Ph.D., Systems and Control, IIT Bombay & Monash University Melbourne.
Master's, Control Systems, National Institute of Technology Tiruchirappalli, Tiruchirappalli.
Bachelor's , Instrumentation & Control, Saranathan College of Engineering, Anna University, Chennai.
Experience
Assistant Professor (Contract): Aug 2022 to Present, Electrical Engineering Department at NIT‑Calicut.
Undergrad Teaching, Lab Instructor and other extra academic responsibilities.
Doctoral Fellow: Jan 2015 to Mar 2022, SysCon IIT Bombay & MAE, Monash Melbourne.
Everybody as a child would have played with Spirograph for drawing. This geometric drawing toy can create intricate patterns by combining two circular motions. These patterns fall into the class of mathematical curves known as trochoids. This trochoidal motion has been observed in planetary movements, biological spiral waves, electron behaviors, and many more naturally occurring phenomena. Engineers have leveraged such patterns for applications like atomic force microscopy, rotary pumps, cam gears, etc. One of the main characteristics of this pattern is that it is contained in a limited area, which is finely swept through by the curves in the pattern. Suppose, if a team of mobile robots wants to trace these patterns, e.g. for surveying an area, what kind of control laws will ensure that the robots trace trochoidal patterns? My doctoral dissertation explores the problem of geometric pattern formation by a group of autonomous agents. The patterns are formed by the trajectories of the agents, which are modelled either as single integrator kinematics or double integrator dynamics. The patterns are annular, which can either be closed or can cover the entire space. The proposed control laws are distributed, which is an extension of the consensus protocol. The class of patterns generated by these control laws falls within trochoidal curves. There is no restriction on the underlying graph topology. The proposed control laws are based on distributed linear consensus protocol. The necessary and sufficiency conditions for pattern formation for a more general graph topology, given that the agents are modelled as single integrators or double integrators are derived. Computer simulations and multi-robot experiments are presented to validate the theoretical findings.
Research Assistant: Jul 2011 to Aug 2014, Department of Instrumentation & Control, NIT Tiruchirappalli.
Noise is generally considered as nuisance in engineering. So engineers typically attemptto filter it out. But, quite remarkably a counter intuitive phenomenon was found in 1980 that changed the perception of noise, giving birth to a phenomenon called StochasticResonance. The main intent of my Master's dissertation is to highlight the benefits of noise from a unified framework of studying stochastic resonance and controlling stochastic resonance. In the dissertation, developed a randomized algorithm for a linear state feedback control law, designed an antilock braking system using stochastic extremum seeking. Also, provided a general scheme of controlling stochastic resonance through extremum seeking and demonstrated using a signal processing application, namely stochastic resonance-based Bayesian detectors.
Summer Intern: May 2010 to Jul 2010, Department of Computer Science & Automation, IISc, Bangalore.
Developed a code on symbolic C++ for Gaussian elimination (for any m×n case) and Bezout’s theorem.