Implementation of ROS-based Feeding Robot System with Intention Detection

In this study, we completed a feeding robot system, which USES a robotic arm combined with image processing to complete automatic feeding, reducing the burden of caregivers and improving the autonomy and dignity of the caregivers. The system detects the user open mouth through the camera, When the user opens mouth continuously for a preset period of time, it means that he wants to eat at this time, the system starts the robot arm to take food and feed, the feeding system continuously controls the robot arm to track the position of the mouth. Make sure that the food is delivered to the user's mouth so that the user can eat easily. The software is based on the robot operating system as a development environment. It contains four nodes, including a mouth detection node, an arm feeding control node, an arm driving node, and an arm status publishing node. In this study, we mainly develop the mouth detection node and the arm feeding control node, and then integrate the arm drive node and arm status publishing node provided by the manufacturer. Since each node operates independently, so even if the speed of image processing is different from the speed of arm feeding control, it does not affect the operation of the system. The experimental results show that the system can correctly complete the user's eating movement without the use of two-handed assistance to achieve practical purposes.