Contribution to the indirect decentralized adaptive control of manipulation robots (Q1343368): Difference between revisions

Adaptive control for manipulation robots is studied. The synthesis of an indirect and decentralized adaptive control law is given. The method consists in the recursive estimation of unknown parameters of the symbolic model of the robot dynamics and of the afterward self-tuning control based on the obtained estimates. Newton-Euler equations are used to derive the symbolic model of robot dynamics and model parameters are then grouped into slow-varying and partially known-ones or fast-varying and a priori unknown-ones (payload). Adaptive versions of well-known robot control schemes such as decentralized control, computed-torque method and sliding-mode control are proposed. Decentralized structures for the feedback and feedforward loops are derived and stability and convergence analysis of the method is given. Real-time aspects of the method are underlined. The validation by simulation of the industrial robot model tracking a specified trajectory under unknown payload, is provided.