Centre for Data Science & AI

As engineering projects become more and more complex, the focus of project teams now includes multiple different aspects as opposed to studying the system in isolation.

With the opportunities afforded by the fourth industrial revolution, specifically new data sources generated through Internet of Things (IoT) and Artificial Intelligence (AI)/Machine Learning (ML), most engineering projects must now take a systems-of-systems approach to design. This necessitates consideration of factors beyond the conventional single system stakeholder interests, such as network-centric operation, sustainability, and robustness in multi-role operation throughout the project lifecycle. Unfortunately, most existing engineering design tools are still discipline-specific and lack the capability to capture all instantiations of the design – from concept of operations and requirements capture through to operational acceptance tests.

In the James Watt School of Engineering, Dr David Anderson, a Reader in the Autonomous Systems and Connectivity section, has led the development of a multi-resolution, multi-agent autonomous systems simulation engine and complementary software toolset called MAVERIC - Modelling of Autonomous Vehicle Environments using Robust, Intelligent Computing – specifically to address this deficiency.

MAVERIC uses concepts from cognitive reasoning and agent modelling to provide a heterogeneous simulation architecture with physical models at multiple levels of fidelity, to assist engineers in assessing current design concepts via simulation of systems-level operational vignettes. This has a huge advantage to engineers as a system design tool, as it allows new models of key components and subsystems to be integrated into operational simulation as soon as they become available, with widespread applications across security, defence, and civilian autonomous systems. MAVERIC also utilises an Automatic Differentiation library, making the physics models in the code base differentiable physics simulation compliant – a key necessary technology for adopting AI/ML into engineering. Furthermore, MAVERIC can also be implemented in system-level design optimisation and supports our research into expanding civilian uses of autonomous systems, for example in energy generation.

Current and recent MAVERIC projects includes:

• Cooperative control between UAS and charging station to enable safe laser wireless power transfer for enhanced UAS endurance.
• Systems design optimisation of high-altitude wind turbines.
• Analysis of helicopter evasive manoeuvres to ground attack from rocket-propelled grenades and small-arms fire.
• Operational analysis of littoral marine helicopter patrols for various sensor and weapons system loadouts.
• Assessing the sensitivity of synthetic aperture radar systems to navigation errors.

---

First published: 1 January 2023