AI on Edge Deploying for Real-Time Analytics

AI on edge's real-time processing capability has driven its popularity, enabling analytics through AI model deployment on edge devices. While research highlights its potential, barriers to adoption persist in areas such as energy efficiency, scalability, explainability, andsecurity. This paper addresses these challenges through three key strategies: adaptable modular design, real-world deployment validation, and dynamic AI specialization. The proposed framework achieves enhanced security and low latency through federated learning, lightweight AI models, and explainable AI techniques, while optimizing energy use via intelligent edge-cloud orchestration. This approach supports versatile AI applications beyond specialized use cases, facilitating cross-industry generalization in healthcare and smart cities. Additionally, the framework promotes sustainable AI implementation by integrating green computing practices for near real-time detection in critical situation, addressing modern technological challenges.