Fuzzy dynamic control of manipulators: a scheduling approach

This paper presents a method for a dynamic control of manipulators. Fuzzy techniques are used to schedule both control torques and multivariable Proportional-Integral-Derivative controller. For each control variable, fuzzy subsets are defined by a set of crisp values and membership functions. Then we consider the control variables space and call crisp point a point in this space whose coordinates are crisp values. For each crisp point the control torque and a Proportional-Integral-Derivative controller are computed and stored. A fuzzy scheduler is in charge of sending torques and controllers to the robot