A HETEROGENEOUS DEEP ENSEMBLE APPROACH FOR ANOMALY DETECTION IN CLASS IMBALANCED ENERGY CONSUMPTION DATA

Abstract

Robotics education is increasingly essential for preparing learners with skills in STEM, programming, and artificial intelligence. Yet, scaling such education equitably remains a global challenge. This study systematically reviews 20 peer reviewed publications (2013–2024) to explore how robotics education is being scaled across five dimensions: technical, pedagogical, institutional, financial, and geographic. Guided by Constructivist Learning Theory, the TPACK framework, and the Digital Divide and Equity Model, the review analyzes the role of cloud platforms, localized content, and connectivity innovations including mobile STEM labs. Cloud tools (e.g., Open Roberta, VEX code VR) support technical scalability, while teacher training through TPACK enhances institutional readiness. Constructivist approaches foster hands on, student centered learning, crucial for pedagogical scalability. However, financial and geographic inclusion remain limited. Solutions like mobile labs and low cost kits show promise but require broader policy support. The study calls for holistic, equity driven strategies to ensure sustainable and inclusive robotics education at scale.