Abstract: The integration of artificial intelligence components and humanoid robotic platforms into adaptive educational software ecosystems requires clear architectural planning, secure data flows and reliable interfaces between user-facing applications, AI services, databases and robotic interaction modules. This paper proposes a formal architecture-driven framework for adaptive educational IoT environments with robotic support. The proposed architecture is represented as a directed graph, where nodes correspond to software, AI, data, security, and robotic components, while edges represent controlled data flows and communication interfaces. This model introduces metrics for end-to-end latency, communication overhead, reliability, traceability, and risk exposure. A small-scale simulation-based validation is presented for evaluating representative learning scenarios involving AI-based personalization, teacher supervision, secure logging, and humanoid feedback robots such as the QTrobot RD-V2 i7 and NAO. This proposed framework supports students with special education needs by combining predictable interaction, multimodal feedback, emotional literacy support, teacher control, and GDPR-compliant data management.

Keywords: artificial intelligence, distributed IoT, humanoid robots, adaptive learning, special educational needs, QTrobot, NAO, secure data flow, educational robotics

Pages: 495-504