Radar Target Classification Using an RCS-Scintillation CNN
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current automotive RADAR sensors face challenges in urban environments due to low reflectivity, overlapping radar cross section (RCS) values, velocity artifacts, and high sensor floor noise levels, which complicate the classification of target signatures.
Innovation Solution
A scintillation-based convolutional neural network (CNN) is employed to classify RADAR targets using RCS scintillation values and velocity metrics, utilizing a Neyman-Pearson criteria for enhanced discrimination and resilience against motion artifacts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional RADAR classification methods are used in urban environments, then the system can operate in cities, but the classification accuracy deteriorates due to low reflectivity, overlapping RCS values, velocity artifacts, and high sensor floor noise
Solution Approach 1:
The patent segments the target classification problem into multiple stages: initial detection, RCS scintillation analysis, velocity artifact filtering, and final classification. By dividing the complex classification task into discrete processing stages, the system can address each harmful factor (clutter, noise, velocity artifacts) at appropriate stages, improving overall classification accuracy in urban environments
Solution Approach 2:
The patent introduces RCS scintillation measurements as an intermediary feature between raw RADAR data and final classification. These scintillation values serve as a mediator that captures target characteristics while being resistant to velocity artifacts and environmental clutter, enabling more accurate classification despite harmful urban factors
2Reliability
If the RADAR sensor attempts to detect all targets in urban environments, then coverage is comprehensive, but false positives increase due to environmental clutter and low reflectivity objects
Solution Approach 1:
The patent performs preliminary RCS scintillation analysis and velocity artifact filtering before final target classification. By preprocessing the data to eliminate velocity artifacts and identify scintillation patterns early in the processing chain, the system reduces false positives from environmental clutter before classification decisions are made
Solution Approach 2:
The patent changes the parameter space by introducing RCS scintillation measurements alongside traditional RCS values and velocity data. This parameter transformation allows the system to distinguish true targets from clutter by analyzing scintillation patterns, improving reliability while reducing false positives from environmental factors
Data Source
AI summary
Disclosed are systems and methods for scintillation-based neural network for RADAR target classification. In some aspects, a method includes generating a point cloud from radio frequency (RF) scene responses received from a radio detection and ranging (RADAR) sensor for a scanned scene; populating a rolling buffer with frame data from the point cloud, the frame data including radar cross section (RCS) values, RCS scintillation measurements, and velocity values for objects in the scanned scene; inputting the RCS scintillation measurements and velocity values for an object of the objects to a convolutional neural network (CNN); and receiving a classification of the object from the CNN, wherein the CNN is to utilize a probability density function (PDF) estimate of the RCS scintillation measurements and the velocity values to determine fits with one or more reference PDFs based on a Neyman-Pearson evaluation, and wherein the fits are assessed to classify the object.