AbstractThe possibility of applying sequential state and parameter estimation procedures in the case of incorrectly modelled dynamic systems has been investigated. Of major concern was the estimation of the modelling errors, as a step to improve the accuracy of the state estimates. The modelling errors are modelled as piece-wise continuous low-order polynomials, the coefficients of which are included in the estimated vector. An effective algorithm is introduced to detect the phenomenon of filter divergence. Once divergence of the filter has been established a feed-back is activated to adapt the filter to the new situation. The feasibility of this approach is demonstrated with the help of a few realistic model problems in the field of satellite attitude determination.
Estimation of modelling errors in satellite attitude determination
Acta Astronautica ; 4 , 9-10 ; 1059-1073
1976-12-28
15 pages
Article (Journal)
Electronic Resource
English