Development of Smart Mobile Manipulator Controlled by a Single Windows PC Equipped with Real-Time Control Software

Abstract

Despite significant advances in robotic technologies, few affordable robots for domestic applications are commercially available. This study aimed to develop an affordable yet practical mobile manipulator for use in household environments. The proposed mobile manipulator features a software-based real-time controller to control the motion of the manipulator and mobile base. The real-time controller was implemented on a single on-board Windows PC and can control the actuator and IO devices via EtherCAT communication. The major functionalities of the mobile manipulator include the pick-and-place of household objects and autonomous navigation within a domestic environment. For pick-and-place tasks, a deep neural network (DNN) was employed to recognize the object to pick up. For autonomous navigation, open-source ROS packages for SLAM and navigation were used along with the measurements from a LiDAR sensor and odometry. The feasibility of using the developed robot for domestic applications was experimentally evaluated.