Passing through Open/Closed Doors: A Solution for 3D Scanning Robots

Area of Science:

• Robotics
• Computer Vision
• 3D Data Processing

Background:

• Autonomous mobile platforms require sophisticated navigation for complex environments.
• Building scanning necessitates efficient methods for interior data acquisition and modeling.

Purpose of the Study:

• To propose a traversing door methodology for mobile platforms.
• To develop a system for autonomous door interaction and passage in buildings.

Main Methods:

• Utilizing dense 3D laser scanners for environmental perception.
• Implementing door state recognition (open, closed, semi-closed).
• Developing robot arm interaction for closed door manipulation.

Main Results:

• The MoPAD platform successfully navigated and passed through open doors in real-world building environments.
• Closed door passage was successfully simulated using a 4DOF robot arm in Gazebo.
• Accurate door identification and state recognition were achieved.

Conclusions:

• The proposed door traversing methodology is effective for autonomous mobile platforms.
• The system enhances the capabilities of mobile platforms for building digitization.
• The solution shows potential for adaptation to various platforms and robot arms.