An Innovative Collision-Free Image-Based Visual Servoing Method for Mobile Robot Navigation Based on the Path...

Area of Science:

• Robotics
• Computer Vision
• Control Systems

Background:

• Visual servoing (VS) is crucial for robot guidance using camera feedback.
• Existing methods often struggle with real-time obstacle avoidance and maintaining field-of-view.
• Monocular vision presents challenges due to depth ambiguity.

Purpose of the Study:

• To develop an innovative 2D visual servoing (IBVS) approach for object guidance.
• To ensure collision avoidance and keep the target within the camera's field of view using only monocular data.
• To demonstrate the method's effectiveness on a two-wheeled mobile robot.

Main Methods:

• Utilizing differential flatness of system dynamics to control image-plane trajectories.
• Linking current and desired robot configurations via image-plane point control.
• Implementing an inverse problem approach for trajectory planning.
• Validating the strategy through numerical simulations.

Main Results:

• The method successfully generates and tracks image-plane trajectories, ensuring obstacle avoidance and field-of-view maintenance.
• The number of required image points for trajectory definition depends on robot control inputs (e.g., one point for a two-wheeled robot).
• Optimal trajectory planning is essential to prevent non-intuitive Cartesian movements and potential collisions.

Conclusions:

• The proposed 2D visual servoing method effectively guides robots using monocular vision.
• Differential flatness provides a robust framework for controlling image-plane dynamics.
• The approach offers a promising solution for safe and reliable robot navigation in complex environments.