Academic framework for teaching robot programming at university

This paper presents a framework for teaching robotics at engineering university degrees in a practical way. It is open source, multiplatform (Linux, Windows, MacOS) and emphasizes the programming of the robot intelligence. It consists of a collection of exercises of several types, which are similar to the new robotic applications (autonomous vehicles, drones, vacuum cleaners...). It uses the Gazebo simulator and the Python programming language. For each exercise an academic application has been created which performs all the auxiliary tasks like graphical interface, connection to the sensors and actuators, timing of the code, etc. and hosts the student's code, who can focus on the perception and control algorithms. Each academic application combines an infrastructure part, which it is provided and remains hidden, and the student part, who simply fills a form file with the robot logic for such exercise. It has been successfully used in several subjects at engineering degree, master and robotics introductory courses.