SCOPING REVIEW OF AI-BASED DETECTION TECHNIQUES IN IOT PROJECT-BASED APPLICATIONS
Keywords:
internet of things, Artificial Intelligence, Machine Learning, Deep Learning, Anomaly Detection, Scoping ReviewAbstract
The Internet of Things (IoT) is rapidly expanding across domains such as smart cities, industrial automation, and precision agriculture, creating an urgent need for reliable and efficient detection systems. Artificial intelligence (AI) provides powerful capabilities to identify anomalies, intrusions, and domain-specific events within IoT ecosystems. However, the diversity of application contexts, techniques, and deployment strategies requires a systematic overview to guide future research. This scoping review mapped and synthesised evidence on AI-based detection techniques in IoT project-based applications, highlighting algorithmic trends, deployment approaches, and cross-cutting challenges. Guided by PRISMA-ScR and the JBI framework, the study reviewed ten publications from 2020 to 2024, covering smart manufacturing, smart buildings, agriculture, and network security. Classical machine learning remained competitive for structured data, while deep learning architectures showed advantages for complex or multimodal inputs. Deployment increasingly utilised edge and fog computing to improve latency and efficiency, though optimisation remains necessary for resource-constrained devices. Key limitations included the lack of integrated explainability and the limited evaluation of robustness against adversarial threats. The findings suggest a shift towards hybrid approaches that balance accuracy, interpretability, and efficiency within decentralised architectures. Future work should embed explainable and secure AI models into operational IoT systems, design lightweight adaptive solutions for edge environments, and conduct large-scale longitudinal studies to ensure scalability and resilience.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2025 Seminar Pembelajaran Sepanjang Hayat (SPSH) 2025

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.