Adaptive cross approximation algorithm accelerated inverse equivalent current method for near-field antenna measurement

The adaptive cross approximation algorithm (ACA) is adopted to efficiently compute the discretized forward integral equation operator in the NFFT by utilizing the rank-deficient nature of the well-separated interactions between the equivalent sources and the sampling probes. The threshold of the ACA compression is carefully investigated in order to achieve acceptable accuracy of the NFFT. An octree-based geometrical-adaptive grouping scheme is also proposed where two different grouping strategies are designed for the planar and spherical measurements respectively. Besides, considering the different spatial distribution density of the equivalent sources and the sampling probes, the group sizes for the sources and probes are different. As a result, the numerical accuracy and computational efficiency are significantly improved.