Novel Multi-Axis Cross-Correlation Algorithm for Precise Azimuth Angle Estimation in MEMS Underwater Acoustic Vector Sensors

Resumen

Underwater detection of acoustic sources, tracking movement of targets, and acquiring directional information requires an accurate estimation of the azimuth angle (i.e. angle in the horizontal plane). A new algorithm for use with cross-shaped piezoelectric MEMS acoustic vector sensors is described here. This new Multi-Axis Cross-Correlation (MACC) algorithm takes advantage of the orthogonal cantilever design to permit the simultaneous measurement of the particle velocity components along multiple axes, thereby obtaining a very accurate estimation of the direction-of-arrival (DOA) without requiring large sensor arrays. The MACC algorithm achieves this by combining two techniques: cross-correlated analysis of time series data with the use of phase differences between each component and maximum/minimum amplitude ratios. The resulting azimuth estimation is accurate to within ±2°, across the entire 360° of azimuth, over a frequency range of 20 kHz to 200 kHz. Comparison of MACC to conventional beam forming and MUSIC algorithms through extensive simulation and analysis has shown that MACC achieves a 65% reduction in estimation error and 40% increase in angular resolution compared to these methods. Additionally, because the computational complexity of MACC is low; it is well suited for near real-time implementation using a limited-resource MEMS platform.