Output details

Non-parametric nonlinear system identification: A data-driven orthogonal basis function approach

The problem with identification using basis function is that sufficient a priori information on a unknown system must be available. This is not realistic in most applications. This paper presents a way to choose a basis function that is automatically determined by the data. In other words, basis choice and system identification are integrated parts of system identification. This eliminates the problem of blindly choosing a basis function and automatically finds the basis functions that match the system even though the system is unknown.