Threat Detection and Mitigation In Iot Environments: A Systematic Review Based On Approaches

Internet of Things environments have security challenges due to the diversity of devices, limited resources, and an imminent massive escalation. This systematic review examines different approaches such as: machine and deep learning, hybrid approaches, statistical, cryptographic, mathematical, blockchain and AI, in addi-tion to the results of other systematic reviews, related to threat detection and miti-gation in IoT environments. 413 research and review articles published between 2020 and 2025 were analyzed, out of a total of 17,631 articles, from different sci-entific databases, which were obtained by applying formulas, using Boolean op-erators, then the PRISMA method was applied, the review was divided into five main categories of approaches: machine learning, deep learning, hybrid methods (two or more approaches), emerging approaches, and systematic reviews. The in-formation was divided into four groups and another for systematic reviews. Data relevance was assessed based on: accuracy in binary classification, methodologi-cal innovation, advanced detection, attack coverage, and experimental robustness. The methodological diversity documented in this review demonstrates that there is no single solution for all IoT scenarios, but rather that the selection must con-sider factors such as resource constraints, real-time requirements, types of threats, and specific characteristics of the application domain.