Towards Trustworthy AI-driven Autonomous Systems: Multidisciplinary Design Optimisation

The development of Autonomous Collaborative Platforms (ACPs) represents a significant milestone in military aviation, offering remarkable opportunities for scalability and adaptability. Designed with a focus on mass production and modular configurations, these systems aim to deliver flexible combat capabilities across diverse mission scenarios while maintaining cost-efficiency and operational versatility. Their modular nature enables rapid reconfiguration for missions such as intelligence, surveillance, and reconnaissance (ISR), electronic warfare, ground attack, and air combat operations. However, achieving this adaptability poses challenges, such as optimizing aerodynamic and structural performance and balancing radar cross-section with operational effectiveness. The increasing complexity of modern battlefields necessitates innovative approaches that unify design and operational planning processes to ensure these platforms meet evolving strategic and tactical requirement but the existing frameworks often fall short in addressing the interconnected demands. The proposed research aims to explore the feasibility of using AI to advance the state-of-the-art in air vehicle multidisciplinary design optimization (MDO) by developing a design framework applicable to a variety of military air vehicle types. This framework will support the development and application of algorithms and methods during the conceptual and preliminary design phases, where navigating a complex multidisciplinary trade-space is essential.

The core objectives of this study are outlined as follows:
1 Identify and define the aircraft class and configuration for next-generation advanced unmanned aerial vehicles (UAVs), ensuring alignment with the operational requirements and capabilities envisioned for future Autonomous Collaborative Platforms (ACPs).
2 Conduct a comprehensive literature review focusing on multidisciplinary design optimization (MDO) techniques and the integration of artificial intelligence (AI), with particular emphasis on surrogate modelling methodologies for conceptual and preliminary UAV design.
3 Develop robust and accurate AI-based surrogate models, particularly leveraging artificial neural networks (ANNs), to significantly reduce computational complexity and analysis time, thereby enabling rapid exploration and efficient execution within multidisciplinary optimization processes.
4 Establish an integrated multidisciplinary design optimization framework including aerodynamic analysis (lift, drag, and aerodynamic efficiency), radar cross-section (RCS) evaluation (stealth characteristics and detectability), structural modelling (weight estimation, structural integrity, and payload optimisation), propulsion and weight estimation modules.
5 Demonstrate a modular and parametric UAV design framework using the developed AI-driven optimization methodologies, showcasing the capability for rapid adaptation of UAV configurations across diverse operational scenarios.
6 Demonstrate the effectiveness of the developed AI-driven design optimization framework through mission-based simulations, evaluating UAV designs in representative operational scenarios to quantify improvements in mission performance, adaptability, and operational effectiveness.
This research is expected to include theoretical analysis, modelling, computational work, and potential physical implementation in collaboration with an industry partner. It is anticipated that the physical demonstration may involve real and/or synthetic elements. The outcomes of this work are expected to help build credibility among users and customers, supporting the future deployment of AI-driven MDO capabilities within design and development toolsets.