Session: 02-01-01: Aero-, Servo-, Thermo-Elasticity of Aircraft, Rotorcraft and Spacecraft

Paper Number: 152272

152272 - Aeroservoelastic Modeling for Trajectory Optimization of Morphing Aircrafts 

Morphing aerial vehicles, that can change their physical configurations in flight, achieves drastically increased maneuverability, energy efficiency and durability, and multifunctionality and task versatility, when compared to their non-morphing counterparts. Yet, despite the recent strides in the development of morphing aerial vehicles, these vehicles still fall short compared to the aerial prowess displayed by birds.  Birds showcase extraordinary versatility in flight, largely attributed to an instinctive coupling between neurological perception and control mechanisms, and a physical morphology for dynamic shape changes. This motivates us to undertake a concurrent design approach for morphing aerial vehicles that accounts for the coupling between autonomous functionalities (e.g., control, sensing, path planning) and vehicle physics (e.g., dynamics of fluids, structures, morphing, and flight).  To facilitate such a control co-design approach, it is critical to develop a morphing aerial vehicle model that is of sufficiently high fidelity to capture dynamic aeroservoelastic and morphing effects, as well as computationally efficient for control engineering analysis.  Only with such a model one could perform control co-design iterations in a computationally tractable manner.

In this paper, we aim to first develop an aeroservoelastic model for a nominal morphing aircraft based on the open-source package SHARPy, and apply the model to perform the trajectory optimization (TO) considering the aeroservoelastic and morphing effects.  Specifically, SHARPy is an existing well-validated aeroservoelastic code including flight dynamics, and we will modify the code to include morphing mechanisms such as deflection in dihedral of winglets.  Next, we will integrate the model into a TO framework.  Typical TO requires the knowledge of the gradients of system dynamics, which however is not available in SHARPy.  Hence we intend to use derivative-free stochastic optimization methods such as STOMP (Stochastic Trajectory Optimization for Motion Planning). In this method, a series of noisy trajectories are generated from an initial trajectory, which could be infeasible, and the cost for each of these generated trajectories is calculated by simulating them. These trajectories are then combined to get an updated lower-cost trajectory.  Finally, once the TO capability is established, we will explore the benefits of morphing in and impact of aeroservoelasticity on the maneuver of an aircraft, e.g., in obstacle avoidance scenarios.  Among the different scenarios, we will identify cases where the morphing results in beneficial reduction in the control cost and time, as well as the morphing becomes indispensable for achieving certain maneuver.

Once successful, this study shall not only reveal the benefits of morphing and aeroservoelasticity for certain aircraft configuration and scenarios, but also establish a verified model as a basis for the future exploration in the control co-design optimizations of morphing aircraft.

Presenting Author: Subarna Pudasaini Penn State

Presenting Author Biography: Subarna Pudasaini is a first-year graduate student in Aerospace Engineering at Penn State University. His research interests lie in the intersection of aeroelasticity and machine learning. He aims to develop reduced-order models for these problems and is actively exploring the potential of physics-informed machine-learning techniques to develop these models.