mechanism. This substitution eliminates the need for eigenvalue decomposition and iterative optimization, replacing them with straightforward convolution and correlation operations that are computationally more efficient.
2Measurement precision
If the 2D MUSIC algorithm is used to acquire high-resolution radar images, then the image resolution is improved, but the processing time increases
Solution Approach 1:
The patent performs preliminary signal preprocessing and feature extraction in the time domain before image generation. By pre-processing the radar signals to extract target characteristics and organize data in time-domain format, the invention reduces the computational burden during the actual image generation phase, enabling faster real-time processing while maintaining high resolution.
Solution Approach 2:
The patent replaces the time-consuming iterative spectral estimation process of 2D MUSIC with direct time-domain correlation operations. This substitution eliminates multiple iterative cycles of eigenvalue decomposition and spectral peak searching, replacing them with single-pass correlation computations that execute significantly faster while producing equivalent or superior image resolution.
3Reliability
If the Nyquist sampling method is used to acquire radar signals, then the signal restoration is complete, but the sampling rate requires at least twice the frequency bandwidth
Solution Approach 1:
The patent applies compression sensing theory which allows sampling at a rate lower than the Nyquist rate by exploiting the sparsity of radar signals in certain domains. Instead of requiring complete signal restoration through high-rate sampling, the invention uses partial sampling combined with sparse reconstruction algorithms that can accurately recover the signal from fewer samples, reducing the sampling rate requirement while maintaining reliability.
4Area of stationary object
If distributed radars are used to receive signals, then the coverage area is improved, but the data processing complexity increases
Solution Approach 1:
The patent merges the signals and data from multiple distributed radars into a unified time-domain representation. By combining the received signals through coherent integration and correlation operations in the time domain, the invention processes multi-radar data as a single integrated dataset rather than separately processing each radar's output, thereby reducing overall processing complexity while maintaining expanded coverage capability.
Data Source
AI summary
Disclosed are a radar image generation method and an apparatus for performing the same. The radar image generation method includes receiving a received signal received at each of radars that are distributed and arranged; generating an input signal by processing the received signal; generating a support vector based on the input signal; updating the support vector; updating a coefficient corresponding to the support vector; and generating a radar image based on the support vector and the coefficient.