A fuzzy approximation to dexterity measures of mobile manipulators

In this paper, a fuzzy system is trained to approximate the manipulability index and the inverted condition number of a mobile manipulator. The kinematic model of a mobile platform with the attached manipulator is fused using twist theory and certain dexterity measures are examined throughout the configuration space. A Takagi-Sugeno-Kang fuzzy system is designed to approximate the dexterity measures and it is evaluated both for its error and its efficiency in computational cost. The proposed methodology is proved as a very efficient approximation that can be utilized in real-time motion planning algorithms without significantly adding computational cost to the controller.