An Improved Gray Wolf Algorithm for Radiation Near-Field Source Localization in Exact Model

Spectral search algorithms are widely used in radiation near-field (RNF) source locations because of their high estimation accuracy, but they have disadvantages such as large computation and long running time. To solve this problem, the Fresnel model is mainly adopted in the existing works of literature to realize the source location estimation in the algorithm of spectral search. Although the operation time is effectively reduced, there is a mismatch between the Fresnel model and the exact model, resulting in the existing method being not suitable for the source location under the exact model. In order to shorten the running time of the RNF source spectrum search algorithm under the exact model, an improved gray wolf optimal (IGWO) algorithm based on the exact model is proposed to estimate the RNF source position, which guarantees the high precision and shorten the operation time. The algorithm uses the noise subspace fitting method as the cost function and the power function as the convergence factor, which shortens the time required for source location and improves the estimation accuracy of the algorithm. The simulation results show that the IGWO algorithm takes less time and has better estimation accuracy than the existing source location algorithms.