Smart engineered systems that make decisions and take actions autonomously are becoming commonplace in many applications, such as self-driving cars and drones. The broader adoption of such systems is contingent on their ability to demonstrate merited trust in relation to their safe, reliable, resilient and ethical operation. Achieving success with autonomous systems deployment at scale is conditioned upon the ability to bring transformative and trustworthy innovation within their systems while meeting societal expectations for environmental sustainability and market driven cost constraints. For ubiquitous autonomous systems, this brings significant challenges not just from a design and development perspective, but also from the point of view of the fundamental uncertainty driven modelling of robustness, reliability and resilience of complex systems to demonstrate the required levels of safety.
Sustainable autonomous systems will also be expected to be durable, which becomes a significant challenge given the open heterogeneous architecture of autonomous systems. Taking the example of an automotive system, some physical systems will be expected to last 10-15 years, while some other physical components and software systems will be replaced or refreshed a lot more often.
This requires a dynamic approach to the evaluation of reliability of cyber-physical mechatronic systems, to account for the (i) evolution of the architecture of the system; (ii) differential degradation within a heterogeneous composition of the system – including both hardware and software; and (iii) situational awareness – including both the environment / mission and the current condition of the system.
Specific research projects will focus on the following topics:
(1) Global dynamic reliability and availability modelling of complex autonomous systems, based on data driven (Machine Learning) diagnostics and prognostics;
(2) Robust decisions under uncertainty considering the availability and reliability prognostics of systems;
(3) AI based management of reliability knowledge for design stage analysis – combining data extracted from human centred engineering systems analysis with machine learning based insight from system usage experience.
The University of Bradford has a well established track record of collaboration with the global automotive and aerospace industry spanning over 25 years.
You will be joining a multidisciplinary engineering and computer science team in the Advanced Automotive Analytics research unit, and you will be exposed to collaboration with our industrial partners in developing and validation of the outcomes of this research.
Funding might be available from collaborative research with industry, this will be advertised as allocated competitively; self funded students are welcome to join the lab at any time.