Design of an Adaptive Controller to Improve the Condition Number of the Inertia Matrix of Serial Manipulators
Mariana de Paula Assis Fonseca1; Bruno Vilhena Adorno1; Philippe Fraisse2
1 Universidade Federal de Minas Gerais; 2 Université de Montpellier
Baixar PDF doi:10.20906/CPS/CBA2018-0982
Resumo
The ill-conditioning of the inertia matrix of serial manipulators is a problem intrinsic to multi-link open serial chains, which may potentially reduce the accuracy and performance of most motion control techniques based on the robot's dynamic model. In more extreme cases, the ill-conditioning can even result in unstable behavior. In order to solve this, an adaptive control law is applied to the robot to improve the matrix conditioning while ensuring that the well-conditioned inertia matrix is positive definite and hence continues to have physical meaning. Simulation results on a serial manipulator with seven degrees of freedom show that the proposed control law outperforms other commonly used techniques in terms of smoother behavior, smaller steady-state error, and smaller condition number of the inertia matrix.
Palavras-chave: Robot manipulator; Adaptive control; Dynamics; Inertia matrix; Ill-conditioning