Application of Tensor Decomposition Methods to Antenna Array Calibration Measurements for De-Noising and Narrowband Modelling

Narrowband modelling of an antenna array's response is often considered for parameter estimation or modelling purposes. Because the narrowband assumption has to be carefully assigned, a new criterion to justify the narrowband assumption based on array calibration data will be presented. Subsequently, tensor decomposition methods are presented, which are applicable to estimate a narrowband array model from array calibration measurements, while accounting for various assumptions regarding the calibration measurement system or the antenna array itself. The performance of the proposed decomposition methods will be compared in Monte-Carlo simulations using synthetic as well as measurement data.