A predictive maintenance system for wireless sensor networks: a machine learning approach

<p>Predictive maintenance system (PdM) is a new concept that helps operators evaluate the current status of their systems, and it also assists in predicting future quality these systems scheduling action. This paper proposes PdM model utilizes machine learning to predict system’s operational after M active steps based on L previous observations implemented by feedforward neural network (FFNN). We use quantization encoding schemes reduce complexity system. apply proposed build for wireless sensors networks (WSNs), where our concern state as far data transfer concerned. The FFNN provides forward prediction consecutive time future, readings service (QoS) requirements WSN. demonstrate relation between accuracy. found larger leads higher error, entails smaller error. investigate how can implement real-time system.</p>