Robotic Assistance for Physical Human–Robot Interaction Using a Fuzzy RBF Hand Impedance Compensator and a Neural Network Based Human Motion Intention Estimator

Abstract

This paper proposes a robotic assistance control scheme for intuitive teaching tasks by integrating the motion intention of human and real-time hand impedance compensation based upon a fuzzy RBF (Radial Basis Function) compensator. The motion intention of a human is estimated using a feedforward neural network. The parameters of the proposed fuzzy RBF hand impedance compensator are adjusted by considering hand impedance and robot dynamics. Three robotic assistance control schemes are compared: 1) robot without an impedance compensator; 2) robot with a constant-parameter impedance compensator; 3) robot with a fuzzy RBF hand impedance compensator. Several experiments have been conducted to verify the effectiveness of the proposed approach by comparing the contour error, exerted force and task time spent on teaching tasks. Experimental results indicate that the proposed fuzzy RBF hand impedance compensator has the best assistance results among the tested robotic assistance control schemes.

A Rapid Base Parameter Physical Feasibility Test Algorithm for Industrial Robot Manipulator Identification Using a Recurrent Neural Network

“…= L (3) xy − a (4) l (4) z − 6 i=4 a (4) d ( 4) M (i) , β (11) = L (3) xz , β (12) = L (3) yz , β (13) = L (3) zz + L (4) yy + 2d (4)…”

“…= l (3) x + 6 i=4 a (4) M (i) , β (15) = l (3) y + l (4) z + 6 i=4 d ( 4) M (i) , β (16) = L (4) xx − L (4) yy + L (5) yy , β (17) = L (4) xy , β (18) = L (4) xz , β (19) = L (4) yz , β (20) = L (4) zz + L (5) yy , β (21) = l (4) x , β (22) = l (4) y − l (5) z , β (23) = L (5) xx − L (5) yy + L (6) yy , β (24) = L (5) xy , β (25) = L (5) xz , β (26) = L (5) yz , β (27) = L (5) zz + L (6) yy , β (28) = l (5) x , β (29) = l (5) y + l (6) z , β (30) = L (6) xx − L (6) yy , β (31) = L (6) xy , β (32) = L (6) xz , β (33) = L (6) yz ,...…”

“…Model-based approaches for industrial robot manipulators such as advanced controller and observer design [1], [2], human-robot/robot-environment interaction [3], [4] and simulation of virtual robot motion [5], [6] all rely on accurate dynamic models. However, modeling accuracy of a dynamic model depends on its sensitivity with respect to environmental noise, especially non-Gaussian noise commonly seen in measurements [7].…”