Dynamic movement primitives and force feedback: Teleoperation in precision grinding process

摘要

In this work, the Dynamic Movement Primitives (DMPs) concept for teleoperation on a robotic deburring machine is investigated. The correspondence problem occurs when human link and joint structures are different than target robot. Thus, transferring human movements to robot is generally problematic. Here, a teleoperation scheme is followed where human movements are directly converted to target robot while imposing capabilities of robot to the operator. A single degree of freedom haptic interface which relays the process forces to the human participant is built for controlling the deburring machine remotely. Human participant rotates the knob of the haptic device to keep the normal force constant. The knob motion is converted to the piezoelectric actuator via teleoperation. The DMPs of participants for different experiments are extracted and the ones with the lowest time span are collected in a single database which then used for simulation on different force values.