My work focuses on the development and analysis of Cyber-Physical Systems (CPSs).
These are systems of highly interacting cyber/software and physical/hardware/mechanical components while in an uncertain environment, often with human-in-the-loop.
My doctoral research is on the validation of computer simulations. This requires the precise definition and modelling of validation experiments. Once an experimental frame is defined, based on validation criteria, it can be can be compared with the target system, and deemed fit for use. In this domain we are dealing with the concepts of experimental design, modeling theory, computer simulation, and software validation.
In the hardware domain, I have worked on the development and subsequent hardware implementation of the Modified-Maximum-Mean-Minimum (MoMaMeMi) filter. The MoMaMeMi filter aims to be an optimized replacement for traditional band-limited filters. We used this filter with optimized time-sliced architecture to signficantly reduce resource overheads, for multi-channel EEG signal processing. We are working on more efficient and complete implementations of pipelines incorporating the MoMaMeMi filter.
For analyzing the control and data flow of the software system which governs the behavior of the CPS a formal model has to be first developed. The Architecture Analysis and Design Language (AADL) is suitable for the safety-critical analysis of an embedded CPS with physical components like actuators and sensors tightly-coupled to embedded platform components (like threads, bus, memory, and processors) and software components. We are developing the Open-Source AADL Tool Environment-based Declarative Instance Mapping (OSATE-DIM) tool that greatly simplifies the development of other AADL-based tools. We borrow the concepts of graph theory, bidirectional transformations, object-oriented programming, and modelling, to develop OSATE-DIM.
Once a formal model has been established, the next step is formal verification of system properties like equality, safety, correctness, reachability, etc. using this formal model. We are currently working on proving semantic equivalence of two programs, through the formal model of Petri net. The Petri net is a graph, that has the inherent ability to model parallelism, something which cannot be represented easily by finite-state-machine-with-data paths (FSMDs). We have developed a scalable extension of the classical Petri net model, using graph theory and compiler concepts.
The third view of my work with cyber-physical systems is from the education-application point-of-view. In this work, we are trying to understand the role of cyber-physical systems in complementing and enhancing the modern education system. This is done through platform-based technological developments, and developments on the affective-computing front. We are currently reviewing the two domains of this important but oft overlooked interaction between humans and CPS.
Organizations that have funded my research:
Organizations that I have collaborated with:
