The proposed work focuses on the formulation of a framework for the control of chemofluidic actuators for self-powered robots. The control of these actuators is markedly different from standard types of actuators and as such requires the development of unique approaches to their motion and force control. In order to make such chemofluidic actuators broadly available to the robotics community, a formalized theoretical framework for the low-level (i.e., position, force, impedance) control of such actuators will be developed. The proposed work includes system identification for establishing a first-principles lumped-parameter model of reaction dynamics, energetic transport delays, hydraulic flow dynamics, pneumatic flow dynamics, and the transduction of thermal energy through the constitutive behavior of a gas into the mechanical domain. Nonlinear observers will be developed to enable elimination or minimization of sensor requirements, which in turn will reduce the cost of implementation of such a system. Thirdly, nonlinear control design will be fused with nonlinear optimization methods to render the multi-input single-output system energetically efficient. The result of this work will be a chemofluidic actuator modeling and control framework that will enable others in the robotics community to safely and effectively incorporate similar actuators into self-powered robots.
