Efficient Visual Feedback Method to Control a Three-Dimensional Overhead Crane

This paper presents an efficient method to capture the dynamic movement of a three-dimensional (3-D) overhead crane, enabling it to be controlled by visual feedback in real time with two cheap handy cameras. Two tracking areas and one positioning block in each frame are used to search image features and determine the useful vision information. The depth information of 3-D crane system in images can be also obtained by a tag on the dynamic plant. The presented visual tracking method involves comparison of the lightest or darkest points in the tracking or positioning area of a dynamic object and then computes the necessary trolley position and load swing in 3-D space. Upon tracking, the sensing data are sent to an adaptive fuzzy sliding-mode controller (AFSMC) to derive control power for the crane system. Accordingly, the merits of this AFSMC approach, including robustness of the sliding-mode control and the model-free property of the fuzzy logic for the 3-D crane system, are confirmed. Experimental results verify the improvement of the proposed methodology.