Adaptive control of mechanical manipulators

Implementação Em Hardware De Um Controlador Neural Para Robô

“…Na tentativa de superar estas dificuldades, vários algoritmos de controle usando as características dinâmicas do robô manipulador têm sido propostos [1], [4], [5], [6], [9].…”

Online Incremental Learning of Inverse Dynamics Incorporating Prior Knowledge

“…However, this performance is highly dependant upon the accurate representation of the robot's dynamics, which includes precise knowledge of the inertial parameters of link mass, centre of mass and moments of inertia, and friction parameters [2]. In practice, obtaining such a model is a challenging task This research is funded in part by the Natural Sciences and Engineering Research Council of Canada which involves modeling physical processes that are not well understood or difficult to model, such as friction [3] and backlash.…”

“…Traditionally, adaptive control strategies have been used to cope with parameter uncertainty [2], [5]. This allows parameters of the dynamic model such as end effector load, inertia and friction to be estimated online [2].…”

“…This allows parameters of the dynamic model such as end effector load, inertia and friction to be estimated online [2]. However, such methods are still reliant upon adequate knowledge of the structure of the dynamic model and are thus susceptible to modeling errors and changes in the model structure.…”

“…However, such methods are still reliant upon adequate knowledge of the structure of the dynamic model and are thus susceptible to modeling errors and changes in the model structure. Furthermore, accurate estimation of the model parameters requires sufficiently rich data sets [2], [6], and even when available, may result in physically inconsistent parameters meaning that constraints must be applied to the identified parameters [2], [7].…”