A case study involving continuous system methods of inverse simulation for an unmanned aerial vehicle application

摘要

Inverse simulation allows input time histories to be found that generate specified outputs for non-linear dynamic models in cases where analytical methods of inversion present difficulties. The two approaches considered involve continuous system simulation principles. One is an approximate differentiation method while the second involves feedback principles. These approaches are compared for a non-linear six-degrees-of-freedom flight-mechanics model of a fixed-wing unmanned aerial vehicle incorporating actuator sub-models with saturation and rate limits. Additional insight is provided through analysis of a linearised version of the vehicle model. It is concluded that both the continuous system simulation methods for finding inverse solutions, for the type of application described in this paper, provide a useful alternative to more conventional iterative methods of inverse simulation based on discrete models. In many cases, including those involving hard non-linearities in control surface actuator sub-systems, they allow issues of vehicle handling and manoeuvrability to be addressed in a more direct fashion than is possible using conventional simulation methods alone.