Identification and control of hydraulic system using visual feedback

This paper presents identification and control of a nonlinear hydraulic system based on visual feedback provided by a simple camera and computer vision methods that result in an ability to measure the height of levels of liquid in a three-tank system. Visual measurement channel is realized using a Web camera from which the picture is exported to Matlab and using Image processing Toolbox functions the height of the levels is evaluated. Such a visual feedback has been first applied for an identification of the three-tank system and later to prove the real time implementation it has been successfully used for different control methods with the real hydraulic system. The results have been compared with traditional hydrostatic pressure sensors and the question of noise related with appropriate lightening has been discussed in details.