Home

A Novel Approach for Odometry Calibration on Wheeled Robots

Odometry calibration is a field widely studied in the literature for different steering geometries, such as the differential drive, tricycle, and omnidirectional ones. The calibration methods either estimate the kinematic parameters (e.g., the diameter of the wheels) or adjust directly the robot's model to improve the odometry accuracy of the wheeled odometry. To the best of our knowledge, it is not available any method intended for two or more distinct steering geometries.

So, our method intends to generalize the odometry calibration problem independently of the steering geometry considered. First, both forward and inverse kinematics are defined for the differential drive, Ackerman/tricycle, and omnidirectional geometries that are used in the experiments. Next, it is developed an optimization-based approach that uses the improved Resilient Propagation without weight-backtracking (iRprop-) due to its accuracy and robustness while being only a first-order optimization algorithm for estimating the kinematic parameters using only the position data of the robot. Even though the proposed algorithm is not path-specific, it is also suggested a test path for retrieving odometry data for the calibration procedure.

The results obtained were evaluated in terms of maximum position and absolute orientation errors over the path and at the final point using four robots on the suggested test path, the square, and on an arbitrary path. The OptiTrack motion capture system was used as a ground-truth. Also, the proposed method is compared with several works proposed in the literature for the four robots used in the experiments.

Overall, the proposed method led to improvements on the odometry accuracy over the existent literature while being a generalized approach to the odometry calibration problem. The method implemented in MATLAB is available in GitHub as an open-source repository.

Table of Contents

• Kinematics

Contributors

• Ricardo B. Sousa (corresponding author)
  • Affiliations: UTAD, INESC TEC
  • Email: [email protected]
• Marcelo R. Petry
  • Affiliations: INESC TEC
  • Email: [email protected]
• Professor António Paulo Moreira
  • Affiliations: FEUP, INESC TEC
  • Email: [email protected]
• Professor Paulo Costa
  • Affiliations: FEUP, INESC TEC
  • Email: [email protected]

Institutions

• CRIIS - Centre for Robotics in Industry and Intelligent Systems from INESC TEC - Institute for Systems and Computer Engineering, Technology and Science
• Faculty of Engineering, University of Porto (FEUP)
• Universidade de Trás-os-Montes e Alto Douro (UTAD)