Picking using 3D object detection for mobile assistive robots

Abstract

Elderly and disabled persons often have limited mobility and are unable to receive packages. For this reason, assistive robots have been researched and developed. In recent years, the tasks of assistive robots have been controlled using a digital twin. However, the robot’s automated movements are limited to simple actions such as looking around and gesturing. In order for the elderly to perform complex picking tasks, they must be able to perform picking with simple operations. In this paper, a Head Mount Display (HMD) is used to run to the package and pick it at the touch of a button. This robot is equipped with an RGB-D camera at its hand and can see the object in three dimensions. The grasping points of the box are calculated by matching the point cloud with a box-shaped picking model. Finally, pick-and-place experiments are conducted to demonstrate the effectiveness of the proposed method.