A Scalable Architecture for the Dynamic Deployment of Multimodal Learning Analytics Applications in Smart Classrooms.

Area of Science:

• Educational Technology
• Computer Science
• Learning Analytics

Background:

• Smart classrooms utilize diverse software, devices, and wearables, generating substantial heterogeneous data.
• Current Multimodal Learning Analytics (MMLA) architectures are often ad-hoc and rigid, limiting scalability in real-world educational contexts.
• Existing MMLA approaches struggle with the dynamic reconfiguration needs of evolving learning environments.

Purpose of the Study:

• To propose a novel, scalable architecture for Multimodal Learning Analytics (MMLA).
• To address the challenges of deploying, dismantling, and reconfiguring MMLA applications in dynamic smart classroom settings.
• To demonstrate the feasibility and performance of the proposed architecture in managing diverse educational tools and devices.

Main Methods:

• Development of a novel MMLA architecture leveraging software-defined networking (SDN) and network function virtualization (NFV).
• Exemplification of the architecture's capability to solve scalability and reconfiguration challenges in MMLA deployments.
• Experimental validation of the architecture's performance and feasibility with various classroom devices and learning tools.

Main Results:

• The proposed architecture effectively addresses the scalability and reconfiguration challenges in MMLA.
• Experiments demonstrate the feasibility and performance of the architecture in dynamic smart classroom environments.
• The system successfully manages diverse learning tools and classroom devices through flexible reconfiguration.

Conclusions:

• The novel MMLA architecture offers a scalable and flexible solution for future smart classrooms.
• This approach can significantly improve the deployment and management of educational technologies and MMLA applications.
• The findings provide a foundation for implementing advanced MMLA in real educational scenarios.