DOE OSTI · 21143557
Neural Network Identification For a C5 Parallel Robot
Abstract
This paper presents the design and analysis of a neural network-based identification of the inverse dynamic model of a C5 parallel robot. The identification structure is designed using the black box form (the dynamic model is completely unknown). This identification uses real data acquired on the C5 parallel robot by applying a nominal control scheme (PD). The desired trajectories of this scheme are based on Fourier series and the coefficients are chosen in a heuristic way. We have used this type of desired trajectories to obtain exciting trajectories for identification procedure. Three identification schemes are tested and compared. The comparison is performed based on the number of parameters used in each architecture and the quality of the generalization error. The used neural network is of MLP type and composed of one hidden layer.
Explore related subjects
Keep this discovery
Explore connections, maps & timelines
Daachi, M E, Chikouche, D, Achili, B, Daachi, B. 2008-06-12. Neural Network Identification For a C5 Parallel Robot. https://doi.org/10.1063/1.2952984. https://searchnasa.com/record/osti%3A21143557/neural-network-identification-for-a-c5-parallel-robot
Citations retain the original authors and DOI and include this site’s record URL. If this discovery service contributed to your research, you can also cite Search NASA by Indiciae: BibTeX · RIS.