| The challenges which face a modern company, large or small, are to improve product quality and production efficiency, ensuring safety and environmental responsibility while at the same time minimizing its costs. The key contributor for achieving these goals is the system automation on which customer satisfaction and company's competitive advantage so much depend. The aim of this research was to show how Neural Controllers, when implemented in real-time to control a plant in closed loop system, provided the required control signal by updating their weights automatically. An adaptive neural network with two layers was used to control the PMDC motor. Brandt-Lin adaptation learning algorithm was used to train the neural network. This algorithm reduced the design complexity as it did not require the neuron model of the plant. Simulation and real-time implementation results were provided to prove the effectiveness of neural controllers in "hybrid" control system. |
