American University of Sharjah (UNITED ARAB EMIRATES)
About this paper:
Appears in: EDULEARN16 Proceedings
Publication year: 2016
Pages: 5547-5557
ISBN: 978-84-608-8860-4
ISSN: 2340-1117
doi: 10.21125/edulearn.2016.2323
Conference name: 8th International Conference on Education and New Learning Technologies
Dates: 4-6 July, 2016
Location: Barcelona, Spain
The emergence of Internet of Things is enabling a host of new possibilities in educational management. One key area of concern for many educational systems are the activities that occur within a classroom. An observation of teacher and students' behavior can be used for diagnostic purposes and for providing feedback and training of teachers. While many observation instruments like CLASS and Stallings has been developed to observe what happens inside a classroom, most of these instruments are expensive to implement as they involve an observer painstakingly recording and coding the activities occurring within a classroom. Classroom videos have also been used to record and code teacher’s behavior. However, most of such methods do not scale due to the high cost. This is especially true for developing countries where the resources are scarce. This paper proposes a system that uses low-cost sensors like microphones, motion detectors, and humidity and temperature sensors to create a profile of what happens within a classroom on a real-time basis. Big data analytics are used to tag the collected patterns of activities into known behaviors. For example, noise and movement patterns can be associated with episodes where the teacher is lecturing. Similarly, quieter episodes suggest that children are engaged in a learning activity like drawing etc. Similarly, rowdy behavioral episodes can also be recognized. The frequency of such episodes within a class period can be used to provide a sketch of teacher's overall teaching practices and classroom management skills. Arduino and Raspberry Pi microcontrollers connected to a host of sensors are employed as edge devices to manage data that is further communicated via the MQTT protocol, sent in JSON format. CouchDB is used as the primary back-end database. SVM is used to construct classifiers that categorize patterns of activity for each teacher's classroom. Each school can subscribe information services based on the sensors in their classrooms. In addition, schools can also be benchmarked against comparative schools. Use of MQTT protocol easily enables red-flags and similar event notifications for the school principal etc. For example, extremely rowdy behavior or a classroom being unexpectedly quiet (or empty) can be subscribed to by appropriate individuals within the school. Finally, the classroom activity profiles can be provided to the Ministry of Education for regulatory compliance and policy formulation.
IoT, Educational Management, Classroom observation.