A fast near-field bearing estimation via spatial exponential kernel distribution

With its wide applications in various contexts of signal processing, the Direction of Arrival (DOA) of signal sources emerges to be a cascading issue. The majority of research, however, emphasizes far-field estimation instead of near-field one. In addition, almost all of the high-resolution DOA algorithms are based on the eigenstructure decomposition which calls for lots of computations. In this paper, we present a new algorithm by using the spatial Exponential Kernel Distribution (EKD) function in conjunction with a source parameter estimation procedure to evaluate the involved instantaneous parameters. Since the EKD function can suppress cross terms effectively, the new algorithm, as justified by simulations, can render precise estimates for near field, multi-component signal sources. Furthermore, it avoids the computationally intensive eigendecomposition and scanning scheme by using the simpler time-frequency distribution function. To further speed up the processing, the kernel can be initially evaluated and stored in a matrix form.