Different approach to robust automatic control for airplanes

Current automatic control system uses linear mathematical models to validate automatic flight control for airplanes. Gain scheduling, non linearity and improved feedback through simulation are also introduced. Very computers operate the actuators in order to keep the airplane on the right path, in the current trim and with the proper safety margin. Some engineers are testing fuzzy control logic to control airplanes and UAVs (Unmanned Aerial Vehicles). The result is brilliant, since very simple controllers are able to fulfill the specification with little "knowledge" about the airplane performances. This means that fuzzy controllers are very robust since they are able to operate with much degraded aerodynamics or with reduced thrust. However no one was able to validate the airplane/fuzzy controller with a mathematical proof. So it is not sure that it will works in any condition. By the way the same happens for the airplane/human pilot model. So a mathematical proof is still required also for this later solution. On the other side, very accurate, time based non linear mathematical models are available for flight simulation. These models are used in several fields ranging from development to training. In recent years computers that can run these accurate models in fractions of seconds were marketed at very low prices. The idea introduced in this paper is to run an accurate mathematical model on some of these fast autopilot computer in order to optimize the sequence of commands to be inputted to the FBW system of the airplane in order to keep the path in the safest way possible. For this purpose it is necessary to have enough computing power to calculate this best solution at a rate compatible to a correct control of the airplane. In this paper we will demonstrate that these computing resources are already available and it is predictable that the computing speed of future years will allow running even more sophisticated simulators. The question may be: why use more complicated systems when current control system fulfills satisfactorily the same task in a cheaper and more reliable way? The answers are several. At first it is a matter of robustness, what happens if the yaw damper fails or the actuator of the left ailerons is unable to fulfill its task or the tail is ripped off? In this case standard systems are not able to take the airplane to the ground safely even if it is indeed possible to control the airplane by a coordinate action of the remaining control surfaces. Optimization means that it is possible to reduce the stress on structures in order to improve aircraft life, to find the control sequence that assure the mean fuel consumption or to prefer the shortest time possible to reach the required trim on the right path. In other words it is more flexible. It is also possible to monitor aircraft performance in order to evaluate external or internal disturbances. Air turbulences, wind gusts may be controlled in order to optimize structural integrity or passenger comfort. Internal disturbances, as defective functioning of components or controls, occasional failure of sensors may be diagnosed, in some cases corrected in other simply reported after landing. The reliability improvement is not the latest benefit. As a rule of the thumb more electronics or more components means less reliability with the exception of redundancy and this is the case of this paper. © 2006-2015 Asian Research Publishing Network (ARPN).