A Sequential Bayesian Algorithm for DOA Tracking in Time-Varying Environments

This paper focuses on the Direction of arrival (DOA) tracking problem in dynamic environments where each source signal is modeled as a Gaussian process with time-varying mean and unknown covariance. In the presence of highly dynamic environment, benchmark algorithms usually have deteriorated performance. By treating the source signals as a function of the arrival angles, a sequential Bayesian tracking approach named Simultaneous angle-source update (SASU) is proposed based on the Maximum a posteriori (MAP) principle. The key feature of the proposed approach is to simultaneously update the arrival angles and the source signals in the Kalman filter step by converting the update process of the state vector into a joint optimization problem. An iterative Newton method to efficiently solve the joint optimization problem is proposed. The accuracy and robustness of the proposed SASU algorithm is demonstrated via simulations.