Machine Learning Radar Angular Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional radar systems face challenges in angular estimation due to size constraints in consumer devices, leading to angular ambiguities and reduced operational effectiveness, especially when antenna element spacings deviate from optimal wavelengths, limiting the radar's ability to accurately determine object positions.
Innovation Solution
A smart-device-based radar system employs machine learning, specifically an angle-estimation module using neural networks or convolutional neural networks, to generate angular probability data based on the radar system's spatial response across a wide field of view, resolving ambiguities and accurately identifying object positions regardless of antenna element spacing or wavelength.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Volume of moving object
If fewer antenna elements and larger or smaller antenna element spacings are used to satisfy size or layout constraints, then the radar can be integrated in consumer devices, but angular ambiguities are caused making it challenging to estimate angular position
Solution Approach 1:
The patent replaces conventional mechanical signal processing methods with machine learning algorithms to resolve angular ambiguities. The angle-estimation module uses trained neural networks to analyze radar returns and determine angular positions, substituting traditional computational approaches with AI-based solutions that can handle ambiguous spatial responses from constrained antenna configurations.
Solution Approach 2:
The patent employs parameter changes by training machine learning models on diverse radar return patterns corresponding to different angular positions and antenna configurations. The system adapts its estimation parameters based on learned patterns from training data, enabling accurate angular estimation despite variations in antenna element spacing and device constraints.
2Length of moving object
If antenna element spacings deviate from optimal wavelengths, then the radar can be miniaturized for consumer devices, but the ability to accurately determine object positions is reduced
Solution Approach 1:
The patent replaces traditional spectral analysis and Fourier transform methods with machine learning-based angle estimation. The neural network models are trained to recognize angular signatures from radar returns generated by non-optimal antenna spacings, substituting conventional signal processing with AI methods that can accurately interpret distorted spatial responses.
Solution Approach 2:
The patent introduces machine learning models as an intermediary between the radar hardware and position estimation. The trained models act as mediators that translate the distorted radar returns from non-optimal antenna configurations into accurate angular position estimates, bridging the gap between hardware constraints and measurement accuracy requirements.
Data Source
AI summary
Techniques and apparatuses are described that implement a smart-device-based radar system capable of performing angular estimation using machine learning. In particular, a radar system 102 includes an angle-estimation module 504 that employs machine learning to estimate an angular position of one or more objects (e.g., users). By analyzing an irregular shape of the radar system 102's spatial response across a wide field of view, the angle-estimation module 504 can resolve angular ambiguities that may be present based on the angle to the object or based on a design of the radar system 102 to correctly identify the angular position of the object. Using machine-learning techniques, the radar system 102 can achieve a high probability of detection and a low false-alarm rate for a variety of different antenna element spacings and frequencies.