Arbitrary Rotation of Spherical Coordinate for Anechoic Chamber Data Analysis

In this paper, a method that allows for arbitrary rotation of the spherical coordinate system is introduced. The rotation maintains the uniformly sampled data points. The rotation process is executed in the Cartesian coordinate system before being transformed back to the spherical coordinate. Subsequently, the transformed raw data is trained through a neural network with two inputs (azimuth and elevation angles) and one output (pattern magnitude). When uniformly sampled angles are fed into the trained network, it yields fitted data suitable for subsequent analysis. The method is demonstrated by rotating the scanning pattern of an antenna array multiple times, with satisfactory results.