A Tension-Sensing and Hysteresis Modeling Platform for Tendon-Driven Continuum Manipulators

Abstract

Tendon-driven continuum manipulators (TDCMs) are useful in minimally invasive surgery because they adapt to anatomical environments unreachable by rigid instruments. However, kinematic models usually assume ideal transmissions and do not adequately compensate cable slack, friction-induced hysteresis or the lack of direct visual feedback during operation. This work presents a low-cost mechatronic platform that records a proximal load signal proportional to the tendon-induced reaction torque, without yet calibrating it as absolute tendon tension. From these data, an open-loop modeling methodology relates the time derivatives of the load signal to the reference angle, encoder angle and distal bending angle to study the hysteresis. Over repeatable cycles, the mean dead-zone width is 2.42 reference degrees, with a standard deviation of 0.02 degrees.