The utilization of miniaturized sensors that are cost-effective and possess enhanced computational power has emerged as a preferred solution for field monitoring due to their energy efficiency. However, the conventional approach of centralizing data processing in the Cloud exhibits limitations concerning energy consumption and environmental sustainability. In light of these concerns, it seems feasible to leverage Machine Learning algorithms directly on IoT devices for field monitoring while minimizing energy expenses. This objective can be attained through diverse optimization approaches, including but not limited to, regulating sensing parameters and employing Federated Learning in a Wireless Sensor Network.