Voice-Controlled Object Pick and Place for Collaborative Robots Employing the ROS2 Framework

Industrial automation has heavily relied on robot arms, mainly when human involvement is not feasible. This paper introduces a robotic arm designed to pick up objects based on a patient's verbal command, precisely the word 'pick.' The system incorporates a perception unit for goal identification, a motion planner, and a robot controller. ROS2 is the middleware for context processing, a speech-based user interface using VOSK, and MoveIt2 for motion planning and control. The VOSK model has been implemented for speech recognition, demonstrating encouraging results through transfer learning from pre-existing weights. A finite state machine (FSM) has been developed to manage the execution flow of the route to the collaborative robotic arm, enabling it to carry out tasks effectively.