Present position: Research Associate Imperial College of London
|Thesis title:||Configurable digital electronic architecture for the solution of complex mathematical models|
|Research area:||Sensors and instrumentation|
In the analysis of dynamics of many real complex systems, descriptive mathematical models permit to simulate their behavior in different specific operating conditions. This is the case of physics, engineering, finance, medicine and so on. Quite all applications involve complex mathematical models that cannot be handled by means of closed form calculations but just through approximated numerical resolution techniques. For instance this is the case of models of very fine grained systems, such as human body organs, which are represented by very sparse matrix of extremely high dimensionality (several millions of elements).
In order to perform an electronic processing of these models, the necessary hardware structures for an efficient implementation of the required algorithms for data handling are based on configurable devices, spatial computing (FPGA) and temporal computing (DSP) architectures. The deep architectural and functional diversity of the two solutions involves specific strategies of algorithm implementation and code writing for performance maximization from each one of them.
The presented research activity is devoted to the design of hardware and software architectures based on configurable devices for treatment of mathematical models of complex systems as, for instance, development and implementation of efficient methodologies for iterative resolution of systems of linear equations.