LQR/RPS design of robust controllers for a one‐link robot

Abstract

In this paper, a design method for robust controllers for a one‐link robot is developed. Based on a linearized model of this nonlinear robot system, the linear quadratic regulator (LQR) design method is incorporated with a reduced parameter sensitivity (RPS) technique to design a robust proportional‐plus‐integral (PI) controller to achieve better performance in position control. This robust controller can tolerate real parameter variations resulting from the mechanical loads of applications. A direction method is further developed and used to evaluate the stability robustness of this controlled system. Combining the direction method with the LQR/RPS method, a robust PI controller can be generated to avoid improper over‐design. Both computer simulation and experimental results demonstrate the feasibility and effectiveness of the proposed approach.

Key words :

• real parameter uncertainty
• LQG/RPS
• direction method
• stability region

Notes

Correspondence addressee