Artificial neural network for identification and tracking control of a flexible joint single-link robot

A Survey on the Control of Flexible Joint Robots

“…In 1989 Khorasani, et al, claimed that using a neural network could improve the responses of FJRs [81]. References [82][83][84] used the benefits of neural networks to design a controller. Reference [85] provided a neural network approach which requires no off-line learning phase and no lengthy and tedious preliminary analysis to find the regression matrices.…”

“…Some combinations of robust and adaptive approaches were proposed in [153,148,149]. A paper providing robust controller based on a neural network was also published in [82]. A simple method with PD action on the rotor position and an integral control action on the link position was shown to provide semiglobal asymptotic stability of the desired link position in [154] …”

Design of a high-speed line following robot that smoothly follows tight curves

“…This approach resulted in smooth control performance without overshoot with just 3 tansig hidden layer neurons. ANN based control strategies for some benchmark robot control tasks that use the function approximation abilities of ANNs are explored in [12][13][14][15][16][17].…”

Tip trajectory control of a flexible-link manipulator using an intelligent proportional integral (iPI) controller

“…Moreover, most model-based controller design methods neglect, on the one hand, the significant non-linear variations in the dynamics of this type of manipulator when operating at high speeds or under rapid payload changes (Bellezza et al, 1991;Li et al, 1998;Mohamed and Tokhi, 2004). On the other hand, most model-based studies do not incorporate the very significant non-linear effects of friction at low speeds: an effect capable of generating additional modes of oscillations in the system (Hisseine and Lohmann, 2001;Kim and Parker, 1993;Rokui and Khorasani, 1997). Besides, several robust control strategies and on-line identification techniques have been reported for improving controller performance under model uncertainties, bonded disturbance effects and friction (Atashzar et al, 2010;Erfanian and Kabganian, 2009;Lee et al, 2007;Mamani et al, 2011); however, there is merit in the continuing evaluation of the real-time cost of implementing these robust schemes.…”