Modeling and vision-based control of a micro catheter head for teleoperated in-pipe inspection

The objective of this project, named MESIA (Multifunctional Micro-Endoscope Head System for Industrial Applications) is to develop an integrated multisensor system and to create intelligent sensor fusion for detection of micro-cracks. This paper is focused on the modeling of a 2-DOF pan-tilt platform actuated by Ni-Ti shape memory alloy (SMA) wires and antagonistic mechanical springs in order to control visually the CCD camera motion. A Preisach model is used and experimentally identified. The derived model is then exploited to design a position controller which compensates for the hysteretic nonlinearity. Finally, by controlling the catheter head orientation through a head motion tracker, in-pipe visual tracking has been experimentally verified from a remote site.