A Novel Velocity-Based Control in a Sensor Space for Parallel Manipulators.

Area of Science:

• Robotics
• Control Systems
• Computer Vision

Background:

• Object tracking with unknown trajectories is challenging for conventional model-based controllers, especially for parallel robots with complex kinematics.
• Existing position-based camera space manipulation (CSM) models are inadequate for tracking dynamic targets due to fixed point mapping.
• Vision-based controllers offer robustness to kinematic uncertainties by correcting trajectories using sensor feedback.

Purpose of the Study:

• To present a novel velocity-based camera space manipulation (CSM) control strategy for enhancing object tracking accuracy.
• To address the limitations of position-based CSM models in dynamic tracking scenarios.
• To validate the proposed control method on a Delta-type parallel robot.

Main Methods:

• Developed a velocity-based CSM control law that establishes a relationship between a movable trajectory and the robot's end effector position.
• Implemented and tested the control strategy on a Delta-type parallel robot.
• Conducted experiments involving static tracking, constant speed linear trajectory tracking, and freehand trajectory tracking.

Main Results:

• Achieved an average static tracking error of 1.09 mm.
• Demonstrated tracking errors of 8.89 mm, 11.76 mm, and 18.65 mm at speeds of 7, 9.5, and 12 cm/s, respectively.
• Recorded maximum tracking errors of 11.79 mm during motion and 1.44 mm for static positioning, with a control cycle time of 48 ms.

Conclusions:

• The novel velocity-based CSM control significantly reduces tracking errors for parallel robots compared to previous methods.
• The proposed approach enhances the robustness and precision of robotic object tracking, particularly for unknown trajectories.
• This work provides a more effective control strategy for vision-guided robotic manipulation tasks.