Fast RCS computation in both spatial and frequency domains using adaptive MBPE technique

Sampling is one of the key techniques of the model based parameter estimation (MBPE). The ill-conditioning of the data matrix caused by the high polynomial order of the fitting-model occurs frequently. How to divide the oversized fitting-model into smaller ones is discussed and an adaptive sampling scheme is proposed in this paper. Since the equispaced approach is firstly applied and there is no need to add additional checking points, the searching time is greatly reduced. Numerical results have proved the validity and the efficiency of the method.