作者: V. Zahn , R. Maass , M. Dapper , R. Eckmiller
DOI: 10.1109/ROBOT.1999.774018
关键词:
摘要: We present a novel type of friction estimation applied to the field sensorless force/position control. As part position based neural force control (NFC-P) and external allows without using sensor. NFC-P consists hybrid controller that accurately generates contact forces objects with arbitrary flexibility uncertain distance or shape. performs by modifying desired joint angle changes in direction before they are fed into computed torque controller. The inverse dynamics manipulator is modeled Kinematic mappings guarantee singularity robustness entire workspace. Results from real time experiments presented 6-DOF industrial as testbed.