Comparative analisys of neural control systems within the NASA IFCS F-15 WVU simulator

This paper describes the results of a study comparing the performance of three different approaches for the fault tolerant flight control system of the F-15 aircraft within the Gen 2 NASA IFCS project.The common feature of the 3 methods is that they are all based on neural algorithms, however neural networks (NN) are used in different ways.The first approach features a Non Linear Dynamic Inversion (NLDI) technique augmented with a neural network to cancel the dynamic inversion errors.The second approach is based on a robust controller using both the Stochastic Optimal Feedforward and Feedback Technique (SOFFT) and a neural network to compensate errors.Finally, the third approach performs a neural network controller based on an adaptive predictor-corrector control strategy.The comparison study has been carried out using the WVU IFCS F-15 simulation environment.The results are given in terms of performance comparison and control activity evaluation for different maneuvers.

Hali tarjima qilinmagan