Reactive Control for Collision Evasion with Extended Obstacles.

Area of Science:

• Robotics
• Autonomous Systems
• Marine Technology

Background:

• Collision evasion is crucial for Autonomous Underwater Vehicle (AUV) exploration in 3D underwater environments.
• AUVs use 3D active sonar to detect obstacles, which can be complex, extended surfaces rather than simple point targets.

Purpose of the Study:

• To develop reactive collision evasion controllers for AUVs navigating 3D underwater spaces.
• To address collision avoidance with extended obstacles, including both static and moving types.
• To integrate hardware limitations, such as speed and acceleration, into the control system.

Main Methods:

• Development of reactive collision evasion controllers for AUVs.
• Consideration of extended obstacles defined by multiple sonar measurements.
• Integration of AUV hardware constraints (e.g., maximum speed, acceleration).
• Simultaneous navigation towards a goal and obstacle avoidance.

Main Results:

• The proposed controllers effectively manage collision evasion with extended obstacles in 3D environments.
• The system successfully guides the AUV towards its goal while avoiding detected obstacles.
• MATLAB simulations validated the effectiveness of the developed controllers.

Conclusions:

• The novel controllers offer a robust solution for AUV collision avoidance with extended 3D obstacles.
• This research advances AUV navigation capabilities in complex underwater environments.
• The approach is effective for AUVs utilizing sonar to map 3D obstacle boundaries.