Machine Learning-Based Iot Sensor Data Analysis

Abstract

Through early machine defect detection, costly breakdown prevention, downtime reduction, and maintenance optimization, predictive maintenance using IoT sensors eventually saves maintenance costs and production losses. Using pre-processed machine steps and speeds data, this study creates a predictive maintenance system that uses AdaBoost machine learning to identify real-time machine stops in knitting machines. It predicts six different sorts of stops: gate, feeder, needle, completed roll, idle, and lycra. Primarily, sensor devices used in energy management systems are low on computing power, memory, and battery capacity, and it is difficult to recharge the battery of sensors. The primary aim of this research work is to explore the problem of energy usage in the scenario of IoT devices with different energy demands. IoT devices execute data, send packets, read sensor values, and control devices. Due to doing these operations, the device will lose some power, which might restrict network operation. There exist various numbers of packets transmitted and received by various communication devices and various packet lengths.