Smart and Sustainable Malware Detection in A Resource-Constrained IoT Environment: A Survey of Continuous and Context-Aware AI Methods

Abstract

Due to The rapid expansion of Internet of Things (IoT) technologies in many different fields in our lives has led to critical challenges of service security, also while most notably the risky threat of malware. Despite the many methods for detecting this malicious software, the resource constraints of IoT devices remain a fundamental challenge that reduces the effectiveness and operational capacity of these solutions. This study conducts a comprehensive survey to identify effective detection strategies in resource-constrained environments and aims to evaluate the feasibility of integrating Context Awareness with machine learning and deep learning approaches to establish a detection system that adapts to device resources and network conditions. This study highlights the fact that most of today's AI solutions are fundamentally defective: while highly accurate, they are unable to adjust and continuously learn to adapt to modern malicious patterns over a long scope. In this survey, we introduce our view, the main reason for this weakness is the omission of the context-awareness factor. and we will provide insights and criteria for designing lightweight, context-aware AI models to achieve the perfect balance between detection accuracy and model flexibility and incremental learning in resource-constrained IoT environments, to counter advanced threats.