Robot Geometric Calibration

Robotic accuracy plays a pivotal role in industrial applications, particularly in high-precision tasks such as aerospace manufacturing and inspection. Inaccuracies due to manufacturing tolerances, assembly errors, joint backlash, and wear should be enhanced by employing a comprehensive calibration process.

2024-2025

Company:
Year:

National Research Council Canada (NRC)

Skills:

Kinematics Modelling, MATLAB, Simulink

Sensor

Using an unknown artifact and the LMI High-Resolution Smart 3D Laser Line Profilers 2430, non-contact measurements of the robot end-effector's distance from an artifact along predefined paths are developed, enabling precise evaluation and adjustment of the robot's Denavit-Hartenberg (DH) parameters.

The Laser Line Profiler is an exceptional tool for this application due to its non-contact, high-precision measurement capabilities. Its compact design and ease of integration make it ideal for robot-mounted applications. The profiler provides dense point cloud data with sub-millimetre accuracy, enabling precise identification of kinematic discrepancies. Unlike traditional calibration methods that rely on external and often costly equipment, this integrated approach streamlines the process and enhances efficiency.

a man riding a skateboard down the side of a ramp
a man riding a skateboard down the side of a ramp

Stages of Robot Geometric Calibration

Established methodologies, the calibration process is structured into the following stages:

  1. Modelling: A mathematical model is developed to define the robot's kinematic structure using the Modified Denavit-Hartenberg (MDH) parameters. This ensures an accurate representation of the robot's geometry, including the tool and world frames.

  2. Measurement: The LMI 2430 profiler measures distances from the end-effector to the artifact across a range of poses.

  3. Error Identification: Errors in the robot's DH parameters are identified by comparing measured poses with theoretical ones derived from the kinematic model. Optimization algorithms minimize the discrepancies.

  4. Compensation: Adjustments are made to the robot's model to incorporate the identified errors. These updates improve path planning accuracy and enhance the robot's overall performance.

  5. Validation: The updated model's accuracy is being tested through repeatability and positional error assessments in the workspace.