Skip to content

Latest commit

 

History

History
58 lines (40 loc) · 3.04 KB

README.md

File metadata and controls

58 lines (40 loc) · 3.04 KB
Raw

A Practical Method For Hand-Eye Calibration Through Rotation Fusion

Overview

Jin, G., Yu, X., Chen, Y., Li, J. (2024), A Practical Method For Hand-Eye Calibration Through Rotation Fusion, submitted to IEEE Transactions on Instrumentation and Measurement.

The hand-eye calibration problem is a fundamental problem in visual assisted robotics. Ensuring the calibration robustness is crucial as the calibration results are used in every task execution. Unlike traditional separate or simultaneous method, we propose a rotation fusion method to further improve the robustness. The proposed method is practical and includes four straightforward steps. First, the forward rotation estimate is obtained using the rotation of AX = XB. Second, the backward rotation estimate is achieved via the translation of AX = XB, which is not involved in existing methods. Third, the two rotation estimates are fused with identity covariance to cope with different noise conditions. Last, the translation estimate is obtained by a least-square minimization.

Figure: Signal flow diagrams of (a) the separate method, (b) the simultaneous method, and (c) our fusion method for hand-eye calibration.

How to use

Dependencies

The code runs on Matlab R2023a without any additional dependencies. The necessary auxiliary functions can be found in the "auxiliary" folder.

Main Instructions

To run the fusion calibration, call

[R_out,t_out,rnti]=HECfuse(RAin,tAin,RBin,tBin)

where

  • RAin (3x3xn): rotation matrix of A,
  • tAin (3xn): translation vector of A (unit: m),
  • RBin (3x3xn): rotation matrix of B,
  • tBin (3x3xn): translation vector of B (unit: m),
  • R_out (3x3): rotation matrix of hand-eye pose,
  • t_out (3x1): translation vector of hand-eye pose (unit: m),
  • rnti (1×1)： runtime (unit: seconds).

The comparison methods include

  • HECrot : forward separate method,
  • HECtran : backward separate method,
  • HECsim : simultaneous method,
  • HECTsai : Tsai's method,
  • HECWu : Wu's method,
  • HECSARA : Sarabandi's method.

Please refer to the submitted article or reference for details.

Demos

Demo main1, main2, and main3 correspond to the accuracy comparison of different types of methods, while Demo main4 is a comparison of computational efficiency. When the program ends, a visual result will be presented.

Reference

  • Tsai R Y, Lenz R K. A new technique for fully autonomous and efficient 3 d robotics hand/eye calibration[J]. IEEE Transactions on robotics and automation, 1989, 5(3): 345-358.
  • Wu J, Sun Y, Wang M, et al. Hand-eye calibration: 4-D procrustes analysis approach[J]. IEEE Transactions on Instrumentation and Measurement, 2019, 69(6): 2966-2981
  • Sarabandi S, Porta J M, Thomas F. Hand-eye calibration made easy through a closed-form two-stage method[J]. IEEE Robotics and Automation Letters, 2022, 7(2): 3679-3686.

Contact

Gumin Jin, Department of Automation, Shanghai Jiao Tong University, Shanghai, jingumin@sjtu.edu.cn