A sliding mode controller with bound estimation for robot manipulators
- South China Univ. of Technology, Guangzhou (China). Dept. of Automation
- Hong Kong Polytechnic, Kowloon (Hong Kong). Dept. of Mechanical and Marine Engineering
In this paper, a sliding mode controller with the bound estimation is proposed for accomplishing trajectory control of robot manipulators. The proposed method is based on the linear parameterization approach, which decomposes the manipulator dynamic equation into the product of two quantities: a constant unknown vector of manipulator parameters, and a known nonlinear function of manipulator dynamics, called the regressor matrix. By using the regressor, a new sliding mode control law is designed. The authors assume here that the system matrix is completely unknown an no information on its possible size is given. The bounds of the unknown parameters, which are needed for calculating the switching gain, are not required known here. Instead, an adaptive algorithm is constructed such that the bounds are generated as a function of the state of the dynamics and the tracking error. The control law then depends on these estimates. Therefore, this results in a new version of the sliding mode control of robot manipulators. Usually, some unmodelled external forces, frictions, etc., may effect the stability. In this paper, the robustness with respect to the uncertainties is analyzed, where the uncertainties are assumed to be unbounded and fast-varying. Analysis results show that the robustness is guaranteed.
- OSTI ID:
- 5711124
- Journal Information:
- IEEE Transactions on Robotics and Automation (Institute of Electrical and Electronics Engineers); (United States), Journal Name: IEEE Transactions on Robotics and Automation (Institute of Electrical and Electronics Engineers); (United States) Vol. 9:2; ISSN 1042-296X; ISSN IRAUEZ
- Country of Publication:
- United States
- Language:
- English
Similar Records
Robust Neural Sliding Mode Control of Robot Manipulators
A robust adaptive robot controller