Bayesian direction-of-arrival estimation for moving sound source in high-interference environments

Area of Science:

• Acoustics and Signal Processing
• Computational Mathematics
• Statistical Inference

Background:

• Direction-of-arrival (DOA) estimation is crucial for locating acoustic sources.
• High-interference environments degrade conventional DOA accuracy due to blurred angular discrimination.
• Existing methods struggle with accurately localizing moving sources amidst strong interference.

Purpose of the Study:

• To develop an accurate and robust DOA estimation method for moving acoustic sources in high-interference scenarios.
• To enhance localization precision by integrating prior target location knowledge.
• To address the limitations of conventional DOA techniques in complex acoustic environments.

Main Methods:

• A Bayesian DOA estimation approach is proposed, integrating tracker-derived prior knowledge into the Multiple Signal Classification (MUSIC) framework.
• The DOA problem is reformulated as a constrained array spatial spectrum optimization.
• A Bayesian MUSIC estimator utilizing a branch-and-bound algorithm is employed for precise DOA measurements.

Main Results:

• The proposed method achieves high-precision DOA measurements by optimizing the spatial spectrum.
• DOA measurements are effectively integrated into an Extended Kalman Filter (EKF) tracker for continuous localization.
• The Bayesian MUSIC estimator demonstrates superior accuracy and robustness compared to benchmark methods.

Conclusions:

• The developed Bayesian DOA estimation method significantly enhances localization accuracy for moving acoustic sources in high-interference environments.
• The integration of prior tracking information into the MUSIC framework provides a robust solution.
• The method offers a computationally efficient and high-performance alternative to existing techniques.