Doctor of Philosophy
School of Mechanical, Materials and Mechatronic Engineering
Mutlu, Rahim, Electroactive polymers as smart and soft robotic actuators: an enhanced modelling methodology inspired by nature, Doctor of Philosophy thesis, School of Mechanical, Materials and Mechatronic Engineering, University of Wollongong, 2013. http://ro.uow.edu.au/theses/3883
Electroactive polymers (EAPs), also called ‘artificial muscles’, offer a significant potential to articulate soft robotic devices, especially bio-inspired robotic devices, due to their remarkable properties of compliance, low electrical energy consumption, suitability for miniaturization, biocompatibility, ability to operate in aqueous environments as well as in air, and high force output to weight ratio. The EAPs are such special materials that they can be tailored to serve as both sensors and actuators. The EAPs as actuators can be employed as either a soft robotic actuator or a soft robotic manipulator especially in the micro domain for cutting edge applications such as micro manipulation systems, medical devices with higher dexterity, soft catheters with built-in actuation, bio-inspired robotics with better mimicking properties and active compliant mechanisms. To realise these applications, there is an increasing need for enhanced modelling methodologies. The subject of this thesis is, therefore, to establish effective modelling and system identification methodologies to predict the time-varying behaviour of the smart actuators typified by polypyrrole (PPy) trilayer laminated actuators, (i.e. ionic type EAP actuators), and to apply these methodologies to predict the behaviour of new planar and spatial mechanisms articulated with these actuators.
This thesis systematically develops a whole structure bending behaviour estimation model based on the differential geometry: the backbone curve approach, which is called the ‘soft robotic model’ of the EAP actuator. Before developing a 3- dimensional (3D) soft robotic kinematic model of the EAP actuator, 2-dimensional (2D) kinematic and dynamic soft robotic models have been derived. 2D soft robotic kinematic and dynamic models have been validated experimentally for a range of electrical stimuli (0.0 − 1.0 V). Following that, image and video processing algorithms are employed to obtain the tip position data of the PPy-EAP actuator, which are used to estimate the entire bending shape or kinematic configuration of the PPy-EAP actuators as a function of time for a given electrical input. System identification techniques are also employed to identify the dynamic parameters of the PPy-EAP actuator’s soft robotic electromechanical model. A non-linear optimisation method has been proposed to solve the inverse kinematic model in order to estimate the highly non-linear bending behaviour of the PPy-EAP actuator, which is not possible to accurately estimate using the methods proposed in the literature for inverse kinematics. An adaptive boundary-constraint approach has been introduced to the optimisation algorithm to make sure that the estimated configurations match well with the real configurations of the EAP actuators under any applicable electrical input. The PPy-EAP actuators have also been designed and fabricated to form a lamina emergent compliant mechanism with built-in actuation, which generates a motion emerging out of the PPy-EAP actuator’s planar configuration (i.e. laminated). This mechanism with a conical helix topology can find a place in numerous applications such as micro-swimmers with a helix propeller inspired by bacteria swimming, micro-mixers, controllable stages, auto focusing mechanisms and micro soft robotic manipulation systems. The soft robotic kinematic modelling methodology has been applied to estimate the twining motion of this lamina emergent compliant mechanism with different spiral sizes. Further, inspired by Rapson’s slide mechanism, a continuously stable compliant mechanism based on the bending and sliding principle of the PPy-EAP actuators has been established so that the PPy-EAP actuators’ bending motion is converted into a linear motion The modelling methodology has been further validated by successfully estimating the displacement, velocity and acceleration of this mechanism. The experimental and numerical simulation results suggest that the soft robotic kinematic and dynamic modelling methodology based on the backbone curve approach is very effective in estimating the PPy-EAP actuators’ highly non-linear bending behaviour and dynamic behaviour, and in estimating the mechanism’s output articulated with the PPy-EAP actuators. The proposed methodology can easily be extended to other bending type actuators and smart structures. This study has improved our understanding of the ionic bending type EAP actuators in order to widen their application to new soft robotic applications.