Head of the Intelligent Systems Research Laboratory
Automated Control and Systems Engineering
- Intelligent systems based signal processing, modelling and control in Biomedicine
Fuzzy and neural fuzzy systems for modelling and control
The Principle of Incompatibility, as explained by L.A. Zadeh, stipulates that interpretability and precision are incompatible properties for some topologies. The research work we have hitherto undertaken within our Group at ACSE tries to push the capabilities of particularly fuzzy logic based algorithms further in order to ‘elicit’ architectures that are capable of achieving a high level of accuracy without significant compromises on transparency.
Modelling and Decision Support in Biomedicine
We have been active in research in Biomedicine for more than 3 decades now when my mentor, Professor D.A. LINKENS, started exploring feedback control for muscle relaxation therapy in dogs and humans. When I joined his Research Group in 1985 as a research student we became the first researchers to apply successfully fuzzy logic to control the administration of muscle relaxant to humans. This was then followed by our own real-time adaptive version of SISO and MIMO Generalised Predictive Control (GPC) of muscle relaxation and Depth of Anaesthesia to humans in the operating theatre. Subsequent and current research activities close to this area include:
A Systems Engineering Approach to modelling and optimisation for metal processing: Investigations into ‘right-first-time’ production
A major aspect of this research relates to investigations into advanced multi-scale models of materials processing with predictive capabilities of the final micro-macro properties. Such models are embedded within the themes of Cellular Automata (2D and 3D), Finite Element (FE) and linguistic granules oriented neural fuzzy architectures. Further research activities concern model structure parameter optimisation and metal properties control via powerful constrained multi-objective optimisation techniques with economical and societal impacts.
Smart Tensegrity Structures: An Active Reconfigurable Structural Control Concept for Self-Healing Variable Geometry Applications
Tensegrity structures date back to the late 1940s when Buckminster Fuller used the term tensegrity as a contracted form of the two words tension and integrity to describe Kenneth Snelson’s structure. Tensegrity structures consist of two components: components in tension and those in compression, which shall be denoted as cables and bars, respectively. From a control engineering perspective, this class of structures are the ideal candidates for deployable structures] for they are capable of undergoing large displacements and can be of lightweight since these structures are obtained by optimal arrangement of material components. In addition to their lightweight and aesthetic value, tensegrity structures have other very interesting properties which motivate engineering research into these types of structures; these include mass efficiency, scalability, possibility of reliable model and accurate control, among others.