Manav Doshi

Guiding Professor

Prof. Debashish Chatterjee

Project Title

Quasi-interpolation for feedback synthesis of optimal control systems

Field

Control Theory, Optimal Control

Problem Statement Description

We worked on Optimal Control Systems; these systems are concerned with finding feedback to minimize a certain cost function over time. These cost functions are usually represented by an integral over time including state and feedback functions. In particular, we used an LQR system where the system dynamics are governed by a linear differential equation and the cost function is quadratic (technically positive definite in terms of vectors).

What was your motivation for choosing this field?

I was working in UMIC at the time of this project, and I felt that controls was an interesting area to learn. Hence I majorly mailed professors with these research interests

What factors did you consider while choosing professors?

Being a beginner in the area, I didn’t have a lot of options/choices in a project; further exacerbated by the fact that I was a sophomore and not a lot of professors offered projects. At the point, I just took whatever project seemed interesting to learn rather than go for a selection of professors.

Duration of the project

It was a non-credited project; lasted for about 6-7 months

Overall experience

I mainly worked on the numerical simulation of the project. The mathematical aspect of the project was done by a PhD I worked in conjunction with; though the simulation required me to understand intricate details since the approximation had to be coded from scratch. This was done by referring to certain books that my guide advised me to go through.

The time commitment in this case was moderate, though it dipped to quite low during some times. Since I had no real exposure to research and the math was fairly daunting for a sophomore student, it had a steep learning curve for me and this was demotivating at times. The major work arrived around when we were trying to publish to conferences and this required accurate simulation data and multiple iterations to be performed to ensure repeatability. During this time, work was quite high and I enjoyed it during this time since I had a concrete goal with deadlines set.

Overall, this was a pretty good experience but it made me realize that controls weren’t exactly my thing. I still look at the project as a success since we managed to publish and present it at a pretty good conference.

Key learnings

Since I went for ML roles mainly, controls didn’t help a lot but the project did strengthen my grip on linear algebra which was pretty helpful.

What course material had you gone through prior to the project?

N/A

Review of the Professor

Prof. Debashish Chatterjee is probably one of the best profs in the institute in terms of knowledge. He is very well-reputed in his field and is a great person to learn from. A thing to keep in mind while pursuing a project with him is that he expects you to put in a substantial amount of effort while working with him. He’s not strict as such, but if you’re looking for project where you have something to put on your resume; this isn’t what you should be doing.

Future plans following this project

Decided not to pursue any research in this area further