This master thesis’s main objective is to develop an application that makes it easier for a user without prior robotic experience to control a robot. By focusing on intuitive, simplicity and safety for the end-user, the hope is to shorten small companies’ path to invest and use robots. The environment consists of a standard computer and a Universal Robot 3 CB-series. The core of the application consists of acquiring a trajectory by Learning from Demonstration and converting these demonstrations to Dynamic Movement Primitives (DMPs). These DMPs are stored in a library and can generate robot trajectories at the users’ discretion. The project consists of studying methods and investigating support libraries to develop a ROS application that follows the requirements of being: intuitive, safe and straightforward. A graphic user interface has been developed using the PyQT framework to facilitate the intuitive and straightforward requirements. While making the application safe, the generated trajectory can not be performed before a collision check is calculated. The developed application performs connection to the robot, recording and storing demonstrations, computing, and storing Dynamical Movement Primitives in a library while also generating and executing safe trajectories on the robot.
The project package dependencies are found at:
https://github.com/UniversalRobots/Universal_Robots_ROS_Driver