Accuracy and resolution of bearing measurements for crossed-dipole receiver

A well-known system for estimating the bearing to an underwater acoustic source consists of two crossed dipole sensors and an omni-directional sensor. At DRDC Atlantic, several bearing estimation methods for this three-channel system in the presence of ocean noise have been examined with the goal of minimizing the bearing error. Example algorithms are arctangent and maximum likelihood beamformer. These algorithms have trade-offs between estimation accuracy and computational efficiency, because any practical bearing-estimation algorithm must remain efficient enough for implementation in a real-time sonar processor. The accuracy of these algorithms can be demonstrated through comparison with the Cramer-Rao lower bound. Previously the calculation of bearing accuracy was based on the assumption of isotropic noise and a single signal source. However, the case of anisotropic noise and multiple-sources is more realistic, especially for bistatic or multistatic active sonar. Adaptive methods such as the Minimum Variance Distortionless Response (MVDR) beamformer will be explored to address this problem. The paper will discuss the potential improvement of bearing estimate accuracy and bearing resolution through numerical models.