An eye-to-hand vision-based automated USB peg-in-hole assembly approach is proposed in this paper. The system employs two fixed cameras to observe the manipulator gripper and a USB square hole. The features of the gripper and USB square hole are observed to determine their poses. With an appropriate encoded error defined for the assembly task, a closed-loop visual servo controller capable of driving the encoded error to zero is proposed. The velocity control command can thus be computed for driving the manipulator to perform the USB peg-in-hole assembly task. The proposed approach has been successfully validated in experiments and can be used to precisely accomplish automated USB peg-in-hole assembly tasks with uncertain location.
展开▼