Multi-Objective Co-Design for Mission-Specific Development of Unmanned Aerial Systems

In recent years, unmanned aerial systems (UAS) are being utilized for a variety of increasingly complex tasks, including the inspection of offshore installations and the transportation of medical equipment. This has motivated the development of mission-specific dynamic design procedures. The principle of concurrent control and design, also known as co-design, extends the traditional approach of design optimization and trajectory optimization by integrating both into a single treatment. This results in coupled solutions that cannot be achieved using a traditional sequential approach. Studies have demonstrated the effectiveness of combining surrogate-assisted optimization methods, such as Bayesian optimization, with a nested formulation of the co-design problem. In the present work, we extend this approach by simultaneously treating multiple objectives. A reformulation of the Bayesian optimization framework through the use of an alternative acquisition function is fit around a trajectory optimization routine. This results in a novel framework that generates optimized designs that outperform the standard design in various metrics. This allows the designer to select a compromising design based on the system’s application type and confirms the effectiveness of the concurrent design and control procedure. The subsequent methodology is evaluated on the mission-specific design of a fixed-wing unmanned aerial system, with the aim of conducting a survey mission in challenging terrain. To model the dynamics of the aircraft, the differential flatness of the system is utilized.