DSP BASED HARDWARE EMULATOR FOR COMPLEX-ORDER FRACTIONAL DYNAMICAL SYSTEM USING ARTIFICIAL NEURAL NETWORK APPROXIMATION TECHNIQUE.

摘要

Recently, many real-world and man-made systems have been found to be more accurately modelled using the complex-order derivatives. Complex-order derivatives, the derivatives with order as complex numbers, are the extensions of integer-order and fractional-order derivatives. The topic of using these derivatives in modelling and control is very recent. There is some theoretical work reported in this field. The main problem in the hardware implementation of these complex-order models (systems and controllers) is their infinite memory. Because of this issue, the real-time implementation of complex-order systems is not much attempted. The available limited memory integer-order approximations sometimes result into ill-conditioned systems and become numerically unstable. This invention is a DSP based emulator for hardware implementation of complex-order dynamical systems. The setup uses artificial neural network (ANN) for approximation of complex-order derivative operators. The use of ANN makes the resulting approximation very simple and thus the hardware implementation becomes very efficient and stable. The designed system is reconfigurable and can be easily interfacing facility with a computer. The proposed system is capable of Hardware in Loop (HIL) Design. The invention will provide an easy experimentation facility for researchers working on complex-order systems and control.