PBVS

Position-Based Visual Servoing (PBVS) estimates the target pose and controls the robot in Cartesian space.

Table of contents

1. Core idea
2. Typical advantages
3. Typical challenges

---

Core idea

PBVS closes the loop on 3D pose error:

1. Estimate camera/target relative pose from image measurements.
2. Compute pose error in SE(3).
3. Apply a control law to reduce translational and rotational error.

Typical advantages

• Intuitive geometric interpretation
• Smooth Cartesian trajectories
• Easy integration with model-based controllers

Typical challenges

• Sensitive to pose-estimation errors
• Requires camera calibration accuracy
• Can degrade with poor visual features