Visual-based quadrotor control by means of fuzzy cognitive maps

Area of Science:

• Robotics
• Artificial Intelligence
• Control Systems

Background:

• Intelligent control of unmanned aerial vehicles (UAVs) is crucial for autonomous operations.
• Image-based visual servoing (IBVS) offers a vision-guided approach for robotic control.
• Tracking dynamic ground targets presents significant challenges for aerial vehicles.

Purpose of the Study:

• To develop an intelligent IBVS method for quadrotor UAVs to track moving ground targets.
• To utilize Fuzzy Cognitive Maps (FCMs) for enhanced vision-based control.
• To address challenges including target mobility and unknown UAV altitude.

Main Methods:

• An image-based visual servoing (IBVS) framework was implemented on a quadrotor UAV.
• Fuzzy Cognitive Maps (FCMs), a soft computing technique, were employed for intelligent control.
• Perspective image moments were combined to derive control features for translational and yaw motions.

Main Results:

• The proposed FCM-based IBVS method enabled the quadrotor UAV to effectively track a moving ground target.
• The system demonstrated satisfactory performance in controlling both translational motion and yaw rotation.
• Simulation results across various target paths confirmed the controller's efficacy.

Conclusions:

• Fuzzy Cognitive Maps provide a suitable soft computing approach for vision-based intelligent control of UAVs.
• The developed IBVS scheme successfully addresses the complexities of tracking mobile targets with limited information.
• The controller's performance indicates its potential for real-world autonomous aerial tracking applications.