Decentralized integral sliding mode control of reconfigurable manipulator: An enhancement methodology for practical application

A decentralized backstepping integral sliding mode control scheme is proposed for reconfigurable manipulators. Based on the Lyapunov stability theory, the backstepping technique is adopted to simplify the controller design process, and the integral sliding mode control is introduced due to its advantages such as robustness, simplicity, fast response. Furthermore, the Radial Basis Function Neural Networks (RBFNN), whose weights are updated by adaptive laws, are used to approximate or compensate the unknown term and interconnection term. In this scheme, it not only can follow the joint desired trajectories with high accuracy performance within a short time, but also reduce the serious chattering phenomenon harmful to the working life of the actuator. Moreover, the decentralized architecture could reflect the modularization property of reconfigurable manipulators sufficiently. In contrast to the controller with linear sliding mode and terminal sliding mode, the simulation results with two different configurations of reconfigurable manipulators show that the proposed scheme could achieve the satisfactory control performance, as well as easily be applied in practice.