Smart environments rely on distributed IoT sensing devices capable of collecting and interpreting context-aware data, enabling automation, adaptability, and informed decision-making. However, deploying such systems in practice presents challenges related to the coordination and organisation of devices under dynamic conditions. This paper presents a lightweight, self-organising architecture based on distributed Monitoring Units (MUs), each equipped with sensing, processing, and communication capabilities. A distributed coordination protocol is proposed to enable MUs to autonomously determine their roles based on current system conditions and availability. The design is validated through two implementations –targeting air quality and environmental noise monitoring– using different hardware architectures. Results demonstrate the system’s simplicity, adaptability, and suitability for constrained IoT deployments.