Research on Integrated Control Technology and Intelligent Algorithm for Multi-Source Data Fusion

The emergence of Agentic AI, anticipated to become pivotal by 2025, signifies a transformative shift in artificial intelligence from augmenting knowledge to enabling autonomous execution, revolutionizing decision-making and operational automation. This evolution profoundly impacts higher education, particularly innovation and entrepreneurship programs, which face urgent demands to transition from skill-based training to cultivating interdisciplinary, AI-driven competencies. Modern pedagogical tools—such as AI-powered assessments and adaptive learning platforms—enhance students’ problem-solving abilities, creativity, and entrepreneurial readiness. However, traditional evaluation frameworks remain inadequate, necessitating dynamic, data-driven approaches to track learner progress and outcomes. Parallel advancements in intelligent perception and control systems address limitations in real-time data processing and adaptability within complex environments. This study proposes a novel AI-IoT integrated method, combining convolutional neural networks (CNN) and long short-term memory (LSTM) algorithms to optimize multi-sensor data fusion, alongside reinforcement learning (RL)-based adaptive control for dynamic decision-making. These dual developments—educational paradigm shifts and technological innovations—underscore AI’s expanding role in reshaping both human-centric systems and automated infrastructures, while highlighting critical challenges in scalability and ethical governance for future research.