Abstract
The mobile robot is surrounded by various structures. Glass is widely used for constructing walls and doors, and it is difficult for mobile robots to detect glass. Related studies consider the optical properties of glass when reconstructing the position of glass using LiDAR data. However, most of them depend on SLAM (Simultaneous Localization And Mapping) or Localization algorithm and prior information about the environment. This paper proposes a point cloud augmentation algorithm to predict glass panel position. The point cloud outlier is removed with the proposed local standard score, and the boundary of free space is augmented by linear interpolation. It is modularized as a post-processing of LiDAR sensor data that does not require any prior information about the environment. This algorithm was applied to an actual mobile robot. Glass recognition performance was evaluated by SLAM while driving in an office environment.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
