DOA Estimation with Lightweight Network on LLM-Aided Simulated Acoustic Scenes

2025-11-11

AI Summary

Existing DOA estimation models are predominantly trained on synthetic room impulse response (RIR) data, suffering from limited acoustic diversity and consequently poor generalization to real-world environments. To address this, we propose LightDOA—a lightweight, deep separable convolution-based DOA estimator supporting multi-channel input with low computational overhead. Crucially, we introduce, for the first time, large language model (LLM)-assisted generation of diverse spatial audio scenes, significantly enhancing acoustic coverage and realism of training data. Extensive experiments across varying reverberation conditions, noise types, and microphone array configurations demonstrate that LightDOA achieves high estimation accuracy while exhibiting strong robustness. Moreover, it attains 3–5× higher inference efficiency compared to state-of-the-art models, making it particularly suitable for resource-constrained edge devices.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Intelligent Robots: State Estimation
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Search and Retrieval-Augmented AI: Large language models for search
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Direction-of-Arrival (DOA) estimation is critical in spatial audio and acoustic signal processing, with wide-ranging applications in real-world. Most existing DOA models are trained on synthetic data by convolving clean speech with room impulse responses (RIRs), which limits their generalizability due to constrained acoustic diversity. In this paper, we revisit DOA estimation using a recently introduced dataset constructed with the assistance of large language models (LLMs), which provides more realistic and diverse spatial audio scenes. We benchmark several representative neural-based DOA methods on this dataset and propose LightDOA, a lightweight DOA estimation model based on depthwise separable convolutions, specifically designed for mutil-channel input in varying environments. Experimental results show that LightDOA achieves satisfactory accuracy and robustness across various acoustic scenes while maintaining low computational complexity. This study not only highlights the potential of spatial audio synthesized with the assistance of LLMs in advancing robust and efficient DOA estimation research, but also highlights LightDOA as efficient solution for resource-constrained applications.

Problem

• Improving DOA estimation generalizability with LLM-aided diverse acoustic scenes
• Developing lightweight model for robust DOA estimation in varying environments
• Addressing computational complexity limitations in multi-channel DOA estimation systems

Innovation

• LLM-aided simulated acoustic scenes generation
• Lightweight DOA model with depthwise separable convolutions
• Multi-channel input processing for varying environments