Title: Sensorless anti-swing control of automatic gantry crane using Dynamic Recurrent Neural Network-based soft sensor
Authors: Mahmud Iwan Solihin, Wahyudi
Addresses: Department of Mechatronics Engineering, International Islamic University Malaysia, PO Box 10, 50728 Kuala Lumpur, Malaysia. ' Department of Mechatronics Engineering, International Islamic University Malaysia, PO Box 10, 50728 Kuala Lumpur, Malaysia
Abstract: Sensor is an indispensable component in feedback control. In anti-swing feedback control of automatic gantry crane system, sensors are normally employed to detect trolley position and payload swing angle. However, sensing the payload motion of a real gantry crane, in particular, is troublesome and often costly since there is hoisting mechanism on parallel flexible cable. Therefore, sensorless anti-swing control method for automatic gantry crane system is proposed in this study. The anti-swing control is performed in feedback control scheme without using real swing angle sensor. Instead, soft sensor approach is used to substitute the real swing angle sensor. The soft sensor is designed based on Dynamic Recurrent Neural Network (DRNN) as a state estimator. Thus, a DRNN is trained using input-output data to estimate payload swing angle from trolley acceleration and input voltage of trolley actuator. An experimental study using lab-scale automatic gantry crane is carried out to evaluate the effectiveness of the proposed sensorless anti-swing control. The results show that the proposed sensorless method is effective for payload swing suppression since similar performance to the sensor-based feedback anti-swing control is obtained.
Keywords: anti-swing control; DRNN; dynamic recurrent neural networks; automatic gantry cranes; sensorless control; soft sensors; feedback control; payload motion; payload swing angle; trolley acceleration; trolley actuators.
International Journal of Intelligent Systems Technologies and Applications, 2009 Vol.6 No.1/2, pp.112 - 127
Published online: 25 Jan 2009 *Full-text access for editors Access for subscribers Purchase this article Comment on this article