A validation methodology for neural network based flight control systems

A significant problem associated the inclusion of a neural network in an avionics system is validating that system to the required reliability level. To accomplish this, it is necessary to associate a “probability of failure” with the neural network and, ultimately, with the operational flight program. It would be more correct to say that the probability of excitation of the network in an unvalidated portion of its input space is required. In this sense, estimation errors associated with neural networks are like latent hardware faults, and techniques that were previously used to measure the probability of failure of hardware due to fault latency can be used to measure the probability of failure of the network. A methodology was developed and applied to a flight controller designed to operate in a well defined environment. The controller incorporated a network to estimate nonlinear portions of plant performance. The results of the study indicates that the technique could be used to provide a final validation of the network and the controller to a specified reliability level and to evaluate the role of flight test in network validation