The Internet of Things (IoT) and wireless sensor networks (WSNs) rely on large numbers of small devices that have limited energy, memory, and computing power and that are exposed to physical and network attacks. Blockchain and machine learning have been proposed as complementary tools for protecting such systems, but their cost and trust assumptions remain debated. This paper presents a focused review of 35 peer-reviewed studies, 32 of them published between 2022 and 2026, on blockchain-based security for IoT and WSNs. The studies are organised into eight themes: reviews and foundations, WSN intrusion detection, recent integrated frameworks, authentication and access control, consensus and scalability, artificial-intelligence-based trust and detection, privacy and federated learning, and post-quantum security. The paper compares representative frameworks and examines one recent two-layer intrusion detection system in detail, covering its design, threat model, reported results, and limits. The review finds that most solutions are evaluated only in simulations or small prototypes, that many depend on a few trusted nodes, that ledger growth and message overhead are seldom measured, that false alarms and detector cost receive little attention, and that attack coverage is narrow. Seven research gaps and six research directions are derived, and a roadmap for a lightweight, fairly evaluated, blockchain-based intrusion detection framework is outlined. The review is intended to help researchers select design options and evaluation practices for secure IoT systems.

Keywords— Blockchain; Internet of Things; wireless sensor networks; intrusion detection; machine learning; lightweight security