An accurate self-sensing method for the control of shape memory alloy actuated flexures

Abstract

This paper proposes an accurate self-sensing method for the control of shape memory alloy (SMA) actuated flexures. SMA actuators exhibit large strain, high energy density, and can be successfully employed in flexures for miniature automation applications. A promising approach to obtain strain feedback for motion control is through an accurate self-sensing. The presented technique builds a self-sensing model based on the SMA strain to resistance curves. To overcome the inaccuracies resulting from hysteresis, the resistance curves can be influenced by sufficient pretension force to exhibit very small hysteresis gaps. The curve shapes are shown to be robust against external stiffness and temperature variations. The curves are then modeled by fitted polynomials so that strain values are directly obtained from measured SMA resistance. Accurate self-sensing control is demonstrated through step response and sinusoidal tracking experiments. Two flexural mechanisms are illustrated to show how the technique can be successfully implemented to various contexts.