RoboMind. A platform for on-the-fly programming and inspection of behavior-based robot programs

Abstract

Lego Mindstorms is a popular tool used by universities for educational and demonstrational purposes. Lego Mindstorms is a set of buildable and programmable robotic kits, made by Lego. It allows for a high level of participation from the audience, while being easily programmable. However, for demonstrational and recruitment purposes, it is not without shortcomings. When there is a limited time to talk to people, it is difficult to explain and change a running Lego Mindstorm program. Typically the program is explained by looking at the robot's behavior and base the explanation on that. Modifying an already running program involves multiple slow steps that disrupts what the robot was previously doing. A more interactive approach would be to allow the audience to see what is happening inside the "robot's mind" while it is running. Where changes could be made on-the-fly without disrupting what the robot was previously doing. This thesis introduces RoboMind, a platform for on-the-fly programming and inspection of behavior-based robot programs. The idea behind RoboMind is to provide users with an interface where they can both visually inspect and modify a robots information and behaviors at run-time. Through RoboMind, users can add or edit existing behavior modules on-the-fly without disrupting what the robot was previously doing. The interface allows users to inspect which behaviors are running in addition to the collected sensor samples from the robot’s run-time. RoboMind has been thoroughly tested and developed for usage with the latest generation of the Lego Mindstorms, the EV3. RoboMind offers a fully featured Python-programming environment for run-time modification of running EV3 programs with the help of the ev3-python library made by the author.