Shape Tracking of Planar Hyper-Flexible Beams via Embedded PVDF Deflection Sensors

In this paper, we present an embedded shape-sensing method for soft robotics. Soft actuators and structures are paramount for safe human-robot interaction and are of interest in biomimetics, biotechnology and defense. Control of compliant manipulators is challenging because of their large deformations, hence a form of shape feedback is required. We propose a tracking system for continuum structures using deflection sensors made of polyvinylidene fluoride (PVDF), a thin piezoelectric polymer. We demonstrate that hyper-flexible beam shapes may be parameterized with a small set of basis functions and that the deflection sensors can be used to determine the weights of the basis functions. In other words, shape reconstruction of the beam shape is reduced to solving a linear interpolation problem defined by the sensors. Interpreting the sensors with this method can reduce the cumulative errors associated with piecewise methods. We demonstrate shape tracking of a hyper-flexible beam in two-dimensional (2-D) on two experimental test beds: a pneumatic bellows actuator and a hyper-flexible plastic ruler. In addition, the data were used in simulation, for systematic investigation of several aspects: number of sensors, their relative-size, and the type of basis functions used for parameterization. We demonstrate tracking capabilities and propose guidelines for the ideal sensor layout. Our approach facilitates kinematic monitoring and control of continuum manipulators and flexible structures.