Identification of nonlinear dynamic systems using functional link artificial neural networks

Abstract

We have presented an alternate ANN structure called functional link ANN (FLANN) for nonlinear dynamic system identification using the popular backpropagation algorithm. In contrast to a feedforward ANN structure, i.e., a multilayer perceptron (MLP), the FLANN is basically a single layer structure in which nonlinearity is introduced by enhancing the input pattern with nonlinear functional expansion. With proper choice of functional expansion in a FLANN, this network performs as good as and in some cases even better than the MLP structure for the problem of nonlinear system identification.

A flann-based controller for maximum power point tracking in PV systems under rapidly changing conditions

“…Due to the absence of hidden layer in FLANN, it is computationally inexpensive. The characteristic of less computational complexity by FLANN than that by MLP can also be found in [13].…”

“…This experiment compares the computational complexities of FLANN [13] and MLP [4]. The computational complexity is reflected by calculating the average execution time of one iteration during training process.…”

“…The number of additions, multiplications, tanh(.) and trigonometric expansions can also be evaluated as in [13]. Here, we consider a two-layer MLP with {2 -8 -1} and a FLANN with {16 -1} architecture.…”

“…Irradiance and temperature are the inputs of the PV arrays. The functional expansion block is used to transfer the input variable into a higher dimension space [13]. The expanded inputs are connected to the output layer, which is activated by a hyperbolic tangent transfer function.…”

Multivariable Self-Tuning Feedback Linearization Controller for Power Oscillation Damping

“…Mostly, various neural networks models are used by researchers [14]- [19] for the accurate representation of the nonlinear systems which has the advantage of its generalization and learning ability [18]. In recent years, use of multi-layer perceptron (MLP), radial basis function (RBF), recurrent and simultaneous recurrent neural network (RNN and SRN) has been reported for online estimation of input−output mapping of nonlinear systems [18], [20]- [22]. These methods typically use back-propagation (BP) or back-propagation through time (BPTT) to update online the neural network parameters.…”