Neural networks improve direction finding under real array conditions
Learning Array Signal Topologies as Conditional Neural Manifolds
Machine Learning
Summary
Direction of arrival (DoA) estimation helps devices figure out where signals come from using arrays of sensors. Traditional methods assume a fixed pattern for these arrays, but this can cause errors if the real environment is different. The authors propose a neural network approach that learns to adjust the array patterns based on the observed data, improving accuracy even with imperfections like noise or signal interference. This flexible method helps resolve tricky ambiguities that older techniques struggle with.
What this means in practice
- •For wireless communication engineers: Improve signal source localization accuracy in imperfect antenna arrays and noisy environments.
- •For audio signal processing teams: Enhance direction estimation of sound sources in complex acoustic settings with correlated noises.
Authors
Julian P. Merkofer, Vincent van de Schaft, Ruud J. G. van Sloun
Abstract
Subspace methods such as multiple signal classification (MUSIC) achieve super-resolution direction of arrival (DoA) estimation by exploiting the orthogonality between the array manifold and the noise subspace of the measurements. Their accuracy therefore depends on the assumed manifold and degrades under model mismatch, while parameters not identifiable from the spatial manifold cannot be recovered. In this work, we propose the conditional neural manifold (CNM), which replaces the fixed manifold with an observation-conditioned mapping from source parameters to steering vectors. An encoder maps the snapshots to a latent scene representation that conditions a zero-initialized neural field over the parameter space. The manifold is learned without steering-vector supervision by shaping the resulting MUSIC landscape. Since the correction acts on the manifold rather than on the estimator, it can be used by other manifold-based methods without modification. The CNM restores resolution under array imperfections, colored noise, correlated sources, and near-field propagation, and resolves the angle-frequency ambiguity inherent to the nominal spatial manifold.