Hypersonic Vehicle Co-design

Hypersonic vehicles operate over a wide flight envelope, and their trajectories are often limited by dynamic pressure, thermal, and stability constraints. Therefore, optimising a vehicle design without consideration of the overall mission objective results in poor designs, or costly iterations between the design and GNC team.

In collaboration with the University of Melbourne and colleagues at the University of Southern Queensland, I led the development of a computational tractable framework for simultanous vehicle and trajectory optimisation. At its core this was achieved by propagating sensitivities from sub-models into a computational graph, and then combining this graph with automatic differentation tools. In 2024 we presented results of a hypersonic waverider geometry defined by 30 Bezier curve control points, subject to internal volume, static stability, and dynamic pressure constraints [1] [2]. Results show that two very different vehicles are realised for varaying mission objectices; in this case, one a maximum distance objective with obstacles requiring bank maneouvres, and a second minimum time to target objective.

Vehicle 1: Maximum range with obstructions

Vehicle 2: Minimum time to target

Vehicle designs optimised for two mission objectives (longest range with obstruction, and shortest time to target)

Optimal trajectories of both vehicles for the ‘Mission 1’ objective

Flowchart of co-design loop