Designing Autolanding Neuro-Controller Using PID and Optimal Strategies

Designing a Neuro Controller for longitudinal autolanding system of a commercial jet transport has been considered. To train the neuro controller there are so many strategies in selecting the training data. In this paper, first a PID and an optimal controller for autolanding system have been designed. Then, the outputs of these two classic controllers have been used to train the neuro controller separately. Furthermore, the robustness of the controllers has been investigated by applying the gust and changing the flight conditions. Other advantages and disadvantages of the controllers have also been discussed. Simulation results show that PID controllers due to their robustness and Optimal controllers due to their performance are good candidates to train the neuro-controller