The idgrey model has parameters that are shared by both the measured (a,b,c,d) and noise (k) components.\nIn this case,

[fyousefi]

The idgrey model has parameters that are shared by both the measured (a,b,c,d) and noise (k) components.\nIn this case, using ''simulation'' focus or a weighting filter can lead to poor results.\nSome options for achieving reliable results are:\n 1. Ignore the estimation of K by using 'DisturbanceModel'/'none' as option-value pair in the estimation command.\n 2. Estimate K separately from the measured dynamics by using 'DisturbanceModel'/'estimate' as option-value pair in the estimation command.\n 3. If you must use the ODE file to parameterize the K matrix, make sure that the parameters used in the formula for K are different from those used for A, B, C and D matrices.