Low-Frequency Perturbation Estimation based on Low Gain Sliding Perturbation Observer while Contact with Environment

Abstract

Sensor less force estimation is a challenging issue among researchers working in the domain of robot manipulators. Different researchers have proposed a unique algorithm to estimate/ observe reaction/ contact force during robot operation includes grinding, cutting, painting, and assembly. This contact force generates perturbation at each joint of the robot manipulator. Sliding mode control with sliding perturbation observer (SMCSPO) is a robust control scheme which is an integration of sliding mode control (SMC) and sliding perturbation observer (SPO). SMCSPO not only estimate perturbation but also uses this estimated information to cancel the effect of actual perturbation. Perturbation is a sum of contact force/reaction force from the environment, modeling, and parametric uncertainties. In SMCPO bode plot between the estimated and actual perturbation is low pass filter with the breaking frequency of lambda. Estimation accuracy and system performance improved with the larger value of lambda. Contact / Reaction force is usually constant and has low frequency but SMCSPO estimates all categories of perturbation. To extract perturbation at the robot joint due to reaction force low gain sliding perturbation observer (LGSPO) has proposed in this research. In this suggested scheme the small value of lambda (breaking frequency) has been selected to estimate low-frequency perturbation (due to contact force). This LGSPO works in parallel with SMCSPO. Design logic has been implemented on MATLAB/ Simulink, and experiments have been conducted on seven degrees of freedom (DOF) assembly robot manipulator.