Nonlinear Friction Estimation in Elastic Drive Systems Using a Dynamic Neural Network-Based Observer
2013 ◽
Vol 17
(4)
◽
pp. 637-646
◽
Keyword(s):
This paper presents a neural-network based observer for nonlinear elastic drive systems. The proposed nonlinear observer uses a Diagonal Recurrent Neural Network (DRNN) combined with the dynamics of a linear Two-Mass-Model (2MM) system to identify nonlinear characteristics of the drive system such as Coulomb and nonlinear viscous friction torques. Theoretical analysis of the proposed neural-network based observer, including the neural network structure and the training algorithm convergence, are presented and discussed. Simulation results are confirmed experimentally using a 2MM system setup.
1999 ◽
Vol 3
(5)
◽
pp. 427-430
◽
2012 ◽
Vol 605-607
◽
pp. 2175-2178
2015 ◽
Vol 740
◽
pp. 871-874
2019 ◽