Abstract:

Inarguably, STEM (Science - Technology - Engineering - Mathematics) education has become extremely popular. In fact, relative educational robotics frameworks appear as an excellent means to materialise/visualise mathematics, physics or engineering concepts, allowing for direct experimentation via creative thinking. Nevertheless, the procurement of appropriate educational robots is troublesome, since reliable modular systems are too costly and come with their own, ready-to-apply software platforms that have not been planned for tackling extraordinary situations, such as the COVID-19 pandemic; thus, major disruptions in this educational process have been imposed. In order to address such types of challenges, the TekTrain educational framework has been designed and implemented. Tektrain comprises support to open-hardware robotic and IoT (Internet of Things) devices, cloud AI (artificial intelligence) services and an easy-to-use web platform that empower proper STEM education management, problem design, assignment and evaluation, no-code (graphical) problem solving, both on real as well as simulated robots, on physical as well as virtual environments. TekTrain design is the outcome of a co-creation process with elementary school teachers, through a number of requirements elicitation and training workshops.

The Tektrain proof-of-value implementation is deployed on a custom robot with a differential drive 4-wheels chassis, where a number of sensors and effectors can be plugged on top. Such sensors include distance sensors, inertial measurement units, microphones, cameras, line following units, buttons and encoders, whilst the actuators comprise the main motors, LEDs, speakers, a touch screen and a pan-tilt mechanism on top of which the camera is placed. The TekTrain middleware for robots caters for the control of the above-mentioned hardware, provides resource abstraction and enables application execution, where “applications” essentially are the students’ solutions to the instructor-defined problems. The TekTrain proof-of-value AI cloud services (e.g. face detection using the camera, speech to text and text to speech etc) can be seamlessly used as blocks within the applications, this way allowing complex application development.

The TekTrain concept has been tested in two summer hackathons in the Halkidiki region, Greece. Fifty sixth-grade pupils along with their instructors took part in each hackathon. The first hackathon was run remotely due to the COVID-19-related lockdown, where pupils participated from their school locations while the TekTrain staff, along with the robots which were at a remote location. In this first hackathon the children practiced on fundamental logic problems that included serial commands and loops, and tested their solutions to the remote robots which could be seen via teleconferencing. During the second hackathon robots were taken to the school premises, as COVID-19 measures relaxed. The tasks included randomness, parallel execution and preemptions. A computational thinking test was administered before and after the first hackathon, and after the second hackathon, so as to identify improved skills after the TekTrain deployment. Results show that the Tektrain approach can be employed for efficient STEM education.

Keywords:

Educational robotics, STEM, Internet of things, Remote Learning, COVID-19.