Research
Advisor: Dr. Alex Doboli
PhD Thesis: Optimized, Goal-Oriented Programming for Cyber Physical Systems
The objective of my PhD thesis is to build goal-oriented programming models and the supporting optimization algorithms and middleware routines for Cyber Physical Systems (distributed networks of reconfigurable embedded systems). Model testing and validation is done on a 36-node, grid type sensor network shown in the figure with Cypress PSoCs as the sensor nodes and on an efficient SystemC simulator.
The main challenge is to optimize aggregation schemes using Data Mining algorithms and determine dynamic switching rates between 'Design Points' representing multiple Performance-Cost trade-offs for reliable data acquisition and decision making using a Linear Programming problem describing adaptation as a Markov Decision Process (MDP), while considering performance requirements like minimizing data loss and latency, timing constraints, precision and resource constraints.
M.S. Thesis: A Sensor Network for Environmental Monitoring using PSoCs
The objective of my Master's Thesis was to develop a high performance, grid type sensor network with Reconfigurable Embedded nodes (Cypress PSoCs), specifically for environmental monitoring applications. A combination of wireless RF modules and UART links were used for data communication between PSoC nodes. Different middleware routines were programmed to manage the data structures of the sensor nodes. These routines define parameters like regions, aggregation paths, events and actuation procedures which affect the overall functionality of the network. .