Radar Super-Resolution with Virtual Antennas for Noisy Signals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Low-quality radar data, characterized by noise, interference, and low resolution, poses challenges for accurate object detection and classification, particularly in autonomous systems, due to limited antenna/receiver elements, leading to increased errors and missed detections.
Innovation Solution
A processing pipeline utilizing machine learning models to super-resolve radar data, including a first path that hallucinates additional inputs as if from virtual antennas and a second path that applies an encoder-decoder architecture with 3D convolutions to transform data into high-resolution radar data, enhancing the quality of existing radar sensors without adding hardware.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the number of antenna/receiver elements is increased to improve radar data quality and resolution, then measurement precision and reliability are improved, but device complexity and cost increase
Solution Approach 1:
The patent creates virtual copies of antenna/receiver elements through signal processing. The super-resolution model generates synthetic signals that mimic what additional physical antennas would produce, effectively copying the function of hardware elements without the physical cost. This allows the system to achieve higher resolution equivalent to having more antennas while maintaining the original hardware configuration.
Solution Approach 2:
The patent replaces the mechanical/hardware approach of adding more physical antennas with a computational/software-based super-resolution model. Instead of increasing device complexity through additional hardware elements, the system uses machine learning algorithms to process and enhance the signals from existing antennas, substituting computational complexity for hardware complexity.
2Reliability
If the number of antenna/receiver elements is increased to improve radar data quality, then reliability is improved, but cost increases
Solution Approach 1:
The super-resolution model generates virtual signals that replicate what additional physical antennas would capture, creating a cost-effective copy of the desired data quality. This allows the system to achieve reliable, high-quality radar data without the expense of purchasing and installing additional antenna/receiver hardware.
Solution Approach 2:
The patent changes the processing parameters and computational approach rather than hardware parameters. By applying super-resolution algorithms that transform the existing signal data into enhanced representations, the system achieves improved reliability and data quality through parameter transformation in the signal domain rather than through hardware upgrades.
3Measurement precision
If super-resolution processing is applied to enhance radar data, then measurement precision is improved, but processing time and computational complexity increase
Solution Approach 1:
The super-resolution model is trained in advance on large datasets to learn the transformations needed for super-resolution. Once trained, the model can quickly apply learned patterns to new radar data without performing computationally intensive calculations in real-time. This preliminary training phase separates the heavy computational work from the time-critical processing phase, reducing real-time processing time while maintaining high resolution output.
Data Source
AI summary
Systems, methods, and other embodiments described herein relate to improving radar data. In one embodiment, a method includes, responsive to acquiring radar data from a radar sensor, transforming the radar data into improved data according to a super-resolution model. The method includes converting the improved data into a range-azimuth-Doppler map. The method includes generating a high-resolution map from the range-azimuth-Doppler map by applying the super-resolution model to the range-azimuth-Doppler map. The method includes providing the high-resolution map.