Morphing Smart Material Actuator Control Using Reinforcement Learning
9.1 Introduction to Smart Materials
Research in the fields of both structures and controls engineering has led to a point where it becomes necessary for materials to do more than simply provide structural support. Materials are needed that can span the gap between engineering disciplines and enable a variety of functions. Historically, materials have been researched and created for the purpose of filling some of these needs (Lagoudas 2008). The demand for vehicles that can travel through the air and even beyond the atmosphere led to the creation of materials that are both light and strong. Similarly, new demands on aerospace structures require new materials and research for the provision of multifunctional behavior.
Smart materials are a special class of materials that have the capability to act as both actuators and sensors. For example, they provide coupling between mechanical deformation and changes in temperature, electric current, and/or magnetic fields. These materials behave as energy transformers and can be engineered to act as sensors by converting variations in mechanical deformation into a non-mechanical response (e.g., thermal or electrical). Likewise, the opposite relationship allows them to act as actuators. By applying a non-mechanical input such as voltage, a smart material can respond with a mechanical deformation, even under substantial ...