Error analysis for combined estimation and control systems with application to VTOL digital flight control synthesis

This paper considers the application of error analysis algorithms for the linear quadratic gaussian problem to the design of digital control systems for a simple model of a helicopter in hover. Particularly, we investigate the influence on the performance of modelling errors in the stability derivatives and correlation time of the wind gust model. The algorithms presented enable the determination of performance loss in the case of reduced order models. As an example of this possibility we will consider the case when colored system noise is misindentified as a gaussian white noise. The simulated results, for a helicopter in hover, given in the paper indicate that the extra effort expended in parameter identification may not be justified as the optimal trajectory and final performance is practically insensitive to errors in most of the model parameters.