Repetitive Neurocontroller with Disturbance Feedforward

This project demonstrates repetitve process control using gradient-based dynamic optimization tool.

Repetitive neurocontroller (RNC) with disturbance feedforward path active in the pass-to-pass direction (kDFF) represents a novel (2014 as far as kDFF is concerned) approach to repetitive process control. The resulting control scheme is of Disturbance Dual Feed-Forward (DDFF) type. The solution is of model-free type, i.e. no accompanying neural network is needed to model dynamics of the plant. Please see inside the m-files for more information. This submission enables you to play with different system configurations, i.e. several flags are provided to easily reconfigure the system. The solution was inspired by the concept of iterative learning control (ILC). This project might be of your interest if you deal with: repetitive process control, iterative learning control, dynamic optimization problems, and neurocontrollers in the form of online trained neural networks. Such control tasks are often encountered in robotics and power electronics. Refs to relevant papers are included. If something brought you here, it could be probably also interesting to have a look at my gradient-free swarm-based repetitive controller (PDPSRC or PDMSRC on Matlab Central). More info can be found in "Repetitive neurocontroller with disturbance feedforward path active in the pass-to-pass direction for a VSI inverter with an output LC filter" (http://dx.doi.org/10.1515/bpasts-2016-0013).