Differential flatness applied to vehicle trajectory tracking

Differential flatness, a property of some dynamic systems which has been recognized only recently, has made possible the development of new tools to control complex nonlinear dynamic systems. Guidance dynamics of many different systems have been recognized as being explicitly or implicitly differentially flat as it is the case for flight guidance dynamics of conventional aircraft. In this paper, a new control structure is proposed to achieve trajectory tracking for vehicles: a neural network deals with the inversion of nominal guidance dynamics while a linear corrector cope with modelling errors and perturbations so that control directives to conventional autopilot systems can be generated in real time. The proposed approach is illustrated in the case of conventional aircraft flight dynamics.