Multi-Source DOA Estimation With Statistical Coverage Guarantees

Abstract

We consider uncertainty quantification (UQ) for multiple DOAs in an acoustic environment. The performance of DOA estimation is affected due to external uncertainty and yet the methods provide no UQ for the DOAs. Conformal prediction obtains statistically valid prediction intervals from an estimation model. First, we assume that a Gaussian mixture model can parameterize the conditional multi-source DOA distribution. Next, we demonstrate the effectiveness of conformal prediction to generate statistical uncertainty intervals from the mixture model outputs. The performance is validated on plane wave data with different sources of uncertainty using statistical metrics.