Recent work has demonstrated the use of model predictive control (MPC) for stabilizing the Caltech ducted fan. In results published up until now, however, additional stability augmentation was required. It was felt that an inadequate model was the cause of this shortcoming. This paper details an optimization-based parameter identification scheme which was used to identify new parameters for the ducted fan model. In addition to solving a traditional least squares optimization problem over the space of parameters, we introduce the additional concept of optimizing over the space of inputs as in a typical optimal control problem. The projection operator approach to trajectory optimization described is used to perform the trajectory optimization and enables many iterations to be executed in a reasonable time span. Finally, the newly identified parameters enabled us to use pure MPC with no additional stability augmentation. We were unable to achieve the same results consistently with other techniques.
Optimization based parameter identification of the Caltech ducted fan
American Control Conference, 2003 ; 2697-2702
2003
6 Seiten, 9 Quellen
Aufsatz (Konferenz)
Englisch
Parameter Identification and Attitude Control of Ducted Fan UAVs
Trans Tech Publications | 2014
|Caltech-JPL collaboration: complementary strengths
NTRS | 2018
|Technology Transfer at Caltech/JPL
NTRS | 2020
|Theodore von Karman's Caltech Students
British Library Conference Proceedings | 1997
|Ducted Wind Turbine Optimization
AIAA | 2016
|