A Multidimensional Survey of Intelligent IoT Systems for Gunshot Detection

Gunshot Detection Systems (GDS) are increasingly deployed in public, institutional, and protected areas to enable rapid, real-time response to firearm incidents. Conventional acoustic-only solutions suffer from high false-alarm rates, limited adaptability to dynamic environments, and restricted spatial awareness. Recent advances in Machine Learning (ML), Deep Learning (DL), and Internet of Things (IoT) technologies have enabled intelligent GDS capable of context-aware detection, firearm classification, and shooter localization through multi-modal inputs including audio, video, and infrared data. These systems can adapt to background noise, time-of-day conditions, and deployment context, improving detection reliability while reducing latency. This survey focuses on sensor-driven, IoT-enabled gunshot detection pipelines that involve on-site sensing, edge or cloud-based inference, and operational deployment, and it does not address internet-only information collection or general crime analytics frameworks. This survey analyzes 156 peer-reviewed studies published between 1998 and 2025, organized into six analytical dimensions: detection and modeling approaches, application contexts, hardware and deployment infrastructure, communication and security frameworks, environmental and spatial adaptability, and performance evaluation. Across these, we identify persistent challenges spanning both technical and operational domains, including accurate detection in noisy or reverberant environments, reliable localization in GPS-poor or complex terrains, low-latency multimodal fusion, scalability under resource constraints, dataset scarcity, privacy and ethical considerations, vulnerability to spoofing or suppressors, and secure, real-time communication for rapid response. By consolidating technical advances and mapping them to these open problems, the review provides actionable insights for designing scalable, energy-efficient, and trustworthy GDS suitable for diverse real-world deployments.