Abstract:

In the era of the Fourth Industrial Revolution, technologies such as Artificial Intelligence (AI), Educational Robotics (ER) and other innovations have a significant impact on many areas of daily life, including education. This study attempts to examine how robotic assistants (RA)—i.e., ER that incorporates AI cameras, voice recognition sensors etc. —can contribute effectively to the experimental approach of concepts and phenomena of the Natural Sciences (NS).

Purpose of the research:
This study, is a part of a postdoctoral research and is being carried out alongside with other educational activities within the framework of the 13 Innovation Centers. It aims to study how RA can contribute effectively to the experimental approach of concepts and phenomena of NS and facilitate the processing, and analysis of the collected data.

Framework:
The theoretical background of this study is based on ER, as well as the application of AI and Machine Learning (ML). The above-mentioned are examined in the perspective of the three dimensions of the Next Generation Science Standards.

Methodology:
The methodology selected for this study is Design Based Research (DBR). DBR is carried out through repeated cycles of design, implementation, and reflection and allows the researcher to identify potential problems in the theoretical documentation, design and use of the study tools and attempt to resolve those by revising the initial design.

Study toolkit:
The hardware used for this study was based on the Nezha ER Kit for Micro:bit. Sensors, motors and actuators were also used. More specifically, particle and alcohol concentration sensors (such as CO, CO₂, smoke, etc.), and ambient air temperature and humidity sensors were used. The equipment is further enhanced with a voice recognition sensor, which allows the robot to interact directly with students. ER were programmed using the Microsoft MakeCode platform.

Research sample:
This study was conducted in two phases: a pilot phase (on a sample of 9 ninth-grade students) and a main phase (on a class of 24 students). Students recorded their observations on the corresponding activity sheets. This was achieved with their involvement in the construction and programming of the robots, the "training" of the voice recognition sensor, and the execution of the experimental measurements.
The collected data was coded, and processed using ATLAS.ti software. Based on the findings, changes were made to the design of the initial activities in order to enhance student’s experience.

Conclusions:
Data analysis showed that the integration of RA has the potential not only to facilitate and automate the conducting of experiments, but also to promote students' active participation. RA encouraged student engagement in processes such as learning about voice recognition sensors, a feature that helped promote the basic principles of ML. Through their engagement , students applied scientific practices and developed skills in organizing and conducting research, computational thinking, collaboration etc. They applied interdisciplinary principles such as cause-and-effect relationships, creation and utilization of models, and study of changes in phenomena. Finally, they developed technological literacy skills through the processes of assembling and programming ER and sensors.
Future plans include the integration of AI cameras and sensors that support IoT technologies. The above will enrich the learning process and will enable remote management of the RA.

Keywords:

Educational Robotics, Robotic Assistants, Artificial Intelligence, Machine Learning, New Generation Science Standards, Natural Sciences.