Collision Avoidance Route Planning for Autonomous Medical Devices Using Multiple Depth Cameras

Radiography is one of the most widely used imaging techniques in world. Since its inception, it has continued to evolve, leading development intelligent and automated radiography systems that are able perceive parts their environment respond accordingly. However, such do not provide a complete view examination space therefore unable detect multiple objects fully ensure safety patients, staff equipment during execution movement. In this paper, we present system architecture based on ROS (Robot Operating System) solve these challenges integrate an autonomous X-ray device. The retrieves point clouds from range sensors placed at specific locations room. By integrating different subsystems, merges data map space. It also implements downsampling clustering methods identify later distinguish obstacles. A subsystem generates bounding boxes detected obstacles feeds them motion planning framework (MoveIt!) enable collision avoidance execution. At same time, deep neural network model (PointNet) classify Finally, developed provided promising results after being deployed Gazebo simulated use case test platform.