Visual Servoing of a Moving Target by an Unmanned Aerial Vehicle

Area of Science:

• Robotics
• Control Systems
• Computer Vision

Background:

• Tracking moving targets with unknown dynamics presents significant challenges for unmanned aerial vehicles (UAVs).
• Ensuring the target remains within the UAV's camera field of view is critical for continuous tracking.
• Maintaining feasible control inputs within the UAV's motion capabilities is essential for safe operation.

Purpose of the Study:

• To develop an advanced tracking controller for UAVs capable of handling unknown target motions.
• To ensure precise control over the relative position and orientation between the UAV and the target.
• To guarantee that the UAV's control inputs remain within its operational limits and the target stays within the camera's view.

Main Methods:

• Development of a nonlinear-model predictive controller (NMPC).
• Prediction of future target motion using quadratic programming (QP).
• Incorporation of constraints on the feature vector and control input within the optimization problem.

Main Results:

• The developed NMPC effectively controlled both relative position and orientation for unknown target motions.
• The controller successfully bounded control inputs, respecting UAV motion capabilities.
• Simulations demonstrated superior performance compared to traditional image-based visual servoing controllers.

Conclusions:

• The proposed NMPC offers a robust solution for UAV-based target tracking under complex motion scenarios.
• This approach enhances tracking accuracy and operational safety by considering system constraints.
• The method provides a significant advancement over existing visual servoing techniques for dynamic target tracking.