Uncalibrated Visual Tracking Control for Networked Eye-in-Hand Robots by Adaptive Distributed Observer.

Area of Science:

• Robotics
• Control Systems
• Computer Vision

Background:

• Visual tracking of unknown targets by robotic manipulators presents challenges due to uncalibrated cameras and uncertain target motion.
• Maintaining target projection on the image plane is crucial for effective robotic manipulation.

Purpose of the Study:

• To develop a distributed control scheme for a network of robotic manipulators to visually track an unknown moving target.
• To ensure the target's image projection is maintained at a specified position despite uncalibrated camera parameters and uncertain depths.

Main Methods:

• An adaptive distributed observer estimates the target's motion.
• An image-space observer estimates target position and velocity for each robot.
• Adaptive laws are proposed to handle uncertain camera and robot parameters, leveraging depth-independent Jacobian properties.

Main Results:

• The closed-loop system stability is rigorously proven using Lyapunov stability theory.
• Asymptotic convergence of image-space tracking errors is demonstrated.
• Simulations with three-DOF robotic manipulators validate the scheme's effectiveness.

Conclusions:

• The proposed distributed control scheme effectively addresses visual tracking of unknown moving targets by networked robotic manipulators.
• The method robustly handles uncalibrated cameras, uncertain depths, and unknown target dynamics.
• This work contributes to advanced robotic perception and control for complex manipulation tasks.