Design and performance analysis of ANFIS-based and deep learning controllers for tower crane systems

Abstract

Ensuring safety, productivity, and efficiency in construction and industries is a significant task. Loads on tower cranes must be applied with high precision, and the swing angle must be carefully controlled. Intelligent controllers for tower cranes that do not need a mathematical model have been developed and evaluated in this study. Two intelligent control systems were investigated: one using a deep learning controller (DLC) and the other employing an adaptable neuro-fuzzy inference system (ANFIS). These controllers have been evaluated on tower cranes using the commonly used proportional differential (PD) controller. Simulations have been used to test how well the suggested intelligent controllers react to different working situations. The tests showed that the DLC was better at handling disturbances, accurate tracking, and responding quickly. As a result, the DLC made it easier to control the crane's swing angle and position, which improved security, reliability, and operating efficiency. The DLC is ideal for complicated crane control tasks as it can adapt to different operating conditions and disturbances.