Automatic Lifting System for Crane Safety using Machine Learning

Abstract

Crane is extensively used to carry loads. But many safety measures need to be considered while operating a crane. One of these is material overloading and not lifting a load at an appropriate safety angle, leading to accidents. In this paper, a methodology is introduced to automate the safety lifting system in a crane by calculating the safety angle to prevent workplace hazards. A crane model is developed with a servo motor, sensor, and buzzer. The boom of a crane is moved to the desired safety angle with the help of the servo motor MG90S. MPU6050 sensor is used to read the current boom position of a crane. Calculation of the safety angle for the given load is automated with the help of machine learning. The load chart of a rotary crane is used as the dataset for training. The machine learning model is trained using Gaussian process regression in MATLAB with statistics and machine learning toolbox. The trained model predicts the boom radius for the given weight. With the help of mathematical modeling safety angle is obtained. If the boom's current position equals the safety angle predicted with the model, then the buzzer will ring, indicating the appropriate safety angle to lift the given load. Otherwise, it will move the boom of a crane model to the safety angle. For this implementation, the MATLAB support package for Arduino hardware is used.