Machine Learning Direction of Arrival Estimation Non-Uniform Arrays
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining the direction of arrival (DoA) of electromagnetic energy, particularly in communications and radar systems, face inaccuracies when using non-uniform antenna arrays, leading to higher deployment costs and inefficiencies.
Innovation Solution
The use of machine-learning networks, such as Deep Neural Networks, trained to estimate the DoA by processing parameters of transmission signals and updating based on error terms, allowing for accurate identification even with non-uniform arrays, reducing the need for precise antenna array configurations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional methods are used to determine direction of arrival with non-uniform antenna arrays, then manufacturing precision requirements increase, but deployment cost increases and efficacy decreases
Solution Approach 1:
The patent changes the fundamental parameter of DoA estimation from traditional signal processing methods to machine learning-based estimation. The neural network is trained to directly estimate direction of arrival from antenna array outputs, bypassing the need for precise uniform array configurations. This parameter change allows non-uniform arrays to achieve accurate DoA estimation, resolving the contradiction between manufacturing precision requirements and deployment cost.
2Manufacturing precision
If traditional methods are used to determine direction of arrival with non-uniform antenna arrays, then manufacturing precision requirements increase, but measurement precision decreases
Solution Approach 1:
The patent replaces the mechanical/system-based traditional signal processing approach with a machine learning model. The neural network learns complex patterns from training data and directly predicts direction of arrival, substituting the need for precise mechanical array configurations with an intelligent algorithm that can handle non-uniform arrays while maintaining high measurement precision.
3Measurement precision
If machine learning networks are used to estimate direction of arrival, then measurement precision improves, but device complexity increases
Solution Approach 1:
The patent applies preliminary action by training the machine learning model offline before deployment. The neural network is trained on extensive datasets to learn the relationship between antenna array outputs and direction of arrival, then this pre-trained model is deployed for real-time estimation. This preliminary training phase separates the complex learning process from the operational phase, reducing real-time computational complexity while maintaining high measurement precision.
Data Source
AI summary
Methods, systems, and apparatus, including computer programs encoded on computer-storage media, for positioning a radio signal receiver at a first location within a three dimensional space; positioning a transmitter at a second location within the three dimensional space; transmitting a transmission signal from the transmitter to the radio signal receiver; processing, using a machine-learning network, one or more parameters of the transmission signal received at the radio signal receiver; in response to the processing, obtaining, from the machine-learning network, a prediction corresponding to a direction of arrival of the transmission signal transmitted by the transmitter; computing an error term by comparing the prediction to a set of ground truths; and updating the machine-learning network based on the error term.