Predictive Observer based Lyapunov antiswing control for overhead visual crane

Based on the analysis of mathematical model of overhead crane, this research studys the control strategy of load´s anti-swing and rapid locating by obtaining the angle of load with image sensor. Because of the sensor´s image processing and transmission time, the angle signal arrived at controller is sampled and lags behind the current actual angle, affecting the effectiveness of the control strategy. The overhead crane with image sensor can be modeled as nonlinear system with time-varying delay. So the Predictive Observer (PO) is built to recover the real-time actual angle from the discrete and delayed angle signal based on the theory of Smith predictor. With real-time angle signal, the control method is designed based on Lyapunov function. The observer and Lyapunov control strategy have been tested by simulation in Matlab. After analyzing the simulation results, soft-start as an improvement to decrease the swing angle and make angle response more stable has been designed. In the end, a method to apply the control strategy to 3-dimension overhead crane system is proposed.