Application of a hand-eye self-calibration technique in robot vision

Wang, Haixia; Fan, Xianghua; Lu, Xiao

doi:10.1109/ccdc.2013.6561604

“…The calibration of the robotic tool center point using a camera system is sometimes referred to as hand-eye calibration. There are many papers found regarding hand-eye calibration techniques such as (partial list) [56], [57] and [58]. Only few were found to clearly mention the distinction between the robot's hand (end-effector) and the mounted tool such as mentioned by Dornaika and Horaud [55] and shown in Figure 2-9.…”

Section: Tool Center Point (Tcp) Localizationmentioning

confidence: 99%

Robotic tooling calibration based on linear and nonlinear formulations

Helal¹

2023

Preprint

0

View full text Add to dashboard Cite

Industrial robot calibration packages, such as ABB CalibWare, are developed only for robot calibration. As a result, the robotic tooling systems designed and fabricated by the user are often calibrated in an ad-hoc fashion. In this thesis, a systematic way for robotic tooling calibration is presented in order to overcome this problem. The idea is to include the tooling system as an extended body in the robot kinematic model, from which two error models are established. The first error model is associated with the robot, while the second error model is associated with the tooling. Once the robot is fully calibrated, the first error will be reduced to the required accuracy. Thus, the method is focused on the second error model. For the tool error calibration, two formulations were used. The first is a linear formulation based on conventional calibration as well as self-calibration methods while the second is a nonlinear formulation. The conventional linear formulation was extensively investigated and implemented while the self-calibration was proven to be inadequate for the tooling calibration. Moreover, the nonlinear formulation was demonstrated to be very effective and accurate through experimental result. The end-effector position estimation as well as the tool pose estimation were obtained using a 3D vision system as an off-line error measurement technique.

show abstract

“…The calibration of the robotic tool center point using a camera system is sometimes referred to as hand-eye calibration. There are many papers found regarding hand-eye calibration techniques such as (partial list) [56], [57] and [58]. Only few were found to clearly mention the distinction between the robot's hand (end-effector) and the mounted tool such as mentioned by Dornaika and Horaud [55] and shown in Figure 2-9.…”

Section: Tool Center Point (Tcp) Localizationmentioning

confidence: 99%

Robotic tooling calibration based on linear and nonlinear formulations

Helal¹

2023

Preprint

0

View full text Add to dashboard Cite

Industrial robot calibration packages, such as ABB CalibWare, are developed only for robot calibration. As a result, the robotic tooling systems designed and fabricated by the user are often calibrated in an ad-hoc fashion. In this thesis, a systematic way for robotic tooling calibration is presented in order to overcome this problem. The idea is to include the tooling system as an extended body in the robot kinematic model, from which two error models are established. The first error model is associated with the robot, while the second error model is associated with the tooling. Once the robot is fully calibrated, the first error will be reduced to the required accuracy. Thus, the method is focused on the second error model. For the tool error calibration, two formulations were used. The first is a linear formulation based on conventional calibration as well as self-calibration methods while the second is a nonlinear formulation. The conventional linear formulation was extensively investigated and implemented while the self-calibration was proven to be inadequate for the tooling calibration. Moreover, the nonlinear formulation was demonstrated to be very effective and accurate through experimental result. The end-effector position estimation as well as the tool pose estimation were obtained using a 3D vision system as an off-line error measurement technique.

show abstract