The paper introduces a control technique for autonomous formation flight that does not require inter-aircraft distance measures to maintain formation stiffness. In a two aircraft formation, the leader induces a wake on its wingman and a neural network structure is synthesized to estimate the relative position. The possibility to derive wake effects estimates in order to train a neural network to recognize them, is described both with inverse modeling and asymptotic state estimation. The effect of aircraft dynamics modeling errors on neural network training have been also investigated. Finally, computer simulations are shown that describe the formation control results achieved with this technique.
Sensorless Formation Flight
POLLINI, LORENZO;INNOCENTI, MARIO
2001-01-01
Abstract
The paper introduces a control technique for autonomous formation flight that does not require inter-aircraft distance measures to maintain formation stiffness. In a two aircraft formation, the leader induces a wake on its wingman and a neural network structure is synthesized to estimate the relative position. The possibility to derive wake effects estimates in order to train a neural network to recognize them, is described both with inverse modeling and asymptotic state estimation. The effect of aircraft dynamics modeling errors on neural network training have been also investigated. Finally, computer simulations are shown that describe the formation control results achieved with this technique.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
